{"title": "Semi-supervisedly Co-embedding Attributed Networks", "book": "Advances in Neural Information Processing Systems", "page_first": 6507, "page_last": 6516, "abstract": "Deep generative models (DGMs) have achieved remarkable advances. Semi-supervised variational auto-encoders (SVAE) as a classical DGM offers a principled framework to effective generalize from small labelled data to large unlabelled ones, but it is difficult to incorporate rich unstructured relationships within the multiple heterogeneous entities. In this paper, to deal with the problem, we present a semi-supervised co-embedding model for attributed networks (SCAN) based on the generalized SVAE for the heterogeneous data, which collaboratively learns low- dimensional vector representations of both nodes and attributes for partially labelled attributed networks semi-supervisedly. The node and attribute embeddings obtained in a unified manner by our SCAN can benefit not only for capturing the proximities between nodes but also the affinities between nodes and attributes. Moreover, our model also trains a discriminative network to learn the label predictive distribution of nodes. Experimental results on real-world networks demonstrate that our model yields excellent performance in a number of applications such as attribute inference, user profiling and node classification compared to the state-of-the-art baselines.", "full_text": "Semi-supervisedly Co-embedding Attributed\n\nNetworks\n\nZaiqiao Meng\u21e4\n\nDepartment of Computing Science\n\nSun Yat-sen University, and University of Glasgow\n\nzaiqiao.meng@gmail.com\n\nShangsong Liang\u2020\n\nSchool of Data and Computer Science\n\nSun Yat-sen University\n\nliangshangsong@gmail.com\n\nSchool of Data and Computer Science\n\nSchool of Data and Computer Science\n\nJinyuan Fang\n\nSun Yat-sen University\nfangjy6@gmail.com\n\nTeng Xiao\n\nSun Yat-sen University\ntengxiao01@gmail.com\n\nAbstract\n\nDeep generative models (DGMs) have achieved remarkable advances. Semi-\nsupervised variational auto-encoders (SVAE) as a classical DGM offer a principled\nframework to effectively generalize from small labelled data to large unlabelled\nones, but it is dif\ufb01cult to incorporate rich unstructured relationships within the\nmultiple heterogeneous entities. In this paper, to deal with the problem, we present\na semi-supervised co-embedding model for attributed networks (SCAN) based on\nthe generalized SVAE for heterogeneous data, which collaboratively learns low-\ndimensional vector representations of both nodes and attributes for partially labelled\nattributed networks semi-supervisedly. The node and attribute embeddings obtained\nin a uni\ufb01ed manner by our SCAN can bene\ufb01t for capturing not only the proximities\nbetween nodes but also the af\ufb01nities between nodes and attributes. Moreover, our\nmodel also trains a discriminative network to learn the label predictive distribution\nof nodes. Experimental results on real-world networks demonstrate that our model\nyields excellent performance in a number of applications such as attribute inference,\nuser pro\ufb01ling and node classi\ufb01cation compared to the state-of-the-art baselines.\n\n1\n\nIntroduction\n\nNetwork embedding has been receiving signi\ufb01cant attention in recent years due to the ubiquity\nof networks in our daily lives. The goal of network embedding is to encode entities, e.g. nodes\nand attributes, of a speci\ufb01c network into low-dimensional representation vectors, while features\nof the network, e.g. nodes\u2019 topological structure [1] and their communities [2], can be decoded\nfrom the inferred embedding vectors. Various graph-based applications, e.g. node classi\ufb01cation and\nclustering [3, 4], community detection [2, 5], link prediction [6] and expert cognition [7], have been\nshown to be able to bene\ufb01t from network embedding techniques.\nAttributed networks as one category of the most important networks are ubiquitous in a myriad of\ncritical domains, ranging from online social networks to academic collaborative networks, where rich\nattributes, e.g. nationalities of the users and journals/conferences the authors published at, describing\nthe properties of the nodes, are available. Many embedding methods for attributed networks [8, 9, 10]\nhave been proposed to learn the low-dimensional vector representations of nodes via leveraging their\n\n33rd Conference on Neural Information Processing Systems (NeurIPS 2019), Vancouver, Canada.\n\n\u21e4This work was done when the \ufb01rst author was with the Sun Yat-sen University.\n\u2020Shangsong Liang is the corresponding author.\n\n\ftopological structure and the associated attributes. In this paper, to further enhance the effectiveness\nof the learned representations, we jointly consider whatever information of the network we have\nin most real-world scenarios: topological structure and attributes of all nodes, and a few labels\nassociated with some of the nodes. We refer to such kind of networks owning these information as\npartially labelled attributed networks. Embeddings trained with partially label information can be\nused to boost the performance of related tasks, where not all the label data are available. For instance,\nin tag recommendation for social network users, tags manually labelled for the expertise of some\nusers can be utilized to train an embedding model to predict tags of those users missing their tags.\nA number of works has been proposed to learn low-dimensional representations for networks either in\nunsupervised [8, 9, 10] or semi-supervised [11, 12, 13, 14] ways. However, both of the existing unsu-\npervised and semi-supervised attributed network embedding methods learn representations for nodes\nonly, and thus are not able to capture the af\ufb01nities/similarities between nodes and attributes, which\nare the key to the success of many attributed network applications, such as attribute inference [15, 16]\nand user pro\ufb01ling [17]. Moreover, a vast majority of existing works predominately represent node\nembeddings by a single point in a low-dimensional continuous space, resulting in the fact that the\nuncertainty of nodes\u2019 representations can not be captured.\nDeep generative models (DGM) have achieved remarkable advances due to their profound basis in\ntheoretic probability and \ufb02exible and scalable optimizing ability of deep neural networks. Semi-\nsupervised variational auto-encoders [18, 19] as a classical DGM offer a principled framework to\neffective generalize from small labelled data to large unlabelled ones, but have limited applicability\nto incorporate rich unstructured relationships within the multiple heterogeneous entities. To alleviate\nthe aforementioned problems, we introduce a Semi-supervised Co-embedding Attributed Network\nalgorithm, SCAN, built based on the generalized SVAE for the heterogeneous data, that is able\nto co-embed both attributes and nodes of partial labelled networks in the same semantic space.\nSCAN collaboratively learns low-dimensional vector representations of both attributes and nodes\nin the same semantic space in a semi-supervised way, such that the af\ufb01nities between them can be\neffectively measured and the partial labels of nodes can be fully utilized. The learned nodes\u2019 and\nattributes\u2019 embeddings in the same semantic space are, in turn, utilized to boost the performance of\nmany down-stream applications, e.g., user pro\ufb01ling [17], where the relevance of users (nodes) and\nkeywords (attributes) can be directly measured by, e.g., cosine similarity. In our SCAN, we infer\nthe embeddings of both nodes and attributes and represent them by means of Gaussian distributions.\nWith the natural property of latent presentations (i.e., Gaussian embeddings in our case), SCAN\ninnately represents their uncertainty with the corresponding variances of the inferred embeddings.\nOur contributions can be summarized as follows: 1(1) We generalize SVAE to the heterogeneous\ndata and propose a novel semi-supervised co-embedding model for attributed networks, SCAN, to\ncollaboratively learn representations of nodes and attributes in the same space such that proximities\nbetween nodes as well as af\ufb01nities between nodes and attributes of networks can be effectively\nmeasured. (2) Our SVAE model jointly optimizes a variational evidence lower bound consisting\nof \ufb01ve atomic observations based on two entities and two relationships, to obtain the Gaussian\nembeddings of nodes and attributes and a discriminator for node classi\ufb01cation, where the mean\nvectors denote the position of nodes and attributes and the variances capture uncertainty of their\nrepresentations. (3) We perform extensive experiments on real-world attributed networks to verify\nthe effectiveness of our embedding model in terms of three network mining tasks, and the results\ndemonstrate that our model is able to signi\ufb01cantly outperforms state-of-the-art methods.\n\n2 Related Work\n\nMany unsupervised representation learning methods have been proposed to embed various networks\ninto low-dimensional vectors of nodes. Approaches such as DeepWalk [1], node2vec [2] and LINE [3]\nlearn embeddings for plain networks, where only topological structure information is utilized based\non random walks or edge sampling. The Structural Deep Network Embedding [20] model embeds a\nnetwork by capturing the highly non-linear network structure so as to preserve the global and local\nstructure of the network. Wang et al. [5] incorporate the community structure of network into result\nembeddings to preserve both of the microscopic and community structures. Some other work obtains\nembeddings for non-plain networks with rich auxiliary information, such as labels, node attributes\n\n1The code of our SCAN is publicly available from: https://github.com/mengzaiqiao/SCAN.\n\n2\n\n\fand text contents, in addition to the topological structure networks [21, 22]. Hamilton et al. [10]\npropose the GraphSAGE model that learns node representations by sampling and aggregating features\nfrom nodes\u2019 local neighbourhoods. Zhang et al. [9] and Gao et al. [23] propose their customized\ndeep neural network architectures to learn node embeddings, while capturing the underlying high\nnon-linearity in both topological structure and attributes. CAN [24] is a model that unsupervisedly\nlearns embedding for both nodes and attributes by their customized DGM. Their results show that\ncombining different types of auxiliary information, rather than using only the topological features,\ncan provide different insights of embedding of nodes. Recently, a few approaches embed nodes by\ndistributions and capture the uncertainty of the embeddings [25, 26, 6]. For example, the KG2E\nmodel [25] represents each entity/relation of knowledge graphs as a Gaussian distribution. In terms\nof embedding attributed networks, Aleksandar et al. [6] embed each node as Gaussian distribution\naccording to the energy-based loss of personalized ranking formulation.\nRecently, learning representations of entities for networks by semi-supervised ways has also been\nwidely studied. Planetoid [11] is a network representation learning model for semi-supervised node\nclassi\ufb01cation but not for capturing af\ufb01nities between embeddings of nodes and attributes. Kipf\net.al [12] propose a graph convolutional neural network model for attributed networks for semi-\nsupervised classi\ufb01cation task, which also outputs embedding for nodes only. Liang et.al [13] propose\na semi-supervised learning model, called SEANO, which utilizes dual-input and dual-output deep\nneural networks to learn node embedding encompassing information related to structure, attributes,\nand labels explicitly alleviating noise effects from outliers. More recently, a semi-supervised deep\ngenerative model for attributed network embedding has been proposed [14], which applies the\ngenerative adversarial nets (GANs) to generates fake samples in low-density areas in networks and\nleverages clustering property to help classi\ufb01cation. However, it still learns embeddings for nodes\nonly. Therefore, the embeddings obtained by these models might not be directly utilized in other\napplications such as attribute inference and user pro\ufb01ling.\n\n3 Semi-supervised Variational Auto-encoder\n\nThe Semi-supervised Variational Auto-encoders (SVAEs) are a kind of generative semi-supervised\nmodels for partially labelled data that learn representations of data by jointly training a probabilistic\nencoder, probabilistic decoder as well as a label predictive neural networks [18, 19], while the small\nlabelled data sets can be generalized to large unlabelled ones. SVAEs have received a lot of attention\ndue to their wide applicability in domains such as text classi\ufb01cation [27], machine translation [28] and\nsyntactic annotation [29]. In this section, we \ufb01st review the construction of an SVAE for homogeneous\nentities, and then extend it to the setting of heterogeneous data.\n\n3.1 Semi-supervised Learning for Homogeneous Data\nWe \ufb01rst consider the homogeneous data that appear as pairs O = {(x1, y1) , . . . , (xN , yN )}, with\nyi 2{ 0, 1}1\u21e5K being the one-hot vector representing the label of i-th observation xi 2 RF where K\nis the number of classes and F is the feature dimension. In order to incorporate unlabeled data in the\nlearning process, previous work on deep generative semi-supervised models optimize a variational\nbound J on the marginal likelihood for N1 labelled data points and N2 unlabelled data points:\n\nJ =\n\nL(xi, yi) +\n\nU(xj),\n\n(1)\n\nN1Xi=1\n\nN2Xj=1\n\nwhere L(xi, yi) is the evidence lower bound (ELBO) for a labelled data point and U(xj) is the\nELBO for an unlabelled one. Normally, a SVAE model also incorporates a classi\ufb01cation loss into Eq.\n1 to impose the label posterior inference model q(y|x) to be able to act as a classi\ufb01er:\n(2)\nwhere hyper-parameter \u21b5 controls the weight between generative and purely discriminative learning.\nWhile these models are conceptually simple and easy to train, one potential limitation of this approach\nis that marginalization over all K classes becomes prohibitively expensive for a large number of\nclasses. We leave the discussion of the solution for this limitation to subsection 4.2.\nAnother limitation is that they only learn representations for the homogeneous observations, which\nare generated independently according to a homogeneous prior, ignoring that heterogeneous data\n\nJ \u21b5 = J + \u21b5 \u00b7 Eepl(x,y) [ log q(y|x)] ,\n\n3\n\n\fobservations are ubiquitous in the real world. For example, in recommender system items and users\nare two independent observations but they are associated with purchase or interaction behaviours.\nMany efforts have been paid to co-embedding multiple entities for heterogeneous systems in a fully\nunsupervised learning procedure [24, 30]. However, to our knowledge, no model in the literature can\nlearn embeddings for heterogeneous data with multiple entities in a semi-supervised way. Hence, in\nthe next subsection we will present a model that allows to learn multiple entity observations having\narbitrary conditional dependency structures and arbitrary label for each type of entity.\n\n3.2 Semi-supervised Learning for Heterogeneous Data\n\nWe consider a generalized form of semi-supervised model for heterogeneous data that appear as\ntriples instead of pairs. Let O = (X ,Y,R) be a triple set of the heterogeneous observation data,\nwhere X = (X1,\u00b7\u00b7\u00b7 , XT ) is the entity set with multiple types (e.g. T types), R = {rij | xg\ni 2\nj 2 Xh} is a set containing relationships with rij being the relationship strength of two entities\nXg, xh\nin same/different types, and Y = (Y1,\u00b7\u00b7\u00b7 , YT ) is the partially labelled set for all types of entities.\nWithout loss of generality, we can formulate the semi-supervised learning model by \ufb01rst considering\nonly two entity types, e.g. Xg and Xh. We let Oij = (xg\nj , rij, Yl) be an atomic data point of the\nobservation where Yl are the labels of entities if there are, Zij = (zg\nj , Yu) be the collection of\nlatent variables of the two entities where Yu are the predicted labels for entities without a label, and\nY = (Yu, Yl). Fij = ((xg\nj )) is the conditional variables of variational posterior where \ni ), (xh\nis a function taking entities\u2019 feature as input for \ufb01ltering posteriors. Then, the logarithm marginal\nlikelihood of Oij can be written as:\n\ni , xh\n\ni , zh\n\nlog p(Oij) Eq(Zij|Fij )[log p\u2713(Oij,Zij) log q(Zij | Fij)],\n\n(3)\n\nwhere the joint distribution, i.e. p\u2713(Oij,Zij), can be represented as:\n\ni , xh\n\np\u2713(Oij,Zij) = p\u2713(xg\n\n(4)\nIn most real-world scenarios, such as recommender systems [31] and network embeddings [24],\npeople only concern about the reconstruction of relations between entities rather than the feature\nof the entities, thus p\u2713(xg\nj , rij | Zij, Y) can be simpli\ufb01ed to be p\u2713(rij | Zij, Y). With the\nmean-\ufb01eld approximate inference of the variational posterior q(Zij | Fij), Eq. 3 can be written as:\n\nj , rij | Zij, Y)p(Y)p(zg\n\ni )p(zh\nj ).\n\ni , xh\n\nlog p(Oij) Eq(Zij|Fij )[log p\u2713(rij |Zij, Y)] KL(q(Zij | Fij) | p(zg\n\ni )p(zh\n\nj ))\n\n, L(Oij),\n\n(5)\nwhere L(Oij) is denoted as the ELBO of Oij. Note that can be different types depending on\nthe heterogeneity of the observation entities. Incorporating an additional weighted discriminative\ncomponent as in Eq. 2 leads to the following lower bound for overall observations:\n\nJ \u21b5(O) = Xrij2R\n\nL(Oij) + \u21b5 \u00b7 Eepl(x,y) [ log q(y|x)] .\n\n(6)\n\n4 Deep Semi-supervised Attributed Co-embedding\n\nWe now turn to the semi-supervised learning problem for the partially labelled attributed networks.\n\n4.1 Problem De\ufb01nition\nLet G = (V,A, A, X, Yl) be a Partially Labelled Attributed Network, with V and A being the\nsets of nodes and attributes, respectively, A 2 RN\u21e5N and X 2 RN\u21e5M being the weighted adjacency\nmatrix and node attribute matrix, respectively, where N = |V| is the number of nodes and M = |A|\nis the number of attributes. Yl is the label matrix representing the node labels. Since most nodes\u2019\nlabels are unknown, V can be divided into two subsets: labelled nodes V l and unlabelled nodes V u\nwith their label matrix being Yl and Yu respectively.\nThe problem of Semi-supervised Attributed Network Co-embedding is de\ufb01ned as:\n\n4\n\n\fProblem. Given a partially labelled attributed network G, learn a mapping function \u2305 that satis\ufb01es\nthe following in a semi-supervised way:\n\nG = (V,A, A, X, Yl) \u2305! Zn, Za, Yu,\n\n(7)\nsuch that both network structure, node attributes and the partial labels can be preserved as much as\npossible by Zn, Za and Yu, where Zn 2 RN\u21e5D and Za 2 RM\u21e5D represent the latent representation\nmatrices for all the nodes and attributes, respectively, and Yu represents the learned labels for all\nthe unlabelled nodes. Here D is the size of the embeddings.\n\n4.2 The Semi-supervised Co-embedding Model\n\nInput Data\n\nEmbeddings & Labels\n\nReconstructed Data\n\nNode Label: Yl\ni\n\nAAACC3icbZC7SgNBFIZnvcZ4i1raDAmCVdgVQbEK2FiIRDAXSdYwOzmbDJm9MHNWDEt6G1/FxkIRW1/AzrdxskmhiT8MfPznHM6c34ul0Gjb39bC4tLyympuLb++sbm1XdjZresoURxqPJKRanpMgxQh1FCghGasgAWehIY3OB/XG/egtIjCGxzG4AasFwpfcIbG6hSKbYQHTK+iLtBL5oE8oyOaeZ6f3o7uZEd0CiW7bGei8+BMoUSmqnYKX+1uxJMAQuSSad1y7BjdlCkUXMIo3040xIwPWA9aBkMWgHbT7JYRPTBOl/qRMi9Emrm/J1IWaD0MPNMZMOzr2drY/K/WStA/dVMRxglCyCeL/ERSjOg4GNoVCjjKoQHGlTB/pbzPFONo4subEJzZk+ehflR2DF8flyr2NI4c2SdFckgcckIq5IJUSY1w8kieySt5s56sF+vd+pi0LljTmT3yR9bnD8n7mtI=\nAAACC3icbZC7SgNBFIZnvcZ4i1raDAmCVdgVQbEK2FiIRDAXSdYwOzmbDJm9MHNWDEt6G1/FxkIRW1/AzrdxskmhiT8MfPznHM6c34ul0Gjb39bC4tLyympuLb++sbm1XdjZresoURxqPJKRanpMgxQh1FCghGasgAWehIY3OB/XG/egtIjCGxzG4AasFwpfcIbG6hSKbYQHTK+iLtBL5oE8oyOaeZ6f3o7uZEd0CiW7bGei8+BMoUSmqnYKX+1uxJMAQuSSad1y7BjdlCkUXMIo3040xIwPWA9aBkMWgHbT7JYRPTBOl/qRMi9Emrm/J1IWaD0MPNMZMOzr2drY/K/WStA/dVMRxglCyCeL/ERSjOg4GNoVCjjKoQHGlTB/pbzPFONo4subEJzZk+ehflR2DF8flyr2NI4c2SdFckgcckIq5IJUSY1w8kieySt5s56sF+vd+pi0LljTmT3yR9bnD8n7mtI=\nAAACC3icbZC7SgNBFIZnvcZ4i1raDAmCVdgVQbEK2FiIRDAXSdYwOzmbDJm9MHNWDEt6G1/FxkIRW1/AzrdxskmhiT8MfPznHM6c34ul0Gjb39bC4tLyympuLb++sbm1XdjZresoURxqPJKRanpMgxQh1FCghGasgAWehIY3OB/XG/egtIjCGxzG4AasFwpfcIbG6hSKbYQHTK+iLtBL5oE8oyOaeZ6f3o7uZEd0CiW7bGei8+BMoUSmqnYKX+1uxJMAQuSSad1y7BjdlCkUXMIo3040xIwPWA9aBkMWgHbT7JYRPTBOl/qRMi9Emrm/J1IWaD0MPNMZMOzr2drY/K/WStA/dVMRxglCyCeL/ERSjOg4GNoVCjjKoQHGlTB/pbzPFONo4subEJzZk+ehflR2DF8flyr2NI4c2SdFckgcckIq5IJUSY1w8kieySt5s56sF+vd+pi0LljTmT3yR9bnD8n7mtI=\nAAACC3icbZC7SgNBFIZnvcZ4i1raDAmCVdgVQbEK2FiIRDAXSdYwOzmbDJm9MHNWDEt6G1/FxkIRW1/AzrdxskmhiT8MfPznHM6c34ul0Gjb39bC4tLyympuLb++sbm1XdjZresoURxqPJKRanpMgxQh1FCghGasgAWehIY3OB/XG/egtIjCGxzG4AasFwpfcIbG6hSKbYQHTK+iLtBL5oE8oyOaeZ6f3o7uZEd0CiW7bGei8+BMoUSmqnYKX+1uxJMAQuSSad1y7BjdlCkUXMIo3040xIwPWA9aBkMWgHbT7JYRPTBOl/qRMi9Emrm/J1IWaD0MPNMZMOzr2drY/K/WStA/dVMRxglCyCeL/ERSjOg4GNoVCjjKoQHGlTB/pbzPFONo4subEJzZk+ehflR2DF8flyr2NI4c2SdFckgcckIq5IJUSY1w8kieySt5s56sF+vd+pi0LljTmT3yR9bnD8n7mtI=\n\nAAACMnicbVBNSwMxEM3Wr1q/qh69BIvgQWRXBMVTwYtepIKtH20t2XTWRrPZJZkVy7K/yYu/RPCgB0W8+iNMaxFrfRB4894MmXl+LIVB1312cmPjE5NT+enCzOzc/EJxcalmokRzqPJIRvrMZwakUFBFgRLOYg0s9CWc+jf7Pf/0FrQRkTrBbgzNkF0pEQjO0Eqt4mED4Q7To6gNezSjjZBhxw/Si+xStcTGUHn9U55nl/KX2y+vW8WSu+n2QUeJNyAlMkClVXxstCOehKCQS2ZM3XNjbKZMo+ASskIjMRAzfsOuoG6pYiGYZto/OaNrVmnTINL2KaR99fdEykJjuqFvO3tbmr9eT/zPqycY7DZToeIEQfHvj4JEUoxoLz/aFho4yq4ljGthd6W8wzTjaFMu2BC8vyePktrWpmf58Xap7A7iyJMVskrWiUd2SJkckAqpEk7uyRN5JW/Og/PivDsf3605ZzCzTIbgfH4BJZKr3A==\nAAACMnicbVBNSwMxEM3Wr1q/qh69BIvgQWRXBMVTwYtepIKtH20t2XTWRrPZJZkVy7K/yYu/RPCgB0W8+iNMaxFrfRB4894MmXl+LIVB1312cmPjE5NT+enCzOzc/EJxcalmokRzqPJIRvrMZwakUFBFgRLOYg0s9CWc+jf7Pf/0FrQRkTrBbgzNkF0pEQjO0Eqt4mED4Q7To6gNezSjjZBhxw/Si+xStcTGUHn9U55nl/KX2y+vW8WSu+n2QUeJNyAlMkClVXxstCOehKCQS2ZM3XNjbKZMo+ASskIjMRAzfsOuoG6pYiGYZto/OaNrVmnTINL2KaR99fdEykJjuqFvO3tbmr9eT/zPqycY7DZToeIEQfHvj4JEUoxoLz/aFho4yq4ljGthd6W8wzTjaFMu2BC8vyePktrWpmf58Xap7A7iyJMVskrWiUd2SJkckAqpEk7uyRN5JW/Og/PivDsf3605ZzCzTIbgfH4BJZKr3A==\nAAACMnicbVBNSwMxEM3Wr1q/qh69BIvgQWRXBMVTwYtepIKtH20t2XTWRrPZJZkVy7K/yYu/RPCgB0W8+iNMaxFrfRB4894MmXl+LIVB1312cmPjE5NT+enCzOzc/EJxcalmokRzqPJIRvrMZwakUFBFgRLOYg0s9CWc+jf7Pf/0FrQRkTrBbgzNkF0pEQjO0Eqt4mED4Q7To6gNezSjjZBhxw/Si+xStcTGUHn9U55nl/KX2y+vW8WSu+n2QUeJNyAlMkClVXxstCOehKCQS2ZM3XNjbKZMo+ASskIjMRAzfsOuoG6pYiGYZto/OaNrVmnTINL2KaR99fdEykJjuqFvO3tbmr9eT/zPqycY7DZToeIEQfHvj4JEUoxoLz/aFho4yq4ljGthd6W8wzTjaFMu2BC8vyePktrWpmf58Xap7A7iyJMVskrWiUd2SJkckAqpEk7uyRN5JW/Og/PivDsf3605ZzCzTIbgfH4BJZKr3A==\nAAAB2XicbZDNSgMxFIXv1L86Vq1rN8EiuCozbnQpuHFZwbZCO5RM5k4bmskMyR2hDH0BF25EfC93vo3pz0JbDwQ+zknIvSculLQUBN9ebWd3b/+gfugfNfzjk9Nmo2fz0gjsilzl5jnmFpXU2CVJCp8LgzyLFfbj6f0i77+gsTLXTzQrMMr4WMtUCk7O6oyaraAdLMW2IVxDC9YaNb+GSS7KDDUJxa0dhEFBUcUNSaFw7g9LiwUXUz7GgUPNM7RRtRxzzi6dk7A0N+5oYkv394uKZ9bOstjdzDhN7Ga2MP/LBiWlt1EldVESarH6KC0Vo5wtdmaJNChIzRxwYaSblYkJN1yQa8Z3HYSbG29D77odOn4MoA7ncAFXEMIN3MEDdKALAhJ4hXdv4r15H6uuat66tDP4I+/zBzjGijg=\nAAACJ3icbVDLSgMxFL3j2/qqbt0Ei+BCZMaN4kpwoxtRsFpta8mkdzSayQzJHbEM801u/BLBhS4U8UdMH0h9HAicR0LuPWGqpCXff/FGRsfGJyanpkszs3PzC+XF2VObZEZgVSQqMbWQW1RSY5UkKaylBnkcKjwLb/e6+dkdGisTfUKdFJsxv9IykoKTs1rlgwbhPeWHSRt3WMEaMafrMMovikvdkus/5M23PC8u1VDakzetcsXf8Htgf0kwIBUY4KhVfmq0E5HFqEkobm098FNq5tyQFAqLUiOzmHJxy6+w7qjmMdpm3lu5YKvOabMoMe5oYj13+EXOY2s7cehudqe0v7Ou+V9WzyjabuZSpxmhFv2PokwxSli3P9aWBgWpjiNcGOlmZeKaGy7ItVxyJQS/V/5LTjc3AsePfZiCZViBNQhgC3ZhH46gCgIe4Bne4N179F69j35dI96gtyX4Ae/zC88eqlE=\nAAACJ3icbVDLSgMxFL3j2/qqbt0Ei+BCZMaN4kpwoxtRsFpta8mkdzSayQzJHbEM801u/BLBhS4U8UdMH0h9HAicR0LuPWGqpCXff/FGRsfGJyanpkszs3PzC+XF2VObZEZgVSQqMbWQW1RSY5UkKaylBnkcKjwLb/e6+dkdGisTfUKdFJsxv9IykoKTs1rlgwbhPeWHSRt3WMEaMafrMMovikvdkus/5M23PC8u1VDakzetcsXf8Htgf0kwIBUY4KhVfmq0E5HFqEkobm098FNq5tyQFAqLUiOzmHJxy6+w7qjmMdpm3lu5YKvOabMoMe5oYj13+EXOY2s7cehudqe0v7Ou+V9WzyjabuZSpxmhFv2PokwxSli3P9aWBgWpjiNcGOlmZeKaGy7ItVxyJQS/V/5LTjc3AsePfZiCZViBNQhgC3ZhH46gCgIe4Bne4N179F69j35dI96gtyX4Ae/zC88eqlE=\nAAACMnicbVBNSwMxEM36WetX1aOXYBE8iOx6UTwVvOhFKtj60daSTWdt2mx2SWbFsuxv8uIvETzoQRGv/gjTWsRWHwTevDdDZp4fS2HQdZ+dicmp6ZnZ3Fx+fmFxabmwslo1UaI5VHgkI33hMwNSKKigQAkXsQYW+hLO/e5h3z+/BW1EpM6wF0MjZDdKBIIztFKzcFxHuMP0JGrBAc1oPWTY9oP0KrtWTbE9UnZ+ysvsWv5yB2WnWSi6O+4A9C/xhqRIhig3C4/1VsSTEBRyyYypeW6MjZRpFFxClq8nBmLGu+wGapYqFoJppIOTM7pplRYNIm2fQjpQf0+kLDSmF/q2s7+lGff64n9eLcFgv5EKFScIin9/FCSSYkT7+dGW0MBR9ixhXAu7K+VtphlHm3LehuCNn/yXVHd3PMtP3WLJHcaRI+tkg2wRj+yREjkiZVIhnNyTJ/JK3pwH58V5dz6+Wyec4cwaGYHz+QUkUqvY\nAAACMnicbVBNSwMxEM3Wr1q/qh69BIvgQWRXBMVTwYtepIKtH20t2XTWRrPZJZkVy7K/yYu/RPCgB0W8+iNMaxFrfRB4894MmXl+LIVB1312cmPjE5NT+enCzOzc/EJxcalmokRzqPJIRvrMZwakUFBFgRLOYg0s9CWc+jf7Pf/0FrQRkTrBbgzNkF0pEQjO0Eqt4mED4Q7To6gNezSjjZBhxw/Si+xStcTGUHn9U55nl/KX2y+vW8WSu+n2QUeJNyAlMkClVXxstCOehKCQS2ZM3XNjbKZMo+ASskIjMRAzfsOuoG6pYiGYZto/OaNrVmnTINL2KaR99fdEykJjuqFvO3tbmr9eT/zPqycY7DZToeIEQfHvj4JEUoxoLz/aFho4yq4ljGthd6W8wzTjaFMu2BC8vyePktrWpmf58Xap7A7iyJMVskrWiUd2SJkckAqpEk7uyRN5JW/Og/PivDsf3605ZzCzTIbgfH4BJZKr3A==\nAAACMnicbVBNSwMxEM3Wr1q/qh69BIvgQWRXBMVTwYtepIKtH20t2XTWRrPZJZkVy7K/yYu/RPCgB0W8+iNMaxFrfRB4894MmXl+LIVB1312cmPjE5NT+enCzOzc/EJxcalmokRzqPJIRvrMZwakUFBFgRLOYg0s9CWc+jf7Pf/0FrQRkTrBbgzNkF0pEQjO0Eqt4mED4Q7To6gNezSjjZBhxw/Si+xStcTGUHn9U55nl/KX2y+vW8WSu+n2QUeJNyAlMkClVXxstCOehKCQS2ZM3XNjbKZMo+ASskIjMRAzfsOuoG6pYiGYZto/OaNrVmnTINL2KaR99fdEykJjuqFvO3tbmr9eT/zPqycY7DZToeIEQfHvj4JEUoxoLz/aFho4yq4ljGthd6W8wzTjaFMu2BC8vyePktrWpmf58Xap7A7iyJMVskrWiUd2SJkckAqpEk7uyRN5JW/Og/PivDsf3605ZzCzTIbgfH4BJZKr3A==\nAAACMnicbVBNSwMxEM3Wr1q/qh69BIvgQWRXBMVTwYtepIKtH20t2XTWRrPZJZkVy7K/yYu/RPCgB0W8+iNMaxFrfRB4894MmXl+LIVB1312cmPjE5NT+enCzOzc/EJxcalmokRzqPJIRvrMZwakUFBFgRLOYg0s9CWc+jf7Pf/0FrQRkTrBbgzNkF0pEQjO0Eqt4mED4Q7To6gNezSjjZBhxw/Si+xStcTGUHn9U55nl/KX2y+vW8WSu+n2QUeJNyAlMkClVXxstCOehKCQS2ZM3XNjbKZMo+ASskIjMRAzfsOuoG6pYiGYZto/OaNrVmnTINL2KaR99fdEykJjuqFvO3tbmr9eT/zPqycY7DZToeIEQfHvj4JEUoxoLz/aFho4yq4ljGthd6W8wzTjaFMu2BC8vyePktrWpmf58Xap7A7iyJMVskrWiUd2SJkckAqpEk7uyRN5JW/Og/PivDsf3605ZzCzTIbgfH4BJZKr3A==\nAAACMnicbVBNSwMxEM3Wr1q/qh69BIvgQWRXBMVTwYtepIKtH20t2XTWRrPZJZkVy7K/yYu/RPCgB0W8+iNMaxFrfRB4894MmXl+LIVB1312cmPjE5NT+enCzOzc/EJxcalmokRzqPJIRvrMZwakUFBFgRLOYg0s9CWc+jf7Pf/0FrQRkTrBbgzNkF0pEQjO0Eqt4mED4Q7To6gNezSjjZBhxw/Si+xStcTGUHn9U55nl/KX2y+vW8WSu+n2QUeJNyAlMkClVXxstCOehKCQS2ZM3XNjbKZMo+ASskIjMRAzfsOuoG6pYiGYZto/OaNrVmnTINL2KaR99fdEykJjuqFvO3tbmr9eT/zPqycY7DZToeIEQfHvj4JEUoxoLz/aFho4yq4ljGthd6W8wzTjaFMu2BC8vyePktrWpmf58Xap7A7iyJMVskrWiUd2SJkckAqpEk7uyRN5JW/Og/PivDsf3605ZzCzTIbgfH4BJZKr3A==\nAAACMnicbVBNSwMxEM3Wr1q/qh69BIvgQWRXBMVTwYtepIKtH20t2XTWRrPZJZkVy7K/yYu/RPCgB0W8+iNMaxFrfRB4894MmXl+LIVB1312cmPjE5NT+enCzOzc/EJxcalmokRzqPJIRvrMZwakUFBFgRLOYg0s9CWc+jf7Pf/0FrQRkTrBbgzNkF0pEQjO0Eqt4mED4Q7To6gNezSjjZBhxw/Si+xStcTGUHn9U55nl/KX2y+vW8WSu+n2QUeJNyAlMkClVXxstCOehKCQS2ZM3XNjbKZMo+ASskIjMRAzfsOuoG6pYiGYZto/OaNrVmnTINL2KaR99fdEykJjuqFvO3tbmr9eT/zPqycY7DZToeIEQfHvj4JEUoxoLz/aFho4yq4ljGthd6W8wzTjaFMu2BC8vyePktrWpmf58Xap7A7iyJMVskrWiUd2SJkckAqpEk7uyRN5JW/Og/PivDsf3605ZzCzTIbgfH4BJZKr3A==\nAAACMnicbVBNSwMxEM3Wr1q/qh69BIvgQWRXBMVTwYtepIKtH20t2XTWRrPZJZkVy7K/yYu/RPCgB0W8+iNMaxFrfRB4894MmXl+LIVB1312cmPjE5NT+enCzOzc/EJxcalmokRzqPJIRvrMZwakUFBFgRLOYg0s9CWc+jf7Pf/0FrQRkTrBbgzNkF0pEQjO0Eqt4mED4Q7To6gNezSjjZBhxw/Si+xStcTGUHn9U55nl/KX2y+vW8WSu+n2QUeJNyAlMkClVXxstCOehKCQS2ZM3XNjbKZMo+ASskIjMRAzfsOuoG6pYiGYZto/OaNrVmnTINL2KaR99fdEykJjuqFvO3tbmr9eT/zPqycY7DZToeIEQfHvj4JEUoxoLz/aFho4yq4ljGthd6W8wzTjaFMu2BC8vyePktrWpmf58Xap7A7iyJMVskrWiUd2SJkckAqpEk7uyRN5JW/Og/PivDsf3605ZzCzTIbgfH4BJZKr3A==\n\nNode: Zn\n\ni , Zn\n\nj , Yl\n\ni, Yl\nj\n\nAAAACAHicbVDLSsNAFL3xWeMr6sKFm8EiuCpJN7oRK25cVrAPaEKZTCbt0MmDmYlQQkD8FTcuFHEl+Bnu/BsnbRfaemGYwzn3cs89fsqZVLb9bSwtr6yurVc2zM2t7Z1da2+/LZNMENoiCU9E18eSchbTlmKK024qKI58Tjv+6LrUO/dUSJbEd2qcUi/Cg5iFjGClqb516PoJD+Q40l/uRlgN/TC/Koq+VbVr9qTQInBmoHr5YV48AECzb325QUKyiMaKcCxlz7FT5eVYKEY4LUw3kzTFZIQHtKdhjCMqvXxyQIFONBOgMBH6xQpN2N8TOY5k6VF3lhblvFaS/2m9TIXnXs7iNFM0JtNFYcaRSlCZBgqYoETxsQaYCKa9IjLEAhOlMzN1CM78yYugXa85ds25daqNOkyrAkdwDKfgwBk04Aaa0AICBTzBC7waj8az8Wa8T1uXjNnMAfwp4/MHFTSYzQ==\n\nAAACAHicbVC7TsMwFHXKq4RXgIGBxaJCYqqSLrAgilgYi0QfUhtVjuO0Vh07sh2kKsrCr3RhACEmJD6DBfE3OG0HaLmS5aNz7tU99wQJo0q77rdVWlldW98ob9pb2zu7e87+QUuJVGLSxIIJ2QmQIoxy0tRUM9JJJEFxwEg7GN0UevuBSEUFv9fjhPgxGnAaUYy0ofrOUS8QLFTj2HxZL0Z6GETZdZ73nYpbdacFl4E3B5Wrd/symXzZjb7z2QsFTmPCNWZIqa7nJtrPkNQUM5LbvVSRBOERGpCugRzFRPnZ9IAcnhomhJGQ5nENp+zviQzFqvBoOguLalEryP+0bqqjCz+jPEk14Xi2KEoZ1AIWacCQSoI1GxuAsKTGK8RDJBHWJjPbhOAtnrwMWrWq51a9O69Sr4FZlcExOAFnwAPnoA5uQQM0AQY5mIBn8GI9Wk/Wq/U2ay1Z85lD8Kesjx8Gw5pB\nAAACAHicbVC7TsMwFHXKq4RXgIGBxaJCYqqSLrAgilgYi0QfUhtVjuO0Vh07sh2kKsrCr3RhACEmJD6DBfE3OG0HaLmS5aNz7tU99wQJo0q77rdVWlldW98ob9pb2zu7e87+QUuJVGLSxIIJ2QmQIoxy0tRUM9JJJEFxwEg7GN0UevuBSEUFv9fjhPgxGnAaUYy0ofrOUS8QLFTj2HxZL0Z6GETZdZ73nYpbdacFl4E3B5Wrd/symXzZjb7z2QsFTmPCNWZIqa7nJtrPkNQUM5LbvVSRBOERGpCugRzFRPnZ9IAcnhomhJGQ5nENp+zviQzFqvBoOguLalEryP+0bqqjCz+jPEk14Xi2KEoZ1AIWacCQSoI1GxuAsKTGK8RDJBHWJjPbhOAtnrwMWrWq51a9O69Sr4FZlcExOAFnwAPnoA5uQQM0AQY5mIBn8GI9Wk/Wq/U2ay1Z85lD8Kesjx8Gw5pB\nAAAB2XicbZDNSgMxFIXv1L86Vq1rN8EiuCozbnQpuHFZwbZCO5RM5k4bmskMyR2hDH0BF25EfC93vo3pz0JbDwQ+zknIvSculLQUBN9ebWd3b/+gfugfNfzjk9Nmo2fz0gjsilzl5jnmFpXU2CVJCp8LgzyLFfbj6f0i77+gsTLXTzQrMMr4WMtUCk7O6oyaraAdLMW2IVxDC9YaNb+GSS7KDDUJxa0dhEFBUcUNSaFw7g9LiwUXUz7GgUPNM7RRtRxzzi6dk7A0N+5oYkv394uKZ9bOstjdzDhN7Ga2MP/LBiWlt1EldVESarH6KC0Vo5wtdmaJNChIzRxwYaSblYkJN1yQa8Z3HYSbG29D77odBu3wMYA6nMMFXEEIN3AHD9CBLghI4BXevYn35n2suqp569LO4I+8zx84xIo4\nAAAB9XicbVC9TsMwGPxS/kopEFgYWCwqJKYqYYERxMJYJPojNVHlOE5r1bEj20Gqoiy8CgsDCPEqbLwNTtsBWk6yfLrzJ993UcaZNp737dQ2Nre2d+q7jb3m/sGhe9TsaZkrQrtEcqkGEdaUM0G7hhlOB5miOI047UfTu8rvP1GlmRSPZpbRMMVjwRJGsLHSyD0JIsljPUvtVQQpNpMoKW7LcuS2vLY3B1on/pK0YInOyP0KYknylApDONZ66HuZCQusDCOclo0g1zTDZIrHdGipwCnVYTFfoETnVolRIpU9wqC5+nuiwKmuMtqXVUS96lXif94wN8l1WDCR5YYKsvgoyTkyElVtoJgpSgyfWYKJYjYrIhOsMDG2s4YtwV9deZ30Ltu+1/YfPKjDKZzBBfhwBTdwDx3oAoESXuAN3p1n59X5WNRVc5a9HcMfOJ8/QVqVuA==\nAAAB9XicbVC9TsMwGPxS/kopEFgYWCwqJKYqYYERxMJYJPojNVHlOE5r1bEj20Gqoiy8CgsDCPEqbLwNTtsBWk6yfLrzJ993UcaZNp737dQ2Nre2d+q7jb3m/sGhe9TsaZkrQrtEcqkGEdaUM0G7hhlOB5miOI047UfTu8rvP1GlmRSPZpbRMMVjwRJGsLHSyD0JIsljPUvtVQQpNpMoKW7LcuS2vLY3B1on/pK0YInOyP0KYknylApDONZ66HuZCQusDCOclo0g1zTDZIrHdGipwCnVYTFfoETnVolRIpU9wqC5+nuiwKmuMtqXVUS96lXif94wN8l1WDCR5YYKsvgoyTkyElVtoJgpSgyfWYKJYjYrIhOsMDG2s4YtwV9deZ30Ltu+1/YfPKjDKZzBBfhwBTdwDx3oAoESXuAN3p1n59X5WNRVc5a9HcMfOJ8/QVqVuA==\nAAACAHicbVC7TsMwFL3hWcorwMDAElEhMVVJFxiLWBiLRB9SE1WO47RWHTuyHaQqysKvsDCAECufwcbf4LQZoOVKlo/OuVf33BOmjCrtut/W2vrG5tZ2bae+u7d/cGgfHfeUyCQmXSyYkIMQKcIoJ11NNSODVBKUhIz0w+ltqfcfiVRU8Ac9S0mQoDGnMcVIG2pkn/qhYJGaJebL/QTpSRjnN0Uxshtu052Xswq8CjSgqs7I/vIjgbOEcI0ZUmrouakOciQ1xYwUdT9TJEV4isZkaCBHCVFBPj+gcC4MEzmxkOZx7czZ3xM5SlTp0XSWFtWyVpL/acNMx9dBTnmaacLxYlGcMUcLp0zDiagkWLOZAQhLarw6eIIkwtpkVjcheMsnr4Jeq+m5Te/ebbRbVRw1OINzuAQPrqANd9CBLmAo4Ble4c16sl6sd+tj0bpmVTMn8Keszx+7TZcP\nAAACAHicbVC7TsMwFL3hWcorwMDAElEhMVVJFxiLWBiLRB9SE1WO47RWHTuyHaQqysKvsDCAECufwcbf4LQZoOVKlo/OuVf33BOmjCrtut/W2vrG5tZ2bae+u7d/cGgfHfeUyCQmXSyYkIMQKcIoJ11NNSODVBKUhIz0w+ltqfcfiVRU8Ac9S0mQoDGnMcVIG2pkn/qhYJGaJebL/QTpSRjnN0Uxshtu052Xswq8CjSgqs7I/vIjgbOEcI0ZUmrouakOciQ1xYwUdT9TJEV4isZkaCBHCVFBPj+gcC4MEzmxkOZx7czZ3xM5SlTp0XSWFtWyVpL/acNMx9dBTnmaacLxYlGcMUcLp0zDiagkWLOZAQhLarw6eIIkwtpkVjcheMsnr4Jeq+m5Te/ea7RbVRw1OINzuAQPrqANd9CBLmAo4Ble4c16sl6sd+tj0bpmVTMn8Keszx+7nZcQ\nAAACAHicbVC7TsMwFHXKq4RXgIGBxaJCYqqSLrAgilgYi0QfUhtVjuO0Vh07sh2kKsrCr3RhACEmJD6DBfE3OG0HaLmS5aNz7tU99wQJo0q77rdVWlldW98ob9pb2zu7e87+QUuJVGLSxIIJ2QmQIoxy0tRUM9JJJEFxwEg7GN0UevuBSEUFv9fjhPgxGnAaUYy0ofrOUS8QLFTj2HxZL0Z6GETZdZ73nYpbdacFl4E3B5Wrd/symXzZjb7z2QsFTmPCNWZIqa7nJtrPkNQUM5LbvVSRBOERGpCugRzFRPnZ9IAcnhomhJGQ5nENp+zviQzFqvBoOguLalEryP+0bqqjCz+jPEk14Xi2KEoZ1AIWacCQSoI1GxuAsKTGK8RDJBHWJjPbhOAtnrwMWrWq51a9O69Sr4FZlcExOAFnwAPnoA5uQQM0AQY5mIBn8GI9Wk/Wq/U2ay1Z85lD8Kesjx8Gw5pB\nAAACAHicbVC7TsMwFHXKq4RXgIGBxaJCYqqSLrAgilgYi0QfUhtVjuO0Vh07sh2kKsrCr3RhACEmJD6DBfE3OG0HaLmS5aNz7tU99wQJo0q77rdVWlldW98ob9pb2zu7e87+QUuJVGLSxIIJ2QmQIoxy0tRUM9JJJEFxwEg7GN0UevuBSEUFv9fjhPgxGnAaUYy0ofrOUS8QLFTj2HxZL0Z6GETZdZ73nYpbdacFl4E3B5Wrd/symXzZjb7z2QsFTmPCNWZIqa7nJtrPkNQUM5LbvVSRBOERGpCugRzFRPnZ9IAcnhomhJGQ5nENp+zviQzFqvBoOguLalEryP+0bqqjCz+jPEk14Xi2KEoZ1AIWacCQSoI1GxuAsKTGK8RDJBHWJjPbhOAtnrwMWrWq51a9O69Sr4FZlcExOAFnwAPnoA5uQQM0AQY5mIBn8GI9Wk/Wq/U2ay1Z85lD8Kesjx8Gw5pB\nAAACAHicbVC7TsMwFHXKq4RXgIGBxaJCYqqSLrAgilgYi0QfUhtVjuO0Vh07sh2kKsrCr3RhACEmJD6DBfE3OG0HaLmS5aNz7tU99wQJo0q77rdVWlldW98ob9pb2zu7e87+QUuJVGLSxIIJ2QmQIoxy0tRUM9JJJEFxwEg7GN0UevuBSEUFv9fjhPgxGnAaUYy0ofrOUS8QLFTj2HxZL0Z6GETZdZ73nYpbdacFl4E3B5Wrd/symXzZjb7z2QsFTmPCNWZIqa7nJtrPkNQUM5LbvVSRBOERGpCugRzFRPnZ9IAcnhomhJGQ5nENp+zviQzFqvBoOguLalEryP+0bqqjCz+jPEk14Xi2KEoZ1AIWacCQSoI1GxuAsKTGK8RDJBHWJjPbhOAtnrwMWrWq51a9O69Sr4FZlcExOAFnwAPnoA5uQQM0AQY5mIBn8GI9Wk/Wq/U2ay1Z85lD8Kesjx8Gw5pB\nAAACAHicbVC7TsMwFHXKq4RXgIGBxaJCYqqSLrAgilgYi0QfUhtVjuO0Vh07sh2kKsrCr3RhACEmJD6DBfE3OG0HaLmS5aNz7tU99wQJo0q77rdVWlldW98ob9pb2zu7e87+QUuJVGLSxIIJ2QmQIoxy0tRUM9JJJEFxwEg7GN0UevuBSEUFv9fjhPgxGnAaUYy0ofrOUS8QLFTj2HxZL0Z6GETZdZ73nYpbdacFl4E3B5Wrd/symXzZjb7z2QsFTmPCNWZIqa7nJtrPkNQUM5LbvVSRBOERGpCugRzFRPnZ9IAcnhomhJGQ5nENp+zviQzFqvBoOguLalEryP+0bqqjCz+jPEk14Xi2KEoZ1AIWacCQSoI1GxuAsKTGK8RDJBHWJjPbhOAtnrwMWrWq51a9O69Sr4FZlcExOAFnwAPnoA5uQQM0AQY5mIBn8GI9Wk/Wq/U2ay1Z85lD8Kesjx8Gw5pB\nAAACAHicbVC7TsMwFL3hWcorwMDAElEhMVVJFxiLWBiLRB9SE1WO47RWHTuyHaQqysKvsDCAECufwcbf4LQZoOVKlo/OuVf33BOmjCrtut/W2vrG5tZ2bae+u7d/cGgfHfeUyCQmXSyYkIMQKcIoJ11NNSODVBKUhIz0w+ltqfcfiVRU8Ac9S0mQoDGnMcVIG2pkn/qhYJGaJebL/QTpSRjnN0Uxshtu052Xswq8CjSgqs7I/vIjgbOEcI0ZUmrouakOciQ1xYwUdT9TJEV4isZkaCBHCVFBPj+gcC4MEzmxkOZx7czZ3xM5SlTp0XSWFtWyVpL/acNMx9dBTnmaacLxYlGcMUcLp0zDiagkWLOZAQhLarw6eIIkwtpkVjcheMsnr4Jeq+m5Te/ea7RbVRw1OINzuAQPrqANd9CBLmAo4Ble4c16sl6sd+tj0bpmVTMn8Keszx+7nZcQ\n\nAAACAnicbVDLSsNAFJ3UV62vqCtxM1gEVyUpgi4LblxWsA9pYplMJu3QyUyYmQglBDf+ihsXirj1K9z5N07aLLT1wjCHc+7lnnuChFGlHefbqqysrq1vVDdrW9s7u3v2/kFXiVRi0sGCCdkPkCKMctLRVDPSTyRBccBIL5hcFXrvgUhFBb/V04T4MRpxGlGMtKGG9pEXCBaqaWy+zIuRHgdRdpfn92xo152GMyu4DNwS1EFZ7aH95YUCpzHhGjOk1MB1Eu1nSGqKGclrXqpIgvAEjcjAQI5iovxsdkIOTw0TwkhI87iGM/b3RIZiVbg0nYVJtagV5H/aINXRpZ9RnqSacDxfFKUMagGLPGBIJcGaTQ1AWFLjFeIxkghrk1rNhOAunrwMus2Ga/DNeb3VLOOogmNwAs6ACy5AC1yDNugADB7BM3gFb9aT9WK9Wx/z1opVzhyCP2V9/gBxd5gJ\nAAACAnicbVDLSsNAFJ3UV62vqCtxM1gEVyUpgi4LblxWsA9pYplMJu3QyUyYmQglBDf+ihsXirj1K9z5N07aLLT1wjCHc+7lnnuChFGlHefbqqysrq1vVDdrW9s7u3v2/kFXiVRi0sGCCdkPkCKMctLRVDPSTyRBccBIL5hcFXrvgUhFBb/V04T4MRpxGlGMtKGG9pEXCBaqaWy+zIuRHgdRdpfn92xo152GMyu4DNwS1EFZ7aH95YUCpzHhGjOk1MB1Eu1nSGqKGclrXqpIgvAEjcjAQI5iovxsdkIOTw0TwkhI87iGM/b3RIZiVbg0nYVJtagV5H/aINXRpZ9RnqSacDxfFKUMagGLPGBIJcGaTQ1AWFLjFeIxkghrk1rNhOAunrwMus2Ga/DNeb3VLOOogmNwAs6ACy5AC1yDNugADB7BM3gFb9aT9WK9Wx/z1opVzhyCP2V9/gBxd5gJ\nAAACAnicbVDLSsNAFJ3UV62vqCtxM1gEVyUpgi4LblxWsA9pYplMJu3QyUyYmQglBDf+ihsXirj1K9z5N07aLLT1wjCHc+7lnnuChFGlHefbqqysrq1vVDdrW9s7u3v2/kFXiVRi0sGCCdkPkCKMctLRVDPSTyRBccBIL5hcFXrvgUhFBb/V04T4MRpxGlGMtKGG9pEXCBaqaWy+zIuRHgdRdpfn92xo152GMyu4DNwS1EFZ7aH95YUCpzHhGjOk1MB1Eu1nSGqKGclrXqpIgvAEjcjAQI5iovxsdkIOTw0TwkhI87iGM/b3RIZiVbg0nYVJtagV5H/aINXRpZ9RnqSacDxfFKUMagGLPGBIJcGaTQ1AWFLjFeIxkghrk1rNhOAunrwMus2Ga/DNeb3VLOOogmNwAs6ACy5AC1yDNugADB7BM3gFb9aT9WK9Wx/z1opVzhyCP2V9/gBxd5gJ\nAAACAnicbVDLSsNAFJ3UV62vqCtxM1gEVyUpgi4LblxWsA9pYplMJu3QyUyYmQglBDf+ihsXirj1K9z5N07aLLT1wjCHc+7lnnuChFGlHefbqqysrq1vVDdrW9s7u3v2/kFXiVRi0sGCCdkPkCKMctLRVDPSTyRBccBIL5hcFXrvgUhFBb/V04T4MRpxGlGMtKGG9pEXCBaqaWy+zIuRHgdRdpfn92xo152GMyu4DNwS1EFZ7aH95YUCpzHhGjOk1MB1Eu1nSGqKGclrXqpIgvAEjcjAQI5iovxsdkIOTw0TwkhI87iGM/b3RIZiVbg0nYVJtagV5H/aINXRpZ9RnqSacDxfFKUMagGLPGBIJcGaTQ1AWFLjFeIxkghrk1rNhOAunrwMus2Ga/DNeb3VLOOogmNwAs6ACy5AC1yDNugADB7BM3gFb9aT9WK9Wx/z1opVzhyCP2V9/gBxd5gJ\n\nYl\nYu\n\nAAAB9XicbVDLSgMxFL3xWeur6tJNsAiuykwRdFlw47KCfUg7LZk004ZmMkOSUcow/+HGhSJu/Rd3/o2ZdhbaeiBwOOde7snxY8G1cZxvtLa+sbm1Xdop7+7tHxxWjo7bOkoUZS0aiUh1faKZ4JK1DDeCdWPFSOgL1vGnN7nfeWRK80jem1nMvJCMJQ84JcZKg35IzMQP0odskCbZsFJ1as4ceJW4BalCgeaw8tUfRTQJmTRUEK17rhMbLyXKcCpYVu4nmsWETsmY9SyVJGTaS+epM3xulREOImWfNHiu/t5ISaj1LPTtZJ5SL3u5+J/XS0xw7aVcxolhki4OBYnAJsJ5BXjEFaNGzCwhVHGbFdMJUYQaW1TZluAuf3mVtOs11/K7y2qjXtRRglM4gwtw4QoacAtNaAEFBc/wCm/oCb2gd/SxGF1Dxc4J/AH6/AEoHZLg\nAAAB9XicbVDLSgMxFL3xWeur6tJNsAiuykwRdFlw47KCfUg7LZk004ZmMkOSUcow/+HGhSJu/Rd3/o2ZdhbaeiBwOOde7snxY8G1cZxvtLa+sbm1Xdop7+7tHxxWjo7bOkoUZS0aiUh1faKZ4JK1DDeCdWPFSOgL1vGnN7nfeWRK80jem1nMvJCMJQ84JcZKg35IzMQP0odskCbZsFJ1as4ceJW4BalCgeaw8tUfRTQJmTRUEK17rhMbLyXKcCpYVu4nmsWETsmY9SyVJGTaS+epM3xulREOImWfNHiu/t5ISaj1LPTtZJ5SL3u5+J/XS0xw7aVcxolhki4OBYnAJsJ5BXjEFaNGzCwhVHGbFdMJUYQaW1TZluAuf3mVtOs11/K7y2qjXtRRglM4gwtw4QoacAtNaAEFBc/wCm/oCb2gd/SxGF1Dxc4J/AH6/AEoHZLg\nAAAB9XicbVDLSgMxFL3xWeur6tJNsAiuykwRdFlw47KCfUg7LZk004ZmMkOSUcow/+HGhSJu/Rd3/o2ZdhbaeiBwOOde7snxY8G1cZxvtLa+sbm1Xdop7+7tHxxWjo7bOkoUZS0aiUh1faKZ4JK1DDeCdWPFSOgL1vGnN7nfeWRK80jem1nMvJCMJQ84JcZKg35IzMQP0odskCbZsFJ1as4ceJW4BalCgeaw8tUfRTQJmTRUEK17rhMbLyXKcCpYVu4nmsWETsmY9SyVJGTaS+epM3xulREOImWfNHiu/t5ISaj1LPTtZJ5SL3u5+J/XS0xw7aVcxolhki4OBYnAJsJ5BXjEFaNGzCwhVHGbFdMJUYQaW1TZluAuf3mVtOs11/K7y2qjXtRRglM4gwtw4QoacAtNaAEFBc/wCm/oCb2gd/SxGF1Dxc4J/AH6/AEoHZLg\nAAAB2XicbZDNSgMxFIXv1L86Vq1rN8EiuCozbnQpuHFZwbZCO5RM5k4bmskMyR2hDH0BF25EfC93vo3pz0JbDwQ+zknIvSculLQUBN9ebWd3b/+gfugfNfzjk9Nmo2fz0gjsilzl5jnmFpXU2CVJCp8LgzyLFfbj6f0i77+gsTLXTzQrMMr4WMtUCk7O6oyaraAdLMW2IVxDC9YaNb+GSS7KDDUJxa0dhEFBUcUNSaFw7g9LiwUXUz7GgUPNM7RRtRxzzi6dk7A0N+5oYkv394uKZ9bOstjdzDhN7Ga2MP/LBiWlt1EldVESarH6KC0Vo5wtdmaJNChIzRxwYaSblYkJN1yQa8Z3HYSbG29D77odOn4MoA7ncAFXEMIN3MEDdKALAhJ4hXdv4r15H6uuat66tDP4I+/zBzjGijg=\nAAAB6nicbZC9TsMwFIVvyl8pBQori0WFxFQlLDBWYmEsEv1BbVo57k1r1XEi2wFVUd6DhQGEeCA23gan7QAtV7L06Rxb9/gEieDauO63U9ra3tndK+9XDqqHR8e1k2pHx6li2GaxiFUvoBoFl9g23AjsJQppFAjsBrPbwu8+odI8lg9mnqAf0YnkIWfUWGk4iKiZBmH2mA+zNB/V6m7DXQzZBG8FdVhNa1T7GoxjlkYoDRNU677nJsbPqDKcCcwrg1RjQtmMTrBvUdIItZ8tUufkwipjEsbKHmnIQv39IqOR1vMosDeLlHrdK8T/vH5qwhs/4zJJDUq2XBSmgpiYFBWQMVfIjJhboExxm5WwKVWUGVtUxZbgrX95EzpXDc/yvQtlOINzuAQPrqEJd9CCNjBQ8AJv8O48O6/Ox7KukrPq7RT+jPP5A9kSkZA=\nAAAB6nicbZC9TsMwFIVvyl8pBQori0WFxFQlLDBWYmEsEv1BbVo57k1r1XEi2wFVUd6DhQGEeCA23gan7QAtV7L06Rxb9/gEieDauO63U9ra3tndK+9XDqqHR8e1k2pHx6li2GaxiFUvoBoFl9g23AjsJQppFAjsBrPbwu8+odI8lg9mnqAf0YnkIWfUWGk4iKiZBmH2mA+zNB/V6m7DXQzZBG8FdVhNa1T7GoxjlkYoDRNU677nJsbPqDKcCcwrg1RjQtmMTrBvUdIItZ8tUufkwipjEsbKHmnIQv39IqOR1vMosDeLlHrdK8T/vH5qwhs/4zJJDUq2XBSmgpiYFBWQMVfIjJhboExxm5WwKVWUGVtUxZbgrX95EzpXDc/yvQtlOINzuAQPrqEJd9CCNjBQ8AJv8O48O6/Ox7KukrPq7RT+jPP5A9kSkZA=\nAAAB9XicbVC7TsMwFL0pr1JeBUYWiwqJqUq6wFiJhbFI9IHatHJcp7VqO5HtgKoo/8HCAEKs/Asbf4PTZoCWI1k6Oude3eMTxJxp47rfTmljc2t7p7xb2ds/ODyqHp90dJQoQtsk4pHqBVhTziRtG2Y47cWKYhFw2g1mN7nffaRKs0jem3lMfYEnkoWMYGOl4UBgMw3C9CEbpkk2qtbcursAWideQWpQoDWqfg3GEUkElYZwrHXfc2Pjp1gZRjjNKoNE0xiTGZ7QvqUSC6r9dJE6QxdWGaMwUvZJgxbq740UC63nIrCTeUq96uXif14/MeG1nzIZJ4ZKsjwUJhyZCOUVoDFTlBg+twQTxWxWRKZYYWJsURVbgrf65XXSadQ9y+/cWrNR1FGGMziHS/DgCppwCy1oAwEFz/AKb86T8+K8Ox/L0ZJT7JzCHzifPybdktw=\nAAAB9XicbVDLSgMxFL3xWeur6tJNsAiuykwRdFlw47KCfUg7LZk004ZmMkOSUcow/+HGhSJu/Rd3/o2ZdhbaeiBwOOde7snxY8G1cZxvtLa+sbm1Xdop7+7tHxxWjo7bOkoUZS0aiUh1faKZ4JK1DDeCdWPFSOgL1vGnN7nfeWRK80jem1nMvJCMJQ84JcZKg35IzMQP0odskCbZsFJ1as4ceJW4BalCgeaw8tUfRTQJmTRUEK17rhMbLyXKcCpYVu4nmsWETsmY9SyVJGTaS+epM3xulREOImWfNHiu/t5ISaj1LPTtZJ5SL3u5+J/XS0xw7aVcxolhki4OBYnAJsJ5BXjEFaNGzCwhVHGbFdMJUYQaW1TZluAuf3mVtOs11/K7y2qjXtRRglM4gwtw4QoacAtNaAEFBc/wCm/oCb2gd/SxGF1Dxc4J/AH6/AEoHZLg\nAAAB9XicbVDLSgMxFL3xWeur6tJNsAiuykwRdFlw47KCfUg7LZk004ZmMkOSUcow/+HGhSJu/Rd3/o2ZdhbaeiBwOOde7snxY8G1cZxvtLa+sbm1Xdop7+7tHxxWjo7bOkoUZS0aiUh1faKZ4JK1DDeCdWPFSOgL1vGnN7nfeWRK80jem1nMvJCMJQ84JcZKg35IzMQP0odskCbZsFJ1as4ceJW4BalCgeaw8tUfRTQJmTRUEK17rhMbLyXKcCpYVu4nmsWETsmY9SyVJGTaS+epM3xulREOImWfNHiu/t5ISaj1LPTtZJ5SL3u5+J/XS0xw7aVcxolhki4OBYnAJsJ5BXjEFaNGzCwhVHGbFdMJUYQaW1TZluAuf3mVtOs11/K7y2qjXtRRglM4gwtw4QoacAtNaAEFBc/wCm/oCb2gd/SxGF1Dxc4J/AH6/AEoHZLg\nAAAB9XicbVDLSgMxFL3xWeur6tJNsAiuykwRdFlw47KCfUg7LZk004ZmMkOSUcow/+HGhSJu/Rd3/o2ZdhbaeiBwOOde7snxY8G1cZxvtLa+sbm1Xdop7+7tHxxWjo7bOkoUZS0aiUh1faKZ4JK1DDeCdWPFSOgL1vGnN7nfeWRK80jem1nMvJCMJQ84JcZKg35IzMQP0odskCbZsFJ1as4ceJW4BalCgeaw8tUfRTQJmTRUEK17rhMbLyXKcCpYVu4nmsWETsmY9SyVJGTaS+epM3xulREOImWfNHiu/t5ISaj1LPTtZJ5SL3u5+J/XS0xw7aVcxolhki4OBYnAJsJ5BXjEFaNGzCwhVHGbFdMJUYQaW1TZluAuf3mVtOs11/K7y2qjXtRRglM4gwtw4QoacAtNaAEFBc/wCm/oCb2gd/SxGF1Dxc4J/AH6/AEoHZLg\nAAAB9XicbVDLSgMxFL3xWeur6tJNsAiuykwRdFlw47KCfUg7LZk004ZmMkOSUcow/+HGhSJu/Rd3/o2ZdhbaeiBwOOde7snxY8G1cZxvtLa+sbm1Xdop7+7tHxxWjo7bOkoUZS0aiUh1faKZ4JK1DDeCdWPFSOgL1vGnN7nfeWRK80jem1nMvJCMJQ84JcZKg35IzMQP0odskCbZsFJ1as4ceJW4BalCgeaw8tUfRTQJmTRUEK17rhMbLyXKcCpYVu4nmsWETsmY9SyVJGTaS+epM3xulREOImWfNHiu/t5ISaj1LPTtZJ5SL3u5+J/XS0xw7aVcxolhki4OBYnAJsJ5BXjEFaNGzCwhVHGbFdMJUYQaW1TZluAuf3mVtOs11/K7y2qjXtRRglM4gwtw4QoacAtNaAEFBc/wCm/oCb2gd/SxGF1Dxc4J/AH6/AEoHZLg\nAAAB9XicbVDLSgMxFL3xWeur6tJNsAiuykwRdFlw47KCfUg7LZk004ZmMkOSUcow/+HGhSJu/Rd3/o2ZdhbaeiBwOOde7snxY8G1cZxvtLa+sbm1Xdop7+7tHxxWjo7bOkoUZS0aiUh1faKZ4JK1DDeCdWPFSOgL1vGnN7nfeWRK80jem1nMvJCMJQ84JcZKg35IzMQP0odskCbZsFJ1as4ceJW4BalCgeaw8tUfRTQJmTRUEK17rhMbLyXKcCpYVu4nmsWETsmY9SyVJGTaS+epM3xulREOImWfNHiu/t5ISaj1LPTtZJ5SL3u5+J/XS0xw7aVcxolhki4OBYnAJsJ5BXjEFaNGzCwhVHGbFdMJUYQaW1TZluAuf3mVtOs11/K7y2qjXtRRglM4gwtw4QoacAtNaAEFBc/wCm/oCb2gd/SxGF1Dxc4J/AH6/AEoHZLg\nAAAB9XicbVDLSgMxFL3xWeur6tJNsAiuykwRdFlw47KCfUg7LZk004ZmMkOSUcow/+HGhSJu/Rd3/o2ZdhbaeiBwOOde7snxY8G1cZxvtLa+sbm1Xdop7+7tHxxWjo7bOkoUZS0aiUh1faKZ4JK1DDeCdWPFSOgL1vGnN7nfeWRK80jem1nMvJCMJQ84JcZKg35IzMQP0odskCbZsFJ1as4ceJW4BalCgeaw8tUfRTQJmTRUEK17rhMbLyXKcCpYVu4nmsWETsmY9SyVJGTaS+epM3xulREOImWfNHiu/t5ISaj1LPTtZJ5SL3u5+J/XS0xw7aVcxolhki4OBYnAJsJ5BXjEFaNGzCwhVHGbFdMJUYQaW1TZluAuf3mVtOs11/K7y2qjXtRRglM4gwtw4QoacAtNaAEFBc/wCm/oCb2gd/SxGF1Dxc4J/AH6/AEoHZLg\n\nInfered Label: Yl\ni\n\nAAACDnicbZC7SgNBFIZnvcZ4i1raDIaAVdgVQbEK2ChYRDAXSWKYnZxNhszOLjNnxbDkCWx8FRsLRWyt7XwbJ5dCE38Y+PjPOZw5vx9LYdB1v52FxaXlldXMWnZ9Y3NrO7ezWzVRojlUeCQjXfeZASkUVFCghHqsgYW+hJrfPx/Va/egjYjUDQ5iaIWsq0QgOENrtXOFJsIDppcqAA0desV8kGd0SMe2H6S3wzvZFu1c3i26Y9F58KaQJ1OV27mvZifiSQgKuWTGNDw3xlbKNAouYZhtJgZixvusCw2LioVgWun4nCEtWKdDg0jbp5CO3d8TKQuNGYS+7QwZ9sxsbWT+V2skGJy2UqHiBEHxyaIgkRQjOsqGdoQGjnJggXEt7F8p7zHNONoEszYEb/bkeageFT3L18f5kjuNI0P2yQE5JB45ISVyQcqkQjh5JM/klbw5T86L8+58TFoXnOnMHvkj5/MHPBWcJw==\nAAACDnicbZC7SgNBFIZnvcZ4i1raDIaAVdgVQbEK2ChYRDAXSWKYnZxNhszOLjNnxbDkCWx8FRsLRWyt7XwbJ5dCE38Y+PjPOZw5vx9LYdB1v52FxaXlldXMWnZ9Y3NrO7ezWzVRojlUeCQjXfeZASkUVFCghHqsgYW+hJrfPx/Va/egjYjUDQ5iaIWsq0QgOENrtXOFJsIDppcqAA0desV8kGd0SMe2H6S3wzvZFu1c3i26Y9F58KaQJ1OV27mvZifiSQgKuWTGNDw3xlbKNAouYZhtJgZixvusCw2LioVgWun4nCEtWKdDg0jbp5CO3d8TKQuNGYS+7QwZ9sxsbWT+V2skGJy2UqHiBEHxyaIgkRQjOsqGdoQGjnJggXEt7F8p7zHNONoEszYEb/bkeageFT3L18f5kjuNI0P2yQE5JB45ISVyQcqkQjh5JM/klbw5T86L8+58TFoXnOnMHvkj5/MHPBWcJw==\nAAACDnicbZC7SgNBFIZnvcZ4i1raDIaAVdgVQbEK2ChYRDAXSWKYnZxNhszOLjNnxbDkCWx8FRsLRWyt7XwbJ5dCE38Y+PjPOZw5vx9LYdB1v52FxaXlldXMWnZ9Y3NrO7ezWzVRojlUeCQjXfeZASkUVFCghHqsgYW+hJrfPx/Va/egjYjUDQ5iaIWsq0QgOENrtXOFJsIDppcqAA0desV8kGd0SMe2H6S3wzvZFu1c3i26Y9F58KaQJ1OV27mvZifiSQgKuWTGNDw3xlbKNAouYZhtJgZixvusCw2LioVgWun4nCEtWKdDg0jbp5CO3d8TKQuNGYS+7QwZ9sxsbWT+V2skGJy2UqHiBEHxyaIgkRQjOsqGdoQGjnJggXEt7F8p7zHNONoEszYEb/bkeageFT3L18f5kjuNI0P2yQE5JB45ISVyQcqkQjh5JM/klbw5T86L8+58TFoXnOnMHvkj5/MHPBWcJw==\nAAACDnicbZC7SgNBFIZnvcZ4i1raDIaAVdgVQbEK2ChYRDAXSWKYnZxNhszOLjNnxbDkCWx8FRsLRWyt7XwbJ5dCE38Y+PjPOZw5vx9LYdB1v52FxaXlldXMWnZ9Y3NrO7ezWzVRojlUeCQjXfeZASkUVFCghHqsgYW+hJrfPx/Va/egjYjUDQ5iaIWsq0QgOENrtXOFJsIDppcqAA0desV8kGd0SMe2H6S3wzvZFu1c3i26Y9F58KaQJ1OV27mvZifiSQgKuWTGNDw3xlbKNAouYZhtJgZixvusCw2LioVgWun4nCEtWKdDg0jbp5CO3d8TKQuNGYS+7QwZ9sxsbWT+V2skGJy2UqHiBEHxyaIgkRQjOsqGdoQGjnJggXEt7F8p7zHNONoEszYEb/bkeageFT3L18f5kjuNI0P2yQE5JB45ISVyQcqkQjh5JM/klbw5T86L8+58TFoXnOnMHvkj5/MHPBWcJw==\n\nNode Feature: Fn\ni\n\nAAACDXicbZC7SgNBFIZnvcZ4i1raDEbBKuyKoFgFhGAlEcwFkhhmJ2eTIbOzy8xZMSx5ARtfxcZCEVt7O9/GyaXQxB8GPv5zDmfO78dSGHTdb2dhcWl5ZTWzll3f2Nzazu3sVk2UaA4VHslI131mQAoFFRQooR5rYKEvoeb3L0f12j1oIyJ1i4MYWiHrKhEIztBa7dxhE+EB0+uoA7QEDBMNF3RIx64fpKXhnWqLdi7vFtyx6Dx4U8iTqcrt3FezE/EkBIVcMmManhtjK2UaBZcwzDYTAzHjfdaFhkXFQjCtdHzNkB5Zp0ODSNunkI7d3xMpC40ZhL7tDBn2zGxtZP5XayQYnLdSoeIEQfHJoiCRFCM6ioZ2hAaOcmCBcS3sXynvMc042gCzNgRv9uR5qJ4UPMs3p/miO40jQ/bJATkmHjkjRXJFyqRCOHkkz+SVvDlPzovz7nxMWhec6cwe+SPn8wd9r5vB\nAAACDXicbZC7SgNBFIZnvcZ4i1raDEbBKuyKoFgFhGAlEcwFkhhmJ2eTIbOzy8xZMSx5ARtfxcZCEVt7O9/GyaXQxB8GPv5zDmfO78dSGHTdb2dhcWl5ZTWzll3f2Nzazu3sVk2UaA4VHslI131mQAoFFRQooR5rYKEvoeb3L0f12j1oIyJ1i4MYWiHrKhEIztBa7dxhE+EB0+uoA7QEDBMNF3RIx64fpKXhnWqLdi7vFtyx6Dx4U8iTqcrt3FezE/EkBIVcMmManhtjK2UaBZcwzDYTAzHjfdaFhkXFQjCtdHzNkB5Zp0ODSNunkI7d3xMpC40ZhL7tDBn2zGxtZP5XayQYnLdSoeIEQfHJoiCRFCM6ioZ2hAaOcmCBcS3sXynvMc042gCzNgRv9uR5qJ4UPMs3p/miO40jQ/bJATkmHjkjRXJFyqRCOHkkz+SVvDlPzovz7nxMWhec6cwe+SPn8wd9r5vB\nAAACDXicbZC7SgNBFIZnvcZ4i1raDEbBKuyKoFgFhGAlEcwFkhhmJ2eTIbOzy8xZMSx5ARtfxcZCEVt7O9/GyaXQxB8GPv5zDmfO78dSGHTdb2dhcWl5ZTWzll3f2Nzazu3sVk2UaA4VHslI131mQAoFFRQooR5rYKEvoeb3L0f12j1oIyJ1i4MYWiHrKhEIztBa7dxhE+EB0+uoA7QEDBMNF3RIx64fpKXhnWqLdi7vFtyx6Dx4U8iTqcrt3FezE/EkBIVcMmManhtjK2UaBZcwzDYTAzHjfdaFhkXFQjCtdHzNkB5Zp0ODSNunkI7d3xMpC40ZhL7tDBn2zGxtZP5XayQYnLdSoeIEQfHJoiCRFCM6ioZ2hAaOcmCBcS3sXynvMc042gCzNgRv9uR5qJ4UPMs3p/miO40jQ/bJATkmHjkjRXJFyqRCOHkkz+SVvDlPzovz7nxMWhec6cwe+SPn8wd9r5vB\nAAACDXicbZC7SgNBFIZnvcZ4i1raDEbBKuyKoFgFhGAlEcwFkhhmJ2eTIbOzy8xZMSx5ARtfxcZCEVt7O9/GyaXQxB8GPv5zDmfO78dSGHTdb2dhcWl5ZTWzll3f2Nzazu3sVk2UaA4VHslI131mQAoFFRQooR5rYKEvoeb3L0f12j1oIyJ1i4MYWiHrKhEIztBa7dxhE+EB0+uoA7QEDBMNF3RIx64fpKXhnWqLdi7vFtyx6Dx4U8iTqcrt3FezE/EkBIVcMmManhtjK2UaBZcwzDYTAzHjfdaFhkXFQjCtdHzNkB5Zp0ODSNunkI7d3xMpC40ZhL7tDBn2zGxtZP5XayQYnLdSoeIEQfHJoiCRFCM6ioZ2hAaOcmCBcS3sXynvMc042gCzNgRv9uR5qJ4UPMs3p/miO40jQ/bJATkmHjkjRXJFyqRCOHkkz+SVvDlPzovz7nxMWhec6cwe+SPn8wd9r5vB\n\nNode: Zn\n\ni , Zn\n\nj , Yu\n\ni , Yu\nj\n\nAAACMnicbVBNSwMxEM3Wr1q/qh69BIvgQWRXBMVTwYtepIKtH20t2XTWRrPZJZkVy7K/yYu/RPCgB0W8+iNMaxFrfRB4894MmXl+LIVB1312cmPjE5NT+enCzOzc/EJxcalmokRzqPJIRvrMZwakUFBFgRLOYg0s9CWc+jf7Pf/0FrQRkTrBbgzNkF0pEQjO0Eqt4mED4Q7To6gNezSjjZBhxw/Si+xStcTGUHn9U55nl8kvt19et4old9Ptg44Sb0BKZIBKq/jYaEc8CUEhl8yYuufG2EyZRsElZIVGYiBm/IZdQd1SxUIwzbR/ckbXrNKmQaTtU0j76u+JlIXGdEPfdva2NH+9nvifV08w2G2mQsUJguLfHwWJpBjRXn60LTRwlF1LGNfC7kp5h2nG0aZcsCF4f08eJbWtTc/y4+1S2R3EkScrZJWsE4/skDI5IBVSJZzckyfySt6cB+fFeXc+vltzzmBmmQzB+fwCQYWr7g==\nAAACMnicbVBNSwMxEM3Wr1q/qh69BIvgQWRXBMVTwYtepIKtH20t2XTWRrPZJZkVy7K/yYu/RPCgB0W8+iNMaxFrfRB4894MmXl+LIVB1312cmPjE5NT+enCzOzc/EJxcalmokRzqPJIRvrMZwakUFBFgRLOYg0s9CWc+jf7Pf/0FrQRkTrBbgzNkF0pEQjO0Eqt4mED4Q7To6gNezSjjZBhxw/Si+xStcTGUHn9U55nl8kvt19et4old9Ptg44Sb0BKZIBKq/jYaEc8CUEhl8yYuufG2EyZRsElZIVGYiBm/IZdQd1SxUIwzbR/ckbXrNKmQaTtU0j76u+JlIXGdEPfdva2NH+9nvifV08w2G2mQsUJguLfHwWJpBjRXn60LTRwlF1LGNfC7kp5h2nG0aZcsCF4f08eJbWtTc/y4+1S2R3EkScrZJWsE4/skDI5IBVSJZzckyfySt6cB+fFeXc+vltzzmBmmQzB+fwCQYWr7g==\nAAACMnicbVBNSwMxEM3Wr1q/qh69BIvgQWRXBMVTwYtepIKtH20t2XTWRrPZJZkVy7K/yYu/RPCgB0W8+iNMaxFrfRB4894MmXl+LIVB1312cmPjE5NT+enCzOzc/EJxcalmokRzqPJIRvrMZwakUFBFgRLOYg0s9CWc+jf7Pf/0FrQRkTrBbgzNkF0pEQjO0Eqt4mED4Q7To6gNezSjjZBhxw/Si+xStcTGUHn9U55nl8kvt19et4old9Ptg44Sb0BKZIBKq/jYaEc8CUEhl8yYuufG2EyZRsElZIVGYiBm/IZdQd1SxUIwzbR/ckbXrNKmQaTtU0j76u+JlIXGdEPfdva2NH+9nvifV08w2G2mQsUJguLfHwWJpBjRXn60LTRwlF1LGNfC7kp5h2nG0aZcsCF4f08eJbWtTc/y4+1S2R3EkScrZJWsE4/skDI5IBVSJZzckyfySt6cB+fFeXc+vltzzmBmmQzB+fwCQYWr7g==\nAAACMnicbVBNSwMxEM3Wr1q/qh69BIvgQWRXBMVTwYtepIKtH20t2XTWRrPZJZkVy7K/yYu/RPCgB0W8+iNMaxFrfRB4894MmXl+LIVB1312cmPjE5NT+enCzOzc/EJxcalmokRzqPJIRvrMZwakUFBFgRLOYg0s9CWc+jf7Pf/0FrQRkTrBbgzNkF0pEQjO0Eqt4mED4Q7To6gNezSjjZBhxw/Si+xStcTGUHn9U55nl8kvt19et4old9Ptg44Sb0BKZIBKq/jYaEc8CUEhl8yYuufG2EyZRsElZIVGYiBm/IZdQd1SxUIwzbR/ckbXrNKmQaTtU0j76u+JlIXGdEPfdva2NH+9nvifV08w2G2mQsUJguLfHwWJpBjRXn60LTRwlF1LGNfC7kp5h2nG0aZcsCF4f08eJbWtTc/y4+1S2R3EkScrZJWsE4/skDI5IBVSJZzckyfySt6cB+fFeXc+vltzzmBmmQzB+fwCQYWr7g==\n\nXAAACAHicbVDLSsNAFL2prxpfURcu3ASL4Kok3ehGLLhxWcE+oAllMpm0QyeTMDMRSgiIv+LGhSKuBD/DnX/jpO1CWy8MczjnXu65J0gZlcpxvo3Kyura+kZ109za3tnds/YPOjLJBCZtnLBE9AIkCaOctBVVjPRSQVAcMNINxtel3r0nQtKE36lJSvwYDTmNKEZKUwPryAsSFspJrL/ci5EaBVHeK4qBVXPqzrTsZeDOQe3qw7x8AIDWwPrywgRnMeEKMyRl33VS5edIKIoZKUwvkyRFeIyGpK8hRzGRfj49oLBPNRPaUSL048qesr8nchTL0qPuLC3KRa0k/9P6mYou/JzyNFOE49miKGO2SuwyDTukgmDFJhogLKj2auMREggrnZmpQ3AXT14GnUbdderurVtrNmBWVTiGEzgDF86hCTfQgjZgKOAJXuDVeDSejTfjfdZaMeYzh/CnjM8fOD6Y5A==\n\nAAACAHicbVDLSsNAFJ34rPEVdeHCzWARXJWkG92IBTcuK9gHNKFMJpN26GQmzEyEErLxV7pxoYgrwc9wI/6Nk7YLbb0wzOGce7nnnjBlVGnX/bZWVtfWNzYrW/b2zu7evnNw2FYik5i0sGBCdkOkCKOctDTVjHRTSVASMtIJRzel3nkgUlHB7/U4JUGCBpzGFCNtqL5z7IeCRWqcmC/3E6SHYZx3i6LvVN2aOy24DLw5qF6/21fp5Mtu9p1PPxI4SwjXmCGlep6b6iBHUlPMSGH7mSIpwiM0ID0DOUqICvLpAQU8M0wEYyHN4xpO2d8TOUpU6dF0lhbVolaS/2m9TMeXQU55mmnC8WxRnDGoBSzTgBGVBGs2NgBhSY1XiIdIIqxNZrYJwVs8eRm06zXPrXl3XrVRB7OqgBNwCs6BBy5AA9yCJmgBDAowAc/gxXq0nqxX623WumLNZ47An7I+fgApzZpY\nAAACAHicbVDLSsNAFJ34rPEVdeHCzWARXJWkG92IBTcuK9gHNKFMJpN26GQmzEyEErLxV7pxoYgrwc9wI/6Nk7YLbb0wzOGce7nnnjBlVGnX/bZWVtfWNzYrW/b2zu7evnNw2FYik5i0sGBCdkOkCKOctDTVjHRTSVASMtIJRzel3nkgUlHB7/U4JUGCBpzGFCNtqL5z7IeCRWqcmC/3E6SHYZx3i6LvVN2aOy24DLw5qF6/21fp5Mtu9p1PPxI4SwjXmCGlep6b6iBHUlPMSGH7mSIpwiM0ID0DOUqICvLpAQU8M0wEYyHN4xpO2d8TOUpU6dF0lhbVolaS/2m9TMeXQU55mmnC8WxRnDGoBSzTgBGVBGs2NgBhSY1XiIdIIqxNZrYJwVs8eRm06zXPrXl3XrVRB7OqgBNwCs6BBy5AA9yCJmgBDAowAc/gxXq0nqxX623WumLNZ47An7I+fgApzZpY\nAAAB2XicbZDNSgMxFIXv1L86Vq1rN8EiuCozbnQpuHFZwbZCO5RM5k4bmskMyR2hDH0BF25EfC93vo3pz0JbDwQ+zknIvSculLQUBN9ebWd3b/+gfugfNfzjk9Nmo2fz0gjsilzl5jnmFpXU2CVJCp8LgzyLFfbj6f0i77+gsTLXTzQrMMr4WMtUCk7O6oyaraAdLMW2IVxDC9YaNb+GSS7KDDUJxa0dhEFBUcUNSaFw7g9LiwUXUz7GgUPNM7RRtRxzzi6dk7A0N+5oYkv394uKZ9bOstjdzDhN7Ga2MP/LBiWlt1EldVESarH6KC0Vo5wtdmaJNChIzRxwYaSblYkJN1yQa8Z3HYSbG29D77odBu3wMYA6nMMFXEEIN3AHD9CBLghI4BXevYn35n2suqp569LO4I+8zx84xIo4\nAAAB9XicbVC9TsMwGPxS/kopEFgYWCwqJKYqYaEjEgtjkeiP1ESV4zitVceObAepirLwKiwMIMSrsPE2OG0HaDnJ8unOn3zfRRln2njet1Pb2t7Z3avvNw6ah0fH7kmzr2WuCO0RyaUaRlhTzgTtGWY4HWaK4jTidBDN7ip/8ESVZlI8mnlGwxRPBEsYwcZKY/csiCSP9Ty1VxGk2EyjpBiW5dhteW1vAbRJ/BVpwQrdsfsVxJLkKRWGcKz1yPcyExZYGUY4LRtBrmmGyQxP6MhSgVOqw2KxQIkurRKjRCp7hEEL9fdEgVNdZbQvq4h63avE/7xRbpJOWDCR5YYKsvwoyTkyElVtoJgpSgyfW4KJYjYrIlOsMDG2s4YtwV9feZP0r9u+1/YfPKjDOVzAFfhwA7dwD13oAYESXuAN3p1n59X5WNZVc1a9ncIfOJ8/Y2eVzw==\nAAAB9XicbVC9TsMwGPxS/kopEFgYWCwqJKYqYaEjEgtjkeiP1ESV4zitVceObAepirLwKiwMIMSrsPE2OG0HaDnJ8unOn3zfRRln2njet1Pb2t7Z3avvNw6ah0fH7kmzr2WuCO0RyaUaRlhTzgTtGWY4HWaK4jTidBDN7ip/8ESVZlI8mnlGwxRPBEsYwcZKY/csiCSP9Ty1VxGk2EyjpBiW5dhteW1vAbRJ/BVpwQrdsfsVxJLkKRWGcKz1yPcyExZYGUY4LRtBrmmGyQxP6MhSgVOqw2KxQIkurRKjRCp7hEEL9fdEgVNdZbQvq4h63avE/7xRbpJOWDCR5YYKsvwoyTkyElVtoJgpSgyfW4KJYjYrIlOsMDG2s4YtwV9feZP0r9u+1/YfPKjDOVzAFfhwA7dwD13oAYESXuAN3p1n59X5WNZVc1a9ncIfOJ8/Y2eVzw==\nAAACAHicbVC7TsMwFL0pr1JeAQYGlogKialKusBYiYWxSPQhNVHlOE5r1bEj20Gqoiz8CgsDCLHyGWz8DU6bAVquZPnonHt1zz1hyqjSrvtt1TY2t7Z36ruNvf2DwyP7+KSvRCYx6WHBhByGSBFGOelpqhkZppKgJGRkEM5uS33wSKSigj/oeUqCBE04jSlG2lBj+8wPBYvUPDFf7idIT8M4HxbF2G66LXdRzjrwKtCEqrpj+8uPBM4SwjVmSKmR56Y6yJHUFDNSNPxMkRThGZqQkYEcJUQF+eKAwrk0TOTEQprHtbNgf0/kKFGlR9NZWlSrWkn+p40yHd8EOeVppgnHy0VxxhwtnDINJ6KSYM3mBiAsqfHq4CmSCGuTWcOE4K2evA767Zbntrx7t9lpV3HU4Rwu4Ao8uIYO3EEXeoChgGd4hTfryXqx3q2PZWvNqmZO4U9Znz/eV5cm\nAAACAHicbVC7TsMwFL0pr1JeAQYGlogKialKusBYiYWxSPQhNVHlOE5r1bEj20Gqoiz8CgsDCLHyGWz8DU6bAVquZPnonHt1zz1hyqjSrvtt1TY2t7Z36ruNvf2DwyP7+KSvRCYx6WHBhByGSBFGOelpqhkZppKgJGRkEM5uS33wSKSigj/oeUqCBE04jSlG2lBj+8wPBYvUPDFf7idIT8M4HxbF2G66LXdRzjrwKtCEqrpj+8uPBM4SwjVmSKmR56Y6yJHUFDNSNPxMkRThGZqQkYEcJUQF+eKAwrk0TOTEQprHtbNgf0/kKFGlR9NZWlSrWkn+p40yHd8EOeVppgnHy0VxxhwtnDINJ6KSYM3mBiAsqfHq4CmSCGuTWcOE4K2evA767Zbntrx7r9lpV3HU4Rwu4Ao8uIYO3EEXeoChgGd4hTfryXqx3q2PZWvNqmZO4U9Znz/ep5cn\nAAACAHicbVDLSsNAFJ34rPEVdeHCzWARXJWkG92IBTcuK9gHNKFMJpN26GQmzEyEErLxV7pxoYgrwc9wI/6Nk7YLbb0wzOGce7nnnjBlVGnX/bZWVtfWNzYrW/b2zu7evnNw2FYik5i0sGBCdkOkCKOctDTVjHRTSVASMtIJRzel3nkgUlHB7/U4JUGCBpzGFCNtqL5z7IeCRWqcmC/3E6SHYZx3i6LvVN2aOy24DLw5qF6/21fp5Mtu9p1PPxI4SwjXmCGlep6b6iBHUlPMSGH7mSIpwiM0ID0DOUqICvLpAQU8M0wEYyHN4xpO2d8TOUpU6dF0lhbVolaS/2m9TMeXQU55mmnC8WxRnDGoBSzTgBGVBGs2NgBhSY1XiIdIIqxNZrYJwVs8eRm06zXPrXl3XrVRB7OqgBNwCs6BBy5AA9yCJmgBDAowAc/gxXq0nqxX623WumLNZ47An7I+fgApzZpY\nAAACAHicbVDLSsNAFJ34rPEVdeHCzWARXJWkG92IBTcuK9gHNKFMJpN26GQmzEyEErLxV7pxoYgrwc9wI/6Nk7YLbb0wzOGce7nnnjBlVGnX/bZWVtfWNzYrW/b2zu7evnNw2FYik5i0sGBCdkOkCKOctDTVjHRTSVASMtIJRzel3nkgUlHB7/U4JUGCBpzGFCNtqL5z7IeCRWqcmC/3E6SHYZx3i6LvVN2aOy24DLw5qF6/21fp5Mtu9p1PPxI4SwjXmCGlep6b6iBHUlPMSGH7mSIpwiM0ID0DOUqICvLpAQU8M0wEYyHN4xpO2d8TOUpU6dF0lhbVolaS/2m9TMeXQU55mmnC8WxRnDGoBSzTgBGVBGs2NgBhSY1XiIdIIqxNZrYJwVs8eRm06zXPrXl3XrVRB7OqgBNwCs6BBy5AA9yCJmgBDAowAc/gxXq0nqxX623WumLNZ47An7I+fgApzZpY\nAAACAHicbVDLSsNAFJ34rPEVdeHCzWARXJWkG92IBTcuK9gHNKFMJpN26GQmzEyEErLxV7pxoYgrwc9wI/6Nk7YLbb0wzOGce7nnnjBlVGnX/bZWVtfWNzYrW/b2zu7evnNw2FYik5i0sGBCdkOkCKOctDTVjHRTSVASMtIJRzel3nkgUlHB7/U4JUGCBpzGFCNtqL5z7IeCRWqcmC/3E6SHYZx3i6LvVN2aOy24DLw5qF6/21fp5Mtu9p1PPxI4SwjXmCGlep6b6iBHUlPMSGH7mSIpwiM0ID0DOUqICvLpAQU8M0wEYyHN4xpO2d8TOUpU6dF0lhbVolaS/2m9TMeXQU55mmnC8WxRnDGoBSzTgBGVBGs2NgBhSY1XiIdIIqxNZrYJwVs8eRm06zXPrXl3XrVRB7OqgBNwCs6BBy5AA9yCJmgBDAowAc/gxXq0nqxX623WumLNZ47An7I+fgApzZpY\nAAACAHicbVDLSsNAFJ34rPEVdeHCzWARXJWkG92IBTcuK9gHNKFMJpN26GQmzEyEErLxV7pxoYgrwc9wI/6Nk7YLbb0wzOGce7nnnjBlVGnX/bZWVtfWNzYrW/b2zu7evnNw2FYik5i0sGBCdkOkCKOctDTVjHRTSVASMtIJRzel3nkgUlHB7/U4JUGCBpzGFCNtqL5z7IeCRWqcmC/3E6SHYZx3i6LvVN2aOy24DLw5qF6/21fp5Mtu9p1PPxI4SwjXmCGlep6b6iBHUlPMSGH7mSIpwiM0ID0DOUqICvLpAQU8M0wEYyHN4xpO2d8TOUpU6dF0lhbVolaS/2m9TMeXQU55mmnC8WxRnDGoBSzTgBGVBGs2NgBhSY1XiIdIIqxNZrYJwVs8eRm06zXPrXl3XrVRB7OqgBNwCs6BBy5AA9yCJmgBDAowAc/gxXq0nqxX623WumLNZ47An7I+fgApzZpY\nAAACAHicbVC7TsMwFL0pr1JeAQYGlogKialKusBYiYWxSPQhNVHlOE5r1bEj20Gqoiz8CgsDCLHyGWz8DU6bAVquZPnonHt1zz1hyqjSrvtt1TY2t7Z36ruNvf2DwyP7+KSvRCYx6WHBhByGSBFGOelpqhkZppKgJGRkEM5uS33wSKSigj/oeUqCBE04jSlG2lBj+8wPBYvUPDFf7idIT8M4HxbF2G66LXdRzjrwKtCEqrpj+8uPBM4SwjVmSKmR56Y6yJHUFDNSNPxMkRThGZqQkYEcJUQF+eKAwrk0TOTEQprHtbNgf0/kKFGlR9NZWlSrWkn+p40yHd8EOeVppgnHy0VxxhwtnDINJ6KSYM3mBiAsqfHq4CmSCGuTWcOE4K2evA767Zbntrx7r9lpV3HU4Rwu4Ao8uIYO3EEXeoChgGd4hTfryXqx3q2PZWvNqmZO4U9Znz/ep5cn\n\nInfered Label: Yu\ni\n\nAAACDnicbZC7SgNBFIZnvcZ4i1raDIaAVdgVQbEK2ChYRDAXSWKYnZxNhszOLjNnxbDkCWx8FRsLRWyt7XwbJ5dCE38Y+PjPOZw5vx9LYdB1v52FxaXlldXMWnZ9Y3NrO7ezWzVRojlUeCQjXfeZASkUVFCghHqsgYW+hJrfPx/Va/egjYjUDQ5iaIWsq0QgOENrtXOFJsIDppcqAA0desV8kGd0SMe2H6S3w7ukLdq5vFt0x6Lz4E0hT6Yqt3NfzU7EkxAUcsmMaXhujK2UaRRcwjDbTAzEjPdZFxoWFQvBtNLxOUNasE6HBpG2TyEdu78nUhYaMwh92xky7JnZ2sj8r9ZIMDhtpULFCYLik0VBIilGdJQN7QgNHOXAAuNa2L9S3mOacbQJZm0I3uzJ81A9KnqWr4/zJXcaR4bskwNySDxyQkrkgpRJhXDySJ7JK3lznpwX5935mLQuONOZPfJHzucPScucMA==\nAAACDnicbZC7SgNBFIZnvcZ4i1raDIaAVdgVQbEK2ChYRDAXSWKYnZxNhszOLjNnxbDkCWx8FRsLRWyt7XwbJ5dCE38Y+PjPOZw5vx9LYdB1v52FxaXlldXMWnZ9Y3NrO7ezWzVRojlUeCQjXfeZASkUVFCghHqsgYW+hJrfPx/Va/egjYjUDQ5iaIWsq0QgOENrtXOFJsIDppcqAA0desV8kGd0SMe2H6S3w7ukLdq5vFt0x6Lz4E0hT6Yqt3NfzU7EkxAUcsmMaXhujK2UaRRcwjDbTAzEjPdZFxoWFQvBtNLxOUNasE6HBpG2TyEdu78nUhYaMwh92xky7JnZ2sj8r9ZIMDhtpULFCYLik0VBIilGdJQN7QgNHOXAAuNa2L9S3mOacbQJZm0I3uzJ81A9KnqWr4/zJXcaR4bskwNySDxyQkrkgpRJhXDySJ7JK3lznpwX5935mLQuONOZPfJHzucPScucMA==\nAAACDnicbZC7SgNBFIZnvcZ4i1raDIaAVdgVQbEK2ChYRDAXSWKYnZxNhszOLjNnxbDkCWx8FRsLRWyt7XwbJ5dCE38Y+PjPOZw5vx9LYdB1v52FxaXlldXMWnZ9Y3NrO7ezWzVRojlUeCQjXfeZASkUVFCghHqsgYW+hJrfPx/Va/egjYjUDQ5iaIWsq0QgOENrtXOFJsIDppcqAA0desV8kGd0SMe2H6S3w7ukLdq5vFt0x6Lz4E0hT6Yqt3NfzU7EkxAUcsmMaXhujK2UaRRcwjDbTAzEjPdZFxoWFQvBtNLxOUNasE6HBpG2TyEdu78nUhYaMwh92xky7JnZ2sj8r9ZIMDhtpULFCYLik0VBIilGdJQN7QgNHOXAAuNa2L9S3mOacbQJZm0I3uzJ81A9KnqWr4/zJXcaR4bskwNySDxyQkrkgpRJhXDySJ7JK3lznpwX5935mLQuONOZPfJHzucPScucMA==\nAAACDnicbZC7SgNBFIZnvcZ4i1raDIaAVdgVQbEK2ChYRDAXSWKYnZxNhszOLjNnxbDkCWx8FRsLRWyt7XwbJ5dCE38Y+PjPOZw5vx9LYdB1v52FxaXlldXMWnZ9Y3NrO7ezWzVRojlUeCQjXfeZASkUVFCghHqsgYW+hJrfPx/Va/egjYjUDQ5iaIWsq0QgOENrtXOFJsIDppcqAA0desV8kGd0SMe2H6S3w7ukLdq5vFt0x6Lz4E0hT6Yqt3NfzU7EkxAUcsmMaXhujK2UaRRcwjDbTAzEjPdZFxoWFQvBtNLxOUNasE6HBpG2TyEdu78nUhYaMwh92xky7JnZ2sj8r9ZIMDhtpULFCYLik0VBIilGdJQN7QgNHOXAAuNa2L9S3mOacbQJZm0I3uzJ81A9KnqWr4/zJXcaR4bskwNySDxyQkrkgpRJhXDySJ7JK3lznpwX5935mLQuONOZPfJHzucPScucMA==\n\nAttribute Feature: Fa\na\n\nAAACEnicbVDLSgNBEJz1GeMr6tHLYBD0IrsiKJ4ignhUMFFI1tA76TVDZh/M9Iph2W/w4q948aCIV0/e/Bsnj4OvgoGiqpueqiBV0pDrfjoTk1PTM7OlufL8wuLScmVltWGSTAusi0Ql+ioAg0rGWCdJCq9SjRAFCi+D3vHAv7xFbWQSX1A/RT+Cm1iGUgBZqV3ZbhHeUX5EpGWQEfITBMo0HvKCD60gzE+Ka2hDu1J1d9wh+F/ijUmVjXHWrny0OonIIoxJKDCm6bkp+TlokkJhUW5lBlMQPbjBpqUxRGj8fBip4JtW6fAw0fbFxIfq940cImP6UWAnI6Cu+e0NxP+8ZkbhgZ/LOLVhYzE6FGaKU8IH/fCO1ChI9S0BoaX9Kxdd0CDItli2JXi/I/8ljd0dz/LzvWrNHddRYutsg20xj+2zGjtlZ6zOBLtnj+yZvTgPzpPz6ryNRiec8c4a+wHn/Qu4EJ4M\nAAACEnicbVDLSgNBEJz1GeMr6tHLYBD0IrsiKJ4ignhUMFFI1tA76TVDZh/M9Iph2W/w4q948aCIV0/e/Bsnj4OvgoGiqpueqiBV0pDrfjoTk1PTM7OlufL8wuLScmVltWGSTAusi0Ql+ioAg0rGWCdJCq9SjRAFCi+D3vHAv7xFbWQSX1A/RT+Cm1iGUgBZqV3ZbhHeUX5EpGWQEfITBMo0HvKCD60gzE+Ka2hDu1J1d9wh+F/ijUmVjXHWrny0OonIIoxJKDCm6bkp+TlokkJhUW5lBlMQPbjBpqUxRGj8fBip4JtW6fAw0fbFxIfq940cImP6UWAnI6Cu+e0NxP+8ZkbhgZ/LOLVhYzE6FGaKU8IH/fCO1ChI9S0BoaX9Kxdd0CDItli2JXi/I/8ljd0dz/LzvWrNHddRYutsg20xj+2zGjtlZ6zOBLtnj+yZvTgPzpPz6ryNRiec8c4a+wHn/Qu4EJ4M\nAAACEnicbVDLSgNBEJz1GeMr6tHLYBD0IrsiKJ4ignhUMFFI1tA76TVDZh/M9Iph2W/w4q948aCIV0/e/Bsnj4OvgoGiqpueqiBV0pDrfjoTk1PTM7OlufL8wuLScmVltWGSTAusi0Ql+ioAg0rGWCdJCq9SjRAFCi+D3vHAv7xFbWQSX1A/RT+Cm1iGUgBZqV3ZbhHeUX5EpGWQEfITBMo0HvKCD60gzE+Ka2hDu1J1d9wh+F/ijUmVjXHWrny0OonIIoxJKDCm6bkp+TlokkJhUW5lBlMQPbjBpqUxRGj8fBip4JtW6fAw0fbFxIfq940cImP6UWAnI6Cu+e0NxP+8ZkbhgZ/LOLVhYzE6FGaKU8IH/fCO1ChI9S0BoaX9Kxdd0CDItli2JXi/I/8ljd0dz/LzvWrNHddRYutsg20xj+2zGjtlZ6zOBLtnj+yZvTgPzpPz6ryNRiec8c4a+wHn/Qu4EJ4M\nAAACEnicbVDLSgNBEJz1GeMr6tHLYBD0IrsiKJ4ignhUMFFI1tA76TVDZh/M9Iph2W/w4q948aCIV0/e/Bsnj4OvgoGiqpueqiBV0pDrfjoTk1PTM7OlufL8wuLScmVltWGSTAusi0Ql+ioAg0rGWCdJCq9SjRAFCi+D3vHAv7xFbWQSX1A/RT+Cm1iGUgBZqV3ZbhHeUX5EpGWQEfITBMo0HvKCD60gzE+Ka2hDu1J1d9wh+F/ijUmVjXHWrny0OonIIoxJKDCm6bkp+TlokkJhUW5lBlMQPbjBpqUxRGj8fBip4JtW6fAw0fbFxIfq940cImP6UWAnI6Cu+e0NxP+8ZkbhgZ/LOLVhYzE6FGaKU8IH/fCO1ChI9S0BoaX9Kxdd0CDItli2JXi/I/8ljd0dz/LzvWrNHddRYutsg20xj+2zGjtlZ6zOBLtnj+yZvTgPzpPz6ryNRiec8c4a+wHn/Qu4EJ4M\n\nAttribute: Za\na\n\nAAACCnicbVDLSsNAFJ3UV62vqEs3o0VwVRIRFFcVNy4r2Ae2MUymk3bo5MHMjVhC1m78FTcuFHHrF7jzb5ymWWjrgYHDOfdy5xwvFlyBZX0bpYXFpeWV8mplbX1jc8vc3mmpKJGUNWkkItnxiGKCh6wJHATrxJKRwBOs7Y0uJ377nknFo/AGxjFzAjIIuc8pAS255n4P2AOkFwCSewmwc5zhXPL89Da7Iy5xzapVs3LgeWIXpIoKNFzzq9ePaBKwEKggSnVtKwYnJRI4FSyr9BLFYkJHZMC6moYkYMpJ8ygZPtRKH/uR1C8EnKu/N1ISKDUOPD0ZEBiqWW8i/ud1E/DPnJSHsQ4Z0ukhPxEYIjzpBfe5ZBTEWBNCJdd/xXRIJKGg26voEuzZyPOkdVyzNb8+qdatoo4y2kMH6AjZ6BTV0RVqoCai6BE9o1f0ZjwZL8a78TEdLRnFzi76A+PzB86ZmuQ=\nAAACCnicbVDLSsNAFJ3UV62vqEs3o0VwVRIRFFcVNy4r2Ae2MUymk3bo5MHMjVhC1m78FTcuFHHrF7jzb5ymWWjrgYHDOfdy5xwvFlyBZX0bpYXFpeWV8mplbX1jc8vc3mmpKJGUNWkkItnxiGKCh6wJHATrxJKRwBOs7Y0uJ377nknFo/AGxjFzAjIIuc8pAS255n4P2AOkFwCSewmwc5zhXPL89Da7Iy5xzapVs3LgeWIXpIoKNFzzq9ePaBKwEKggSnVtKwYnJRI4FSyr9BLFYkJHZMC6moYkYMpJ8ygZPtRKH/uR1C8EnKu/N1ISKDUOPD0ZEBiqWW8i/ud1E/DPnJSHsQ4Z0ukhPxEYIjzpBfe5ZBTEWBNCJdd/xXRIJKGg26voEuzZyPOkdVyzNb8+qdatoo4y2kMH6AjZ6BTV0RVqoCai6BE9o1f0ZjwZL8a78TEdLRnFzi76A+PzB86ZmuQ=\nAAACCnicbVDLSsNAFJ3UV62vqEs3o0VwVRIRFFcVNy4r2Ae2MUymk3bo5MHMjVhC1m78FTcuFHHrF7jzb5ymWWjrgYHDOfdy5xwvFlyBZX0bpYXFpeWV8mplbX1jc8vc3mmpKJGUNWkkItnxiGKCh6wJHATrxJKRwBOs7Y0uJ377nknFo/AGxjFzAjIIuc8pAS255n4P2AOkFwCSewmwc5zhXPL89Da7Iy5xzapVs3LgeWIXpIoKNFzzq9ePaBKwEKggSnVtKwYnJRI4FSyr9BLFYkJHZMC6moYkYMpJ8ygZPtRKH/uR1C8EnKu/N1ISKDUOPD0ZEBiqWW8i/ud1E/DPnJSHsQ4Z0ukhPxEYIjzpBfe5ZBTEWBNCJdd/xXRIJKGg26voEuzZyPOkdVyzNb8+qdatoo4y2kMH6AjZ6BTV0RVqoCai6BE9o1f0ZjwZL8a78TEdLRnFzi76A+PzB86ZmuQ=\nAAACCnicbVDLSsNAFJ3UV62vqEs3o0VwVRIRFFcVNy4r2Ae2MUymk3bo5MHMjVhC1m78FTcuFHHrF7jzb5ymWWjrgYHDOfdy5xwvFlyBZX0bpYXFpeWV8mplbX1jc8vc3mmpKJGUNWkkItnxiGKCh6wJHATrxJKRwBOs7Y0uJ377nknFo/AGxjFzAjIIuc8pAS255n4P2AOkFwCSewmwc5zhXPL89Da7Iy5xzapVs3LgeWIXpIoKNFzzq9ePaBKwEKggSnVtKwYnJRI4FSyr9BLFYkJHZMC6moYkYMpJ8ygZPtRKH/uR1C8EnKu/N1ISKDUOPD0ZEBiqWW8i/ud1E/DPnJSHsQ4Z0ukhPxEYIjzpBfe5ZBTEWBNCJdd/xXRIJKGg26voEuzZyPOkdVyzNb8+qdatoo4y2kMH6AjZ6BTV0RVqoCai6BE9o1f0ZjwZL8a78TEdLRnFzi76A+PzB86ZmuQ=\n\nCase 1: Oll\n\nij\n\nAAACDXicbVDLSsNAFJ34rPUVdelmsAquSiKC4qrQjTsr2Ae0MUymk3bs5MHMjVhCfsCNv+LGhSJu3bvzb5y0WWjrgYHDOffeufd4seAKLOvbWFhcWl5ZLa2V1zc2t7bNnd2WihJJWZNGIpIdjygmeMiawEGwTiwZCTzB2t6onvvteyYVj8IbGMfMCcgg5D6nBLTkmoc9YA+Q1vUIbF/grBcQGFIi0qvsNhUic1N+l7lmxapaE+B5Yhekggo0XPOr149oErAQqCBKdW0rBiclEjgVLCv3EsViQkdkwLqahiRgykkn12T4SCt97EdSvxDwRP3dkZJAqXHg6cp8WTXr5eJ/XjcB/9xJeRgnwEI6/chPBIYI59HgPpeMghhrQqjkeldMh0QSCjrAsg7Bnj15nrROqrbm16eVmlXEUUL76AAdIxudoRq6RA3URBQ9omf0it6MJ+PFeDc+pqULRtGzh/7A+PwB+5OcFg==\nAAACDXicbVDLSsNAFJ34rPUVdelmsAquSiKC4qrQjTsr2Ae0MUymk3bs5MHMjVhCfsCNv+LGhSJu3bvzb5y0WWjrgYHDOffeufd4seAKLOvbWFhcWl5ZLa2V1zc2t7bNnd2WihJJWZNGIpIdjygmeMiawEGwTiwZCTzB2t6onvvteyYVj8IbGMfMCcgg5D6nBLTkmoc9YA+Q1vUIbF/grBcQGFIi0qvsNhUic1N+l7lmxapaE+B5Yhekggo0XPOr149oErAQqCBKdW0rBiclEjgVLCv3EsViQkdkwLqahiRgykkn12T4SCt97EdSvxDwRP3dkZJAqXHg6cp8WTXr5eJ/XjcB/9xJeRgnwEI6/chPBIYI59HgPpeMghhrQqjkeldMh0QSCjrAsg7Bnj15nrROqrbm16eVmlXEUUL76AAdIxudoRq6RA3URBQ9omf0it6MJ+PFeDc+pqULRtGzh/7A+PwB+5OcFg==\nAAACDXicbVDLSsNAFJ34rPUVdelmsAquSiKC4qrQjTsr2Ae0MUymk3bs5MHMjVhCfsCNv+LGhSJu3bvzb5y0WWjrgYHDOffeufd4seAKLOvbWFhcWl5ZLa2V1zc2t7bNnd2WihJJWZNGIpIdjygmeMiawEGwTiwZCTzB2t6onvvteyYVj8IbGMfMCcgg5D6nBLTkmoc9YA+Q1vUIbF/grBcQGFIi0qvsNhUic1N+l7lmxapaE+B5Yhekggo0XPOr149oErAQqCBKdW0rBiclEjgVLCv3EsViQkdkwLqahiRgykkn12T4SCt97EdSvxDwRP3dkZJAqXHg6cp8WTXr5eJ/XjcB/9xJeRgnwEI6/chPBIYI59HgPpeMghhrQqjkeldMh0QSCjrAsg7Bnj15nrROqrbm16eVmlXEUUL76AAdIxudoRq6RA3URBQ9omf0it6MJ+PFeDc+pqULRtGzh/7A+PwB+5OcFg==\nAAACDXicbVDLSsNAFJ34rPUVdelmsAquSiKC4qrQjTsr2Ae0MUymk3bs5MHMjVhCfsCNv+LGhSJu3bvzb5y0WWjrgYHDOffeufd4seAKLOvbWFhcWl5ZLa2V1zc2t7bNnd2WihJJWZNGIpIdjygmeMiawEGwTiwZCTzB2t6onvvteyYVj8IbGMfMCcgg5D6nBLTkmoc9YA+Q1vUIbF/grBcQGFIi0qvsNhUic1N+l7lmxapaE+B5Yhekggo0XPOr149oErAQqCBKdW0rBiclEjgVLCv3EsViQkdkwLqahiRgykkn12T4SCt97EdSvxDwRP3dkZJAqXHg6cp8WTXr5eJ/XjcB/9xJeRgnwEI6/chPBIYI59HgPpeMghhrQqjkeldMh0QSCjrAsg7Bnj15nrROqrbm16eVmlXEUUL76AAdIxudoRq6RA3URBQ9omf0it6MJ+PFeDc+pqULRtGzh/7A+PwB+5OcFg==\n\nCase 2: Ouu\n\nij\n\nAAACDXicbVC7SgNBFJ2NrxhfUUubwShYhd0gKFaBNHZGMA/IxjA7mU3GzD6YuSuGYX/Axl+xsVDE1t7Ov3E2SaGJBwYO59x7597jxYIrsO1vK7e0vLK6ll8vbGxube8Ud/eaKkokZQ0aiUi2PaKY4CFrAAfB2rFkJPAEa3mjWua37plUPApvYByzbkAGIfc5JWCkXvHIBfYAumZG4MoFTt2AwJASoa/SW50kaU/zu7RXLNllewK8SJwZKaEZ6r3il9uPaBKwEKggSnUcO4auJhI4FSwtuIliMaEjMmAdQ0MSMNXVk2tSfGyUPvYjaV4IeKL+7tAkUGoceKYyW1bNe5n4n9dJwD/vah7GCbCQTj/yE4Ehwlk0uM8loyDGhhAqudkV0yGRhIIJsGBCcOZPXiTNStkx/Pq0VLVnceTRATpEJ8hBZ6iKLlEdNRBFj+gZvaI368l6sd6tj2lpzpr17KM/sD5/ABj7nCk=\nAAACDXicbVC7SgNBFJ2NrxhfUUubwShYhd0gKFaBNHZGMA/IxjA7mU3GzD6YuSuGYX/Axl+xsVDE1t7Ov3E2SaGJBwYO59x7597jxYIrsO1vK7e0vLK6ll8vbGxube8Ud/eaKkokZQ0aiUi2PaKY4CFrAAfB2rFkJPAEa3mjWua37plUPApvYByzbkAGIfc5JWCkXvHIBfYAumZG4MoFTt2AwJASoa/SW50kaU/zu7RXLNllewK8SJwZKaEZ6r3il9uPaBKwEKggSnUcO4auJhI4FSwtuIliMaEjMmAdQ0MSMNXVk2tSfGyUPvYjaV4IeKL+7tAkUGoceKYyW1bNe5n4n9dJwD/vah7GCbCQTj/yE4Ehwlk0uM8loyDGhhAqudkV0yGRhIIJsGBCcOZPXiTNStkx/Pq0VLVnceTRATpEJ8hBZ6iKLlEdNRBFj+gZvaI368l6sd6tj2lpzpr17KM/sD5/ABj7nCk=\nAAACDXicbVC7SgNBFJ2NrxhfUUubwShYhd0gKFaBNHZGMA/IxjA7mU3GzD6YuSuGYX/Axl+xsVDE1t7Ov3E2SaGJBwYO59x7597jxYIrsO1vK7e0vLK6ll8vbGxube8Ud/eaKkokZQ0aiUi2PaKY4CFrAAfB2rFkJPAEa3mjWua37plUPApvYByzbkAGIfc5JWCkXvHIBfYAumZG4MoFTt2AwJASoa/SW50kaU/zu7RXLNllewK8SJwZKaEZ6r3il9uPaBKwEKggSnUcO4auJhI4FSwtuIliMaEjMmAdQ0MSMNXVk2tSfGyUPvYjaV4IeKL+7tAkUGoceKYyW1bNe5n4n9dJwD/vah7GCbCQTj/yE4Ehwlk0uM8loyDGhhAqudkV0yGRhIIJsGBCcOZPXiTNStkx/Pq0VLVnceTRATpEJ8hBZ6iKLlEdNRBFj+gZvaI368l6sd6tj2lpzpr17KM/sD5/ABj7nCk=\nAAACDXicbVC7SgNBFJ2NrxhfUUubwShYhd0gKFaBNHZGMA/IxjA7mU3GzD6YuSuGYX/Axl+xsVDE1t7Ov3E2SaGJBwYO59x7597jxYIrsO1vK7e0vLK6ll8vbGxube8Ud/eaKkokZQ0aiUi2PaKY4CFrAAfB2rFkJPAEa3mjWua37plUPApvYByzbkAGIfc5JWCkXvHIBfYAumZG4MoFTt2AwJASoa/SW50kaU/zu7RXLNllewK8SJwZKaEZ6r3il9uPaBKwEKggSnUcO4auJhI4FSwtuIliMaEjMmAdQ0MSMNXVk2tSfGyUPvYjaV4IeKL+7tAkUGoceKYyW1bNe5n4n9dJwD/vah7GCbCQTj/yE4Ehwlk0uM8loyDGhhAqudkV0yGRhIIJsGBCcOZPXiTNStkx/Pq0VLVnceTRATpEJ8hBZ6iKLlEdNRBFj+gZvaI368l6sd6tj2lpzpr17KM/sD5/ABj7nCk=\n\nCase 3: Olu\n\nij\n\nAAACDXicbVC7SgNBFJ2Nrxhfq5Y2g1GwCrsqKFaBNHZGMA9I1mV2MknGzD6YuSuGYX/Axl+xsVDE1t7Ov3HyKDTxwMDhnHvv3HuCRHAFjvNt5RYWl5ZX8quFtfWNzS17e6eu4lRSVqOxiGUzIIoJHrEacBCsmUhGwkCwRjCojPzGPZOKx9ENDBPmhaQX8S6nBIzk2wdtYA+gK2YEPrnAWTsk0KdE6KvsVos08zW/y3y76JScMfA8caekiKao+vZXuxPTNGQRUEGUarlOAp4mEjgVLCu0U8USQgekx1qGRiRkytPjazJ8aJQO7sbSvAjwWP3doUmo1DAMTOVoWTXrjcT/vFYK3XNP8yhJgUV08lE3FRhiPIoGd7hkFMTQEEIlN7ti2ieSUDABFkwI7uzJ86R+XHINvz4tlp1pHHm0h/bREXLRGSqjS1RFNUTRI3pGr+jNerJerHfrY1Kas6Y9u+gPrM8fDLScIQ==\nAAACDXicbVC7SgNBFJ2Nrxhfq5Y2g1GwCrsqKFaBNHZGMA9I1mV2MknGzD6YuSuGYX/Axl+xsVDE1t7Ov3HyKDTxwMDhnHvv3HuCRHAFjvNt5RYWl5ZX8quFtfWNzS17e6eu4lRSVqOxiGUzIIoJHrEacBCsmUhGwkCwRjCojPzGPZOKx9ENDBPmhaQX8S6nBIzk2wdtYA+gK2YEPrnAWTsk0KdE6KvsVos08zW/y3y76JScMfA8caekiKao+vZXuxPTNGQRUEGUarlOAp4mEjgVLCu0U8USQgekx1qGRiRkytPjazJ8aJQO7sbSvAjwWP3doUmo1DAMTOVoWTXrjcT/vFYK3XNP8yhJgUV08lE3FRhiPIoGd7hkFMTQEEIlN7ti2ieSUDABFkwI7uzJ86R+XHINvz4tlp1pHHm0h/bREXLRGSqjS1RFNUTRI3pGr+jNerJerHfrY1Kas6Y9u+gPrM8fDLScIQ==\nAAACDXicbVC7SgNBFJ2Nrxhfq5Y2g1GwCrsqKFaBNHZGMA9I1mV2MknGzD6YuSuGYX/Axl+xsVDE1t7Ov3HyKDTxwMDhnHvv3HuCRHAFjvNt5RYWl5ZX8quFtfWNzS17e6eu4lRSVqOxiGUzIIoJHrEacBCsmUhGwkCwRjCojPzGPZOKx9ENDBPmhaQX8S6nBIzk2wdtYA+gK2YEPrnAWTsk0KdE6KvsVos08zW/y3y76JScMfA8caekiKao+vZXuxPTNGQRUEGUarlOAp4mEjgVLCu0U8USQgekx1qGRiRkytPjazJ8aJQO7sbSvAjwWP3doUmo1DAMTOVoWTXrjcT/vFYK3XNP8yhJgUV08lE3FRhiPIoGd7hkFMTQEEIlN7ti2ieSUDABFkwI7uzJ86R+XHINvz4tlp1pHHm0h/bREXLRGSqjS1RFNUTRI3pGr+jNerJerHfrY1Kas6Y9u+gPrM8fDLScIQ==\nAAAB2XicbZDNSgMxFIXv1L86Vq1rN8EiuCozbnQpuHFZwbZCO5RM5k4bmskMyR2hDH0BF25EfC93vo3pz0JbDwQ+zknIvSculLQUBN9ebWd3b/+gfugfNfzjk9Nmo2fz0gjsilzl5jnmFpXU2CVJCp8LgzyLFfbj6f0i77+gsTLXTzQrMMr4WMtUCk7O6oyaraAdLMW2IVxDC9YaNb+GSS7KDDUJxa0dhEFBUcUNSaFw7g9LiwUXUz7GgUPNM7RRtRxzzi6dk7A0N+5oYkv394uKZ9bOstjdzDhN7Ga2MP/LBiWlt1EldVESarH6KC0Vo5wtdmaJNChIzRxwYaSblYkJN1yQa8Z3HYSbG29D77odOn4MoA7ncAFXEMIN3MEDdKALAhJ4hXdv4r15H6uuat66tDP4I+/zBzjGijg=\nAAACAnicbZC7TsMwFIZPuJZSILCyWBQkpiqBAcSE1IWNItGL1JbIcd3W1LnIPkFUUV6AhVdhYQAhHoKNt8FpO0DLkSx9+n/7+Jzfj6XQ6Djf1tLyyuraemGjuFna2t6xd0sNHSWK8TqLZKRaPtVcipDXUaDkrVhxGviSN/1RNfebD1xpEYW3OI55N6CDUPQFo2gkzz7sIH/EtGpakNMLknUCikNGZXqd3aUyybxU3GeeXXYqzqTIIrgzKMOsap791elFLAl4iExSrduuE2M3pQoFkzwrdhLNY8pGdMDbBkMacN1NJ9tk5MgoPdKPlDkhkon6+0VKA63HgW9u5sPqeS8X//PaCfbPu6kI4wR5yKYf9RNJMCJ5NKQnFGcoxwYoU8LMStiQKsrQBFg0IbjzKy9C46TiGr5xoAD7cADH4MIZXMIV1KAODJ7gBd7g3Xq2Xq2PaVxL1iy3PfhT1ucPXLKauw==\nAAACAnicbZC7TsMwFIZPuJZSILCyWBQkpiqBAcSE1IWNItGL1JbIcd3W1LnIPkFUUV6AhVdhYQAhHoKNt8FpO0DLkSx9+n/7+Jzfj6XQ6Djf1tLyyuraemGjuFna2t6xd0sNHSWK8TqLZKRaPtVcipDXUaDkrVhxGviSN/1RNfebD1xpEYW3OI55N6CDUPQFo2gkzz7sIH/EtGpakNMLknUCikNGZXqd3aUyybxU3GeeXXYqzqTIIrgzKMOsap791elFLAl4iExSrduuE2M3pQoFkzwrdhLNY8pGdMDbBkMacN1NJ9tk5MgoPdKPlDkhkon6+0VKA63HgW9u5sPqeS8X//PaCfbPu6kI4wR5yKYf9RNJMCJ5NKQnFGcoxwYoU8LMStiQKsrQBFg0IbjzKy9C46TiGr5xoAD7cADH4MIZXMIV1KAODJ7gBd7g3Xq2Xq2PaVxL1iy3PfhT1ucPXLKauw==\nAAACDXicbVC7TsMwFHV4lvIKMLJYFCSmKoEBxFSpCxtFog+pDZHjOq2p85B9g6ii/AALv8LCAEKs7Gz8DU6bAVqOZOnonHuv7z1eLLgCy/o2FhaXlldWS2vl9Y3NrW1zZ7elokRS1qSRiGTHI4oJHrImcBCsE0tGAk+wtjeq5377nknFo/AGxjFzAjIIuc8pAS255mEP2AOkdT0Cn17grBcQGFIi0qvsNhVJ5qb8LnPNilW1JsDzxC5IBRVouOZXrx/RJGAhUEGU6tpWDE5KJHAqWFbuJYrFhI7IgHU1DUnAlJNOrsnwkVb62I+kfiHgifq7IyWBUuPA05X5smrWy8X/vG4C/rmT8jBOgIV0+pGfCAwRzqPBfS4ZBTHWhFDJ9a6YDokkFHSAZR2CPXvyPGmdVG3Nr61KzSriKKF9dICOkY3OUA1dogZqIooe0TN6RW/Gk/FivBsf09IFo+jZQ39gfP4AC3ScHQ==\nAAACDXicbVC7SgNBFJ2Nrxhfq5Y2g1GwCrsqKFaBNHZGMA9I1mV2MknGzD6YuSuGYX/Axl+xsVDE1t7Ov3HyKDTxwMDhnHvv3HuCRHAFjvNt5RYWl5ZX8quFtfWNzS17e6eu4lRSVqOxiGUzIIoJHrEacBCsmUhGwkCwRjCojPzGPZOKx9ENDBPmhaQX8S6nBIzk2wdtYA+gK2YEPrnAWTsk0KdE6KvsVos08zW/y3y76JScMfA8caekiKao+vZXuxPTNGQRUEGUarlOAp4mEjgVLCu0U8USQgekx1qGRiRkytPjazJ8aJQO7sbSvAjwWP3doUmo1DAMTOVoWTXrjcT/vFYK3XNP8yhJgUV08lE3FRhiPIoGd7hkFMTQEEIlN7ti2ieSUDABFkwI7uzJ86R+XHINvz4tlp1pHHm0h/bREXLRGSqjS1RFNUTRI3pGr+jNerJerHfrY1Kas6Y9u+gPrM8fDLScIQ==\nAAACDXicbVC7SgNBFJ2Nrxhfq5Y2g1GwCrsqKFaBNHZGMA9I1mV2MknGzD6YuSuGYX/Axl+xsVDE1t7Ov3HyKDTxwMDhnHvv3HuCRHAFjvNt5RYWl5ZX8quFtfWNzS17e6eu4lRSVqOxiGUzIIoJHrEacBCsmUhGwkCwRjCojPzGPZOKx9ENDBPmhaQX8S6nBIzk2wdtYA+gK2YEPrnAWTsk0KdE6KvsVos08zW/y3y76JScMfA8caekiKao+vZXuxPTNGQRUEGUarlOAp4mEjgVLCu0U8USQgekx1qGRiRkytPjazJ8aJQO7sbSvAjwWP3doUmo1DAMTOVoWTXrjcT/vFYK3XNP8yhJgUV08lE3FRhiPIoGd7hkFMTQEEIlN7ti2ieSUDABFkwI7uzJ86R+XHINvz4tlp1pHHm0h/bREXLRGSqjS1RFNUTRI3pGr+jNerJerHfrY1Kas6Y9u+gPrM8fDLScIQ==\nAAACDXicbVC7SgNBFJ2Nrxhfq5Y2g1GwCrsqKFaBNHZGMA9I1mV2MknGzD6YuSuGYX/Axl+xsVDE1t7Ov3HyKDTxwMDhnHvv3HuCRHAFjvNt5RYWl5ZX8quFtfWNzS17e6eu4lRSVqOxiGUzIIoJHrEacBCsmUhGwkCwRjCojPzGPZOKx9ENDBPmhaQX8S6nBIzk2wdtYA+gK2YEPrnAWTsk0KdE6KvsVos08zW/y3y76JScMfA8caekiKao+vZXuxPTNGQRUEGUarlOAp4mEjgVLCu0U8USQgekx1qGRiRkytPjazJ8aJQO7sbSvAjwWP3doUmo1DAMTOVoWTXrjcT/vFYK3XNP8yhJgUV08lE3FRhiPIoGd7hkFMTQEEIlN7ti2ieSUDABFkwI7uzJ86R+XHINvz4tlp1pHHm0h/bREXLRGSqjS1RFNUTRI3pGr+jNerJerHfrY1Kas6Y9u+gPrM8fDLScIQ==\nAAACDXicbVC7SgNBFJ2Nrxhfq5Y2g1GwCrsqKFaBNHZGMA9I1mV2MknGzD6YuSuGYX/Axl+xsVDE1t7Ov3HyKDTxwMDhnHvv3HuCRHAFjvNt5RYWl5ZX8quFtfWNzS17e6eu4lRSVqOxiGUzIIoJHrEacBCsmUhGwkCwRjCojPzGPZOKx9ENDBPmhaQX8S6nBIzk2wdtYA+gK2YEPrnAWTsk0KdE6KvsVos08zW/y3y76JScMfA8caekiKao+vZXuxPTNGQRUEGUarlOAp4mEjgVLCu0U8USQgekx1qGRiRkytPjazJ8aJQO7sbSvAjwWP3doUmo1DAMTOVoWTXrjcT/vFYK3XNP8yhJgUV08lE3FRhiPIoGd7hkFMTQEEIlN7ti2ieSUDABFkwI7uzJ86R+XHINvz4tlp1pHHm0h/bREXLRGSqjS1RFNUTRI3pGr+jNerJerHfrY1Kas6Y9u+gPrM8fDLScIQ==\nAAACDXicbVC7SgNBFJ2Nrxhfq5Y2g1GwCrsqKFaBNHZGMA9I1mV2MknGzD6YuSuGYX/Axl+xsVDE1t7Ov3HyKDTxwMDhnHvv3HuCRHAFjvNt5RYWl5ZX8quFtfWNzS17e6eu4lRSVqOxiGUzIIoJHrEacBCsmUhGwkCwRjCojPzGPZOKx9ENDBPmhaQX8S6nBIzk2wdtYA+gK2YEPrnAWTsk0KdE6KvsVos08zW/y3y76JScMfA8caekiKao+vZXuxPTNGQRUEGUarlOAp4mEjgVLCu0U8USQgekx1qGRiRkytPjazJ8aJQO7sbSvAjwWP3doUmo1DAMTOVoWTXrjcT/vFYK3XNP8yhJgUV08lE3FRhiPIoGd7hkFMTQEEIlN7ti2ieSUDABFkwI7uzJ86R+XHINvz4tlp1pHHm0h/bREXLRGSqjS1RFNUTRI3pGr+jNerJerHfrY1Kas6Y9u+gPrM8fDLScIQ==\nAAACDXicbVC7SgNBFJ2Nrxhfq5Y2g1GwCrsqKFaBNHZGMA9I1mV2MknGzD6YuSuGYX/Axl+xsVDE1t7Ov3HyKDTxwMDhnHvv3HuCRHAFjvNt5RYWl5ZX8quFtfWNzS17e6eu4lRSVqOxiGUzIIoJHrEacBCsmUhGwkCwRjCojPzGPZOKx9ENDBPmhaQX8S6nBIzk2wdtYA+gK2YEPrnAWTsk0KdE6KvsVos08zW/y3y76JScMfA8caekiKao+vZXuxPTNGQRUEGUarlOAp4mEjgVLCu0U8USQgekx1qGRiRkytPjazJ8aJQO7sbSvAjwWP3doUmo1DAMTOVoWTXrjcT/vFYK3XNP8yhJgUV08lE3FRhiPIoGd7hkFMTQEEIlN7ti2ieSUDABFkwI7uzJ86R+XHINvz4tlp1pHHm0h/bREXLRGSqjS1RFNUTRI3pGr+jNerJerHfrY1Kas6Y9u+gPrM8fDLScIQ==\n\n\u02c6AAAAB+XicbVDLSsNAFL3xWesr6tJNsAiuSlIEXVbcuKxgH9CUMplO2qGTSZi5KZSQP3HjQhG3/ok7/8ZJm4W2Hhg4nHMv98wJEsE1uu63tbG5tb2zW9mr7h8cHh3bJ6cdHaeKsjaNRax6AdFMcMnayFGwXqIYiQLBusH0vvC7M6Y0j+UTzhM2iMhY8pBTgkYa2rY/IZj5EcFJEGZ3eT60a27dXcBZJ15JalCiNbS//FFM04hJpIJo3ffcBAcZUcipYHnVTzVLCJ2SMesbKknE9CBbJM+dS6OMnDBW5kl0FurvjYxEWs+jwEwWEfWqV4j/ef0Uw9tBxmWSIpN0eShMhYOxU9TgjLhiFMXcEEIVN1kdOiGKUDRlVU0J3uqX10mnUfcMf7yuNRtlHRU4hwu4Ag9uoAkP0II2UJjBM7zCm5VZL9a79bEc3bDKnTP4A+vzB/1ek9M=\n\nAAAB+XicbVDLSsNAFL3xWesr6tJNsAiuSlIEXVbcuKxgH9CUMplO2qGTSZi5KZSQP3HjQhG3/ok7/8ZJm4W2Hhg4nHMv98wJEsE1uu63tbG5tb2zW9mr7h8cHh3bJ6cdHaeKsjaNRax6AdFMcMnayFGwXqIYiQLBusH0vvC7M6Y0j+UTzhM2iMhY8pBTgkYa2rY/IZj5EcFJEGZ3eT60a27dXcBZJ15JalCiNbS//FFM04hJpIJo3ffcBAcZUcipYHnVTzVLCJ2SMesbKknE9CBbJM+dS6OMnDBW5kl0FurvjYxEWs+jwEwWEfWqV4j/ef0Uw9tBxmWSIpN0eShMhYOxU9TgjLhiFMXcEEIVN1kdOiGKUDRlVU0J3uqX10mnUfcMf7yuNRtlHRU4hwu4Ag9uoAkP0II2UJjBM7zCm5VZL9a79bEc3bDKnTP4A+vzB/1ek9M=\nAAAB+XicbVDLSsNAFL3xWesr6tJNsAiuSlIEXVbcuKxgH9CUMplO2qGTSZi5KZSQP3HjQhG3/ok7/8ZJm4W2Hhg4nHMv98wJEsE1uu63tbG5tb2zW9mr7h8cHh3bJ6cdHaeKsjaNRax6AdFMcMnayFGwXqIYiQLBusH0vvC7M6Y0j+UTzhM2iMhY8pBTgkYa2rY/IZj5EcFJEGZ3eT60a27dXcBZJ15JalCiNbS//FFM04hJpIJo3ffcBAcZUcipYHnVTzVLCJ2SMesbKknE9CBbJM+dS6OMnDBW5kl0FurvjYxEWs+jwEwWEfWqV4j/ef0Uw9tBxmWSIpN0eShMhYOxU9TgjLhiFMXcEEIVN1kdOiGKUDRlVU0J3uqX10mnUfcMf7yuNRtlHRU4hwu4Ag9uoAkP0II2UJjBM7zCm5VZL9a79bEc3bDKnTP4A+vzB/1ek9M=\nAAAB+XicbVDLSsNAFL3xWesr6tJNsAiuSlIEXVbcuKxgH9CUMplO2qGTSZi5KZSQP3HjQhG3/ok7/8ZJm4W2Hhg4nHMv98wJEsE1uu63tbG5tb2zW9mr7h8cHh3bJ6cdHaeKsjaNRax6AdFMcMnayFGwXqIYiQLBusH0vvC7M6Y0j+UTzhM2iMhY8pBTgkYa2rY/IZj5EcFJEGZ3eT60a27dXcBZJ15JalCiNbS//FFM04hJpIJo3ffcBAcZUcipYHnVTzVLCJ2SMesbKknE9CBbJM+dS6OMnDBW5kl0FurvjYxEWs+jwEwWEfWqV4j/ef0Uw9tBxmWSIpN0eShMhYOxU9TgjLhiFMXcEEIVN1kdOiGKUDRlVU0J3uqX10mnUfcMf7yuNRtlHRU4hwu4Ag9uoAkP0II2UJjBM7zCm5VZL9a79bEc3bDKnTP4A+vzB/1ek9M=\n\nAAAB/XicbVDLSsNAFL2pr1pf8bFzM1gEVyUpgi4LblxWsA9paplMJ+3QySTMTIQagr/ixoUibv0Pd/6NkzYLbT0wcDjnXu6Z48ecKe0431ZpZXVtfaO8Wdna3tnds/cP2ipKJKEtEvFIdn2sKGeCtjTTnHZjSXHoc9rxJ1e533mgUrFI3OppTPshHgkWMIK1kQb2kTfGOvVCrMd+kN5l2X3Ks4FddWrODGiZuAWpQoHmwP7yhhFJQio04VipnuvEup9iqRnhNKt4iaIxJhM8oj1DBQ6p6qez9Bk6NcoQBZE0T2g0U39vpDhUahr6ZjKPqRa9XPzP6yU6uOynTMSJpoLMDwUJRzpCeRVoyCQlmk8NwUQykxWRMZaYaFNYxZTgLn55mbTrNdfwm/Nqo17UUYZjOIEzcOECGnANTWgBgUd4hld4s56sF+vd+piPlqxi5xD+wPr8AX3+ldU=\nAAAB/XicbVDLSsNAFL2pr1pf8bFzM1gEVyUpgi4LblxWsA9paplMJ+3QySTMTIQagr/ixoUibv0Pd/6NkzYLbT0wcDjnXu6Z48ecKe0431ZpZXVtfaO8Wdna3tnds/cP2ipKJKEtEvFIdn2sKGeCtjTTnHZjSXHoc9rxJ1e533mgUrFI3OppTPshHgkWMIK1kQb2kTfGOvVCrMd+kN5l2X3Ks4FddWrODGiZuAWpQoHmwP7yhhFJQio04VipnuvEup9iqRnhNKt4iaIxJhM8oj1DBQ6p6qez9Bk6NcoQBZE0T2g0U39vpDhUahr6ZjKPqRa9XPzP6yU6uOynTMSJpoLMDwUJRzpCeRVoyCQlmk8NwUQykxWRMZaYaFNYxZTgLn55mbTrNdfwm/Nqo17UUYZjOIEzcOECGnANTWgBgUd4hld4s56sF+vd+piPlqxi5xD+wPr8AX3+ldU=\nAAAB/XicbVDLSsNAFL2pr1pf8bFzM1gEVyUpgi4LblxWsA9paplMJ+3QySTMTIQagr/ixoUibv0Pd/6NkzYLbT0wcDjnXu6Z48ecKe0431ZpZXVtfaO8Wdna3tnds/cP2ipKJKEtEvFIdn2sKGeCtjTTnHZjSXHoc9rxJ1e533mgUrFI3OppTPshHgkWMIK1kQb2kTfGOvVCrMd+kN5l2X3Ks4FddWrODGiZuAWpQoHmwP7yhhFJQio04VipnuvEup9iqRnhNKt4iaIxJhM8oj1DBQ6p6qez9Bk6NcoQBZE0T2g0U39vpDhUahr6ZjKPqRa9XPzP6yU6uOynTMSJpoLMDwUJRzpCeRVoyCQlmk8NwUQykxWRMZaYaFNYxZTgLn55mbTrNdfwm/Nqo17UUYZjOIEzcOECGnANTWgBgUd4hld4s56sF+vd+piPlqxi5xD+wPr8AX3+ldU=\nAAAB/XicbVDLSsNAFL2pr1pf8bFzM1gEVyUpgi4LblxWsA9paplMJ+3QySTMTIQagr/ixoUibv0Pd/6NkzYLbT0wcDjnXu6Z48ecKe0431ZpZXVtfaO8Wdna3tnds/cP2ipKJKEtEvFIdn2sKGeCtjTTnHZjSXHoc9rxJ1e533mgUrFI3OppTPshHgkWMIK1kQb2kTfGOvVCrMd+kN5l2X3Ks4FddWrODGiZuAWpQoHmwP7yhhFJQio04VipnuvEup9iqRnhNKt4iaIxJhM8oj1DBQ6p6qez9Bk6NcoQBZE0T2g0U39vpDhUahr6ZjKPqRa9XPzP6yU6uOynTMSJpoLMDwUJRzpCeRVoyCQlmk8NwUQykxWRMZaYaFNYxZTgLn55mbTrNdfwm/Nqo17UUYZjOIEzcOECGnANTWgBgUd4hld4s56sF+vd+piPlqxi5xD+wPr8AX3+ldU=\n\n\u02c6Yl\nYu\n\nAAAB9XicbVDLSgMxFL3xWeur6tJNsAiuykwRdFlw47KCfUg7LZk004ZmMkOSUcow/+HGhSJu/Rd3/o2ZdhbaeiBwOOde7snxY8G1cZxvtLa+sbm1Xdop7+7tHxxWjo7bOkoUZS0aiUh1faKZ4JK1DDeCdWPFSOgL1vGnN7nfeWRK80jem1nMvJCMJQ84JcZKg35IzMQP0odskCbZsFJ1as4ceJW4BalCgeaw8tUfRTQJmTRUEK17rhMbLyXKcCpYVu4nmsWETsmY9SyVJGTaS+epM3xulREOImWfNHiu/t5ISaj1LPTtZJ5SL3u5+J/XS0xw7aVcxolhki4OBYnAJsJ5BXjEFaNGzCwhVHGbFdMJUYQaW1TZluAuf3mVtOs11/K7y2qjXtRRglM4gwtw4QoacAtNaAEFBc/wCm/oCb2gd/SxGF1Dxc4J/AH6/AEoHZLg\nAAAB9XicbVDLSgMxFL3xWeur6tJNsAiuykwRdFlw47KCfUg7LZk004ZmMkOSUcow/+HGhSJu/Rd3/o2ZdhbaeiBwOOde7snxY8G1cZxvtLa+sbm1Xdop7+7tHxxWjo7bOkoUZS0aiUh1faKZ4JK1DDeCdWPFSOgL1vGnN7nfeWRK80jem1nMvJCMJQ84JcZKg35IzMQP0odskCbZsFJ1as4ceJW4BalCgeaw8tUfRTQJmTRUEK17rhMbLyXKcCpYVu4nmsWETsmY9SyVJGTaS+epM3xulREOImWfNHiu/t5ISaj1LPTtZJ5SL3u5+J/XS0xw7aVcxolhki4OBYnAJsJ5BXjEFaNGzCwhVHGbFdMJUYQaW1TZluAuf3mVtOs11/K7y2qjXtRRglM4gwtw4QoacAtNaAEFBc/wCm/oCb2gd/SxGF1Dxc4J/AH6/AEoHZLg\nAAAB9XicbVDLSgMxFL3xWeur6tJNsAiuykwRdFlw47KCfUg7LZk004ZmMkOSUcow/+HGhSJu/Rd3/o2ZdhbaeiBwOOde7snxY8G1cZxvtLa+sbm1Xdop7+7tHxxWjo7bOkoUZS0aiUh1faKZ4JK1DDeCdWPFSOgL1vGnN7nfeWRK80jem1nMvJCMJQ84JcZKg35IzMQP0odskCbZsFJ1as4ceJW4BalCgeaw8tUfRTQJmTRUEK17rhMbLyXKcCpYVu4nmsWETsmY9SyVJGTaS+epM3xulREOImWfNHiu/t5ISaj1LPTtZJ5SL3u5+J/XS0xw7aVcxolhki4OBYnAJsJ5BXjEFaNGzCwhVHGbFdMJUYQaW1TZluAuf3mVtOs11/K7y2qjXtRRglM4gwtw4QoacAtNaAEFBc/wCm/oCb2gd/SxGF1Dxc4J/AH6/AEoHZLg\nAAAB2XicbZDNSgMxFIXv1L86Vq1rN8EiuCozbnQpuHFZwbZCO5RM5k4bmskMyR2hDH0BF25EfC93vo3pz0JbDwQ+zknIvSculLQUBN9ebWd3b/+gfugfNfzjk9Nmo2fz0gjsilzl5jnmFpXU2CVJCp8LgzyLFfbj6f0i77+gsTLXTzQrMMr4WMtUCk7O6oyaraAdLMW2IVxDC9YaNb+GSS7KDDUJxa0dhEFBUcUNSaFw7g9LiwUXUz7GgUPNM7RRtRxzzi6dk7A0N+5oYkv394uKZ9bOstjdzDhN7Ga2MP/LBiWlt1EldVESarH6KC0Vo5wtdmaJNChIzRxwYaSblYkJN1yQa8Z3HYSbG29D77odOn4MoA7ncAFXEMIN3MEDdKALAhJ4hXdv4r15H6uuat66tDP4I+/zBzjGijg=\nAAAB6nicbZC9TsMwFIVvyl8pBQori0WFxFQlLDBWYmEsEv1BbVo57k1r1XEi2wFVUd6DhQGEeCA23gan7QAtV7L06Rxb9/gEieDauO63U9ra3tndK+9XDqqHR8e1k2pHx6li2GaxiFUvoBoFl9g23AjsJQppFAjsBrPbwu8+odI8lg9mnqAf0YnkIWfUWGk4iKiZBmH2mA+zNB/V6m7DXQzZBG8FdVhNa1T7GoxjlkYoDRNU677nJsbPqDKcCcwrg1RjQtmMTrBvUdIItZ8tUufkwipjEsbKHmnIQv39IqOR1vMosDeLlHrdK8T/vH5qwhs/4zJJDUq2XBSmgpiYFBWQMVfIjJhboExxm5WwKVWUGVtUxZbgrX95EzpXDc/yvQtlOINzuAQPrqEJd9CCNjBQ8AJv8O48O6/Ox7KukrPq7RT+jPP5A9kSkZA=\nAAAB6nicbZC9TsMwFIVvyl8pBQori0WFxFQlLDBWYmEsEv1BbVo57k1r1XEi2wFVUd6DhQGEeCA23gan7QAtV7L06Rxb9/gEieDauO63U9ra3tndK+9XDqqHR8e1k2pHx6li2GaxiFUvoBoFl9g23AjsJQppFAjsBrPbwu8+odI8lg9mnqAf0YnkIWfUWGk4iKiZBmH2mA+zNB/V6m7DXQzZBG8FdVhNa1T7GoxjlkYoDRNU677nJsbPqDKcCcwrg1RjQtmMTrBvUdIItZ8tUufkwipjEsbKHmnIQv39IqOR1vMosDeLlHrdK8T/vH5qwhs/4zJJDUq2XBSmgpiYFBWQMVfIjJhboExxm5WwKVWUGVtUxZbgrX95EzpXDc/yvQtlOINzuAQPrqEJd9CCNjBQ8AJv8O48O6/Ox7KukrPq7RT+jPP5A9kSkZA=\nAAAB9XicbVC7TsMwFL0pr1JeBUYWiwqJqUq6wFiJhbFI9IHatHJcp7VqO5HtgKoo/8HCAEKs/Asbf4PTZoCWI1k6Oude3eMTxJxp47rfTmljc2t7p7xb2ds/ODyqHp90dJQoQtsk4pHqBVhTziRtG2Y47cWKYhFw2g1mN7nffaRKs0jem3lMfYEnkoWMYGOl4UBgMw3C9CEbpkk2qtbcursAWideQWpQoDWqfg3GEUkElYZwrHXfc2Pjp1gZRjjNKoNE0xiTGZ7QvqUSC6r9dJE6QxdWGaMwUvZJgxbq740UC63nIrCTeUq96uXif14/MeG1nzIZJ4ZKsjwUJhyZCOUVoDFTlBg+twQTxWxWRKZYYWJsURVbgrf65XXSadQ9y+/cWrNR1FGGMziHS/DgCppwCy1oAwEFz/AKb86T8+K8Ox/L0ZJT7JzCHzifPybdktw=\nAAAB9XicbVDLSgMxFL3xWeur6tJNsAiuykwRdFlw47KCfUg7LZk004ZmMkOSUcow/+HGhSJu/Rd3/o2ZdhbaeiBwOOde7snxY8G1cZxvtLa+sbm1Xdop7+7tHxxWjo7bOkoUZS0aiUh1faKZ4JK1DDeCdWPFSOgL1vGnN7nfeWRK80jem1nMvJCMJQ84JcZKg35IzMQP0odskCbZsFJ1as4ceJW4BalCgeaw8tUfRTQJmTRUEK17rhMbLyXKcCpYVu4nmsWETsmY9SyVJGTaS+epM3xulREOImWfNHiu/t5ISaj1LPTtZJ5SL3u5+J/XS0xw7aVcxolhki4OBYnAJsJ5BXjEFaNGzCwhVHGbFdMJUYQaW1TZluAuf3mVtOs11/K7y2qjXtRRglM4gwtw4QoacAtNaAEFBc/wCm/oCb2gd/SxGF1Dxc4J/AH6/AEoHZLg\nAAAB9XicbVDLSgMxFL3xWeur6tJNsAiuykwRdFlw47KCfUg7LZk004ZmMkOSUcow/+HGhSJu/Rd3/o2ZdhbaeiBwOOde7snxY8G1cZxvtLa+sbm1Xdop7+7tHxxWjo7bOkoUZS0aiUh1faKZ4JK1DDeCdWPFSOgL1vGnN7nfeWRK80jem1nMvJCMJQ84JcZKg35IzMQP0odskCbZsFJ1as4ceJW4BalCgeaw8tUfRTQJmTRUEK17rhMbLyXKcCpYVu4nmsWETsmY9SyVJGTaS+epM3xulREOImWfNHiu/t5ISaj1LPTtZJ5SL3u5+J/XS0xw7aVcxolhki4OBYnAJsJ5BXjEFaNGzCwhVHGbFdMJUYQaW1TZluAuf3mVtOs11/K7y2qjXtRRglM4gwtw4QoacAtNaAEFBc/wCm/oCb2gd/SxGF1Dxc4J/AH6/AEoHZLg\nAAAB9XicbVDLSgMxFL3xWeur6tJNsAiuykwRdFlw47KCfUg7LZk004ZmMkOSUcow/+HGhSJu/Rd3/o2ZdhbaeiBwOOde7snxY8G1cZxvtLa+sbm1Xdop7+7tHxxWjo7bOkoUZS0aiUh1faKZ4JK1DDeCdWPFSOgL1vGnN7nfeWRK80jem1nMvJCMJQ84JcZKg35IzMQP0odskCbZsFJ1as4ceJW4BalCgeaw8tUfRTQJmTRUEK17rhMbLyXKcCpYVu4nmsWETsmY9SyVJGTaS+epM3xulREOImWfNHiu/t5ISaj1LPTtZJ5SL3u5+J/XS0xw7aVcxolhki4OBYnAJsJ5BXjEFaNGzCwhVHGbFdMJUYQaW1TZluAuf3mVtOs11/K7y2qjXtRRglM4gwtw4QoacAtNaAEFBc/wCm/oCb2gd/SxGF1Dxc4J/AH6/AEoHZLg\nAAAB9XicbVDLSgMxFL3xWeur6tJNsAiuykwRdFlw47KCfUg7LZk004ZmMkOSUcow/+HGhSJu/Rd3/o2ZdhbaeiBwOOde7snxY8G1cZxvtLa+sbm1Xdop7+7tHxxWjo7bOkoUZS0aiUh1faKZ4JK1DDeCdWPFSOgL1vGnN7nfeWRK80jem1nMvJCMJQ84JcZKg35IzMQP0odskCbZsFJ1as4ceJW4BalCgeaw8tUfRTQJmTRUEK17rhMbLyXKcCpYVu4nmsWETsmY9SyVJGTaS+epM3xulREOImWfNHiu/t5ISaj1LPTtZJ5SL3u5+J/XS0xw7aVcxolhki4OBYnAJsJ5BXjEFaNGzCwhVHGbFdMJUYQaW1TZluAuf3mVtOs11/K7y2qjXtRRglM4gwtw4QoacAtNaAEFBc/wCm/oCb2gd/SxGF1Dxc4J/AH6/AEoHZLg\nAAAB9XicbVDLSgMxFL3xWeur6tJNsAiuykwRdFlw47KCfUg7LZk004ZmMkOSUcow/+HGhSJu/Rd3/o2ZdhbaeiBwOOde7snxY8G1cZxvtLa+sbm1Xdop7+7tHxxWjo7bOkoUZS0aiUh1faKZ4JK1DDeCdWPFSOgL1vGnN7nfeWRK80jem1nMvJCMJQ84JcZKg35IzMQP0odskCbZsFJ1as4ceJW4BalCgeaw8tUfRTQJmTRUEK17rhMbLyXKcCpYVu4nmsWETsmY9SyVJGTaS+epM3xulREOImWfNHiu/t5ISaj1LPTtZJ5SL3u5+J/XS0xw7aVcxolhki4OBYnAJsJ5BXjEFaNGzCwhVHGbFdMJUYQaW1TZluAuf3mVtOs11/K7y2qjXtRRglM4gwtw4QoacAtNaAEFBc/wCm/oCb2gd/SxGF1Dxc4J/AH6/AEoHZLg\nAAAB9XicbVDLSgMxFL3xWeur6tJNsAiuykwRdFlw47KCfUg7LZk004ZmMkOSUcow/+HGhSJu/Rd3/o2ZdhbaeiBwOOde7snxY8G1cZxvtLa+sbm1Xdop7+7tHxxWjo7bOkoUZS0aiUh1faKZ4JK1DDeCdWPFSOgL1vGnN7nfeWRK80jem1nMvJCMJQ84JcZKg35IzMQP0odskCbZsFJ1as4ceJW4BalCgeaw8tUfRTQJmTRUEK17rhMbLyXKcCpYVu4nmsWETsmY9SyVJGTaS+epM3xulREOImWfNHiu/t5ISaj1LPTtZJ5SL3u5+J/XS0xw7aVcxolhki4OBYnAJsJ5BXjEFaNGzCwhVHGbFdMJUYQaW1TZluAuf3mVtOs11/K7y2qjXtRRglM4gwtw4QoacAtNaAEFBc/wCm/oCb2gd/SxGF1Dxc4J/AH6/AEoHZLg\n\nCase 4: Ola\n\nia\n\nAAACDXicbVC7TsMwFHV4lvIKMLJYFCSmKkGVQEyVurBRJPqQmlA5rtNadR6ybxBVlB9g4VdYGECIlZ2Nv8FpM0DLkWwdnXOvfe/xYsEVWNa3sbS8srq2Xtoob25t7+yae/ttFSWSshaNRCS7HlFM8JC1gINg3VgyEniCdbxxI/c790wqHoW3MImZG5BhyH1OCWipbx47wB4gbegncO0SZ05AYESJSK+zu1SQrJ9yfZkVq2pNgReJXZAKKtDsm1/OIKJJwEKggijVs60Y3JRI4FSwrOwkisWEjsmQ9TQNScCUm063yfCJVgbYj6Q+IeCp+rsjJYFSk8DTlfmwat7Lxf+8XgL+hZvyME6AhXT2kZ8IDBHOo8EDLhkFMdGEUMn1rJiOiCQUdIBlHYI9v/IiaZ9Vbc1vapW6VcRRQofoCJ0iG52jOrpCTdRCFD2iZ/SK3own48V4Nz5mpUtG0XOA/sD4/AHhzJwF\nAAACDXicbVC7TsMwFHV4lvIKMLJYFCSmKkGVQEyVurBRJPqQmlA5rtNadR6ybxBVlB9g4VdYGECIlZ2Nv8FpM0DLkWwdnXOvfe/xYsEVWNa3sbS8srq2Xtoob25t7+yae/ttFSWSshaNRCS7HlFM8JC1gINg3VgyEniCdbxxI/c790wqHoW3MImZG5BhyH1OCWipbx47wB4gbegncO0SZ05AYESJSK+zu1SQrJ9yfZkVq2pNgReJXZAKKtDsm1/OIKJJwEKggijVs60Y3JRI4FSwrOwkisWEjsmQ9TQNScCUm063yfCJVgbYj6Q+IeCp+rsjJYFSk8DTlfmwat7Lxf+8XgL+hZvyME6AhXT2kZ8IDBHOo8EDLhkFMdGEUMn1rJiOiCQUdIBlHYI9v/IiaZ9Vbc1vapW6VcRRQofoCJ0iG52jOrpCTdRCFD2iZ/SK3own48V4Nz5mpUtG0XOA/sD4/AHhzJwF\nAAACDXicbVC7TsMwFHV4lvIKMLJYFCSmKkGVQEyVurBRJPqQmlA5rtNadR6ybxBVlB9g4VdYGECIlZ2Nv8FpM0DLkWwdnXOvfe/xYsEVWNa3sbS8srq2Xtoob25t7+yae/ttFSWSshaNRCS7HlFM8JC1gINg3VgyEniCdbxxI/c790wqHoW3MImZG5BhyH1OCWipbx47wB4gbegncO0SZ05AYESJSK+zu1SQrJ9yfZkVq2pNgReJXZAKKtDsm1/OIKJJwEKggijVs60Y3JRI4FSwrOwkisWEjsmQ9TQNScCUm063yfCJVgbYj6Q+IeCp+rsjJYFSk8DTlfmwat7Lxf+8XgL+hZvyME6AhXT2kZ8IDBHOo8EDLhkFMdGEUMn1rJiOiCQUdIBlHYI9v/IiaZ9Vbc1vapW6VcRRQofoCJ0iG52jOrpCTdRCFD2iZ/SK3own48V4Nz5mpUtG0XOA/sD4/AHhzJwF\nAAACDXicbVC7TsMwFHV4lvIKMLJYFCSmKkGVQEyVurBRJPqQmlA5rtNadR6ybxBVlB9g4VdYGECIlZ2Nv8FpM0DLkWwdnXOvfe/xYsEVWNa3sbS8srq2Xtoob25t7+yae/ttFSWSshaNRCS7HlFM8JC1gINg3VgyEniCdbxxI/c790wqHoW3MImZG5BhyH1OCWipbx47wB4gbegncO0SZ05AYESJSK+zu1SQrJ9yfZkVq2pNgReJXZAKKtDsm1/OIKJJwEKggijVs60Y3JRI4FSwrOwkisWEjsmQ9TQNScCUm063yfCJVgbYj6Q+IeCp+rsjJYFSk8DTlfmwat7Lxf+8XgL+hZvyME6AhXT2kZ8IDBHOo8EDLhkFMdGEUMn1rJiOiCQUdIBlHYI9v/IiaZ9Vbc1vapW6VcRRQofoCJ0iG52jOrpCTdRCFD2iZ/SK3own48V4Nz5mpUtG0XOA/sD4/AHhzJwF\n\n\u02c6XAAAB+XicbVDLSsNAFL2pr1pfUZduBovgqiRF0GXBjcsK9gFNKJPppB06eTBzUyihf+LGhSJu/RN3/o2TNgttPTBwOOde7pkTpFJodJxvq7K1vbO7V92vHRweHZ/Yp2ddnWSK8Q5LZKL6AdVciph3UKDk/VRxGgWS94LpfeH3ZlxpkcRPOE+5H9FxLELBKBppaNvehGLuRRQnQZj3F4uhXXcazhJkk7glqUOJ9tD+8kYJyyIeI5NU64HrpOjnVKFgki9qXqZ5StmUjvnA0JhGXPv5MvmCXBllRMJEmRcjWaq/N3IaaT2PAjNZRNTrXiH+5w0yDO/8XMRphjxmq0NhJgkmpKiBjITiDOXcEMqUMFkJm1BFGZqyaqYEd/3Lm6TbbLiGP97UW82yjipcwCVcgwu30IIHaEMHGMzgGV7hzcqtF+vd+liNVqxy5xz+wPr8ASB3k+o=\n\nAAAB+XicbVDLSsNAFL2pr1pfUZduBovgqiRF0GXBjcsK9gFNKJPppB06eTBzUyihf+LGhSJu/RN3/o2TNgttPTBwOOde7pkTpFJodJxvq7K1vbO7V92vHRweHZ/Yp2ddnWSK8Q5LZKL6AdVciph3UKDk/VRxGgWS94LpfeH3ZlxpkcRPOE+5H9FxLELBKBppaNvehGLuRRQnQZj3F4uhXXcazhJkk7glqUOJ9tD+8kYJyyIeI5NU64HrpOjnVKFgki9qXqZ5StmUjvnA0JhGXPv5MvmCXBllRMJEmRcjWaq/N3IaaT2PAjNZRNTrXiH+5w0yDO/8XMRphjxmq0NhJgkmpKiBjITiDOXcEMqUMFkJm1BFGZqyaqYEd/3Lm6TbbLiGP97UW82yjipcwCVcgwu30IIHaEMHGMzgGV7hzcqtF+vd+liNVqxy5xz+wPr8ASB3k+o=\nAAAB+XicbVDLSsNAFL2pr1pfUZduBovgqiRF0GXBjcsK9gFNKJPppB06eTBzUyihf+LGhSJu/RN3/o2TNgttPTBwOOde7pkTpFJodJxvq7K1vbO7V92vHRweHZ/Yp2ddnWSK8Q5LZKL6AdVciph3UKDk/VRxGgWS94LpfeH3ZlxpkcRPOE+5H9FxLELBKBppaNvehGLuRRQnQZj3F4uhXXcazhJkk7glqUOJ9tD+8kYJyyIeI5NU64HrpOjnVKFgki9qXqZ5StmUjvnA0JhGXPv5MvmCXBllRMJEmRcjWaq/N3IaaT2PAjNZRNTrXiH+5w0yDO/8XMRphjxmq0NhJgkmpKiBjITiDOXcEMqUMFkJm1BFGZqyaqYEd/3Lm6TbbLiGP97UW82yjipcwCVcgwu30IIHaEMHGMzgGV7hzcqtF+vd+liNVqxy5xz+wPr8ASB3k+o=\nAAAB2XicbZDNSgMxFIXv1L86Vq1rN8EiuCozbnQpuHFZwbZCO5RM5k4bmskMyR2hDH0BF25EfC93vo3pz0JbDwQ+zknIvSculLQUBN9ebWd3b/+gfugfNfzjk9Nmo2fz0gjsilzl5jnmFpXU2CVJCp8LgzyLFfbj6f0i77+gsTLXTzQrMMr4WMtUCk7O6oyaraAdLMW2IVxDC9YaNb+GSS7KDDUJxa0dhEFBUcUNSaFw7g9LiwUXUz7GgUPNM7RRtRxzzi6dk7A0N+5oYkv394uKZ9bOstjdzDhN7Ga2MP/LBiWlt1EldVESarH6KC0Vo5wtdmaJNChIzRxwYaSblYkJN1yQa8Z3HYSbG29D77odOn4MoA7ncAFXEMIN3MEDdKALAhJ4hXdv4r15H6uuat66tDP4I+/zBzjGijg=\nAAAB7nicbZDNSsNAFIVv6l+tVaNbN4NFcFUSN7oU3LisYH+gCWUynbRDJ5Mwc1MoIW/ixoUiPo4738ZJ24W2Xhj4OGeGe+ZEmRQGPe/bqe3s7u0f1A8bR83jk1P3rNkzaa4Z77JUpnoQUcOlULyLAiUfZJrTJJK8H80eKr8/59qIVD3jIuNhQidKxIJRtNLIdYMpxSJIKE6juBiU5chteW1vOWQb/DW0YD2dkfsVjFOWJ1whk9SYoe9lGBZUo2CSl40gNzyjbEYnfGhR0YSbsFgmL8mVVcYkTrU9CslS/f2ioIkxiySyN6uIZtOrxP+8YY7xXVgIleXIFVstinNJMCVVDWQsNGcoFxYo08JmJWxKNWVoy2rYEvzNL29D76btW37yoA4XcAnX4MMt3MMjdKALDObwAm/w7hTOq/OxqqvmrHs7hz/jfP4AxraSlg==\nAAAB7nicbZDNSsNAFIVv6l+tVaNbN4NFcFUSN7oU3LisYH+gCWUynbRDJ5Mwc1MoIW/ixoUiPo4738ZJ24W2Xhj4OGeGe+ZEmRQGPe/bqe3s7u0f1A8bR83jk1P3rNkzaa4Z77JUpnoQUcOlULyLAiUfZJrTJJK8H80eKr8/59qIVD3jIuNhQidKxIJRtNLIdYMpxSJIKE6juBiU5chteW1vOWQb/DW0YD2dkfsVjFOWJ1whk9SYoe9lGBZUo2CSl40gNzyjbEYnfGhR0YSbsFgmL8mVVcYkTrU9CslS/f2ioIkxiySyN6uIZtOrxP+8YY7xXVgIleXIFVstinNJMCVVDWQsNGcoFxYo08JmJWxKNWVoy2rYEvzNL29D76btW37yoA4XcAnX4MMt3MMjdKALDObwAm/w7hTOq/OxqqvmrHs7hz/jfP4AxraSlg==\nAAAB+XicbVDLSsNAFL3xWesr6tJNsAiuStKNLgtuXFawD2hCmUwn7dDJJMzcFErIn7hxoYhb/8Sdf+OkzUJbDwwczrmXe+aEqeAaXffb2tre2d3brx3UD4+OT07ts/OeTjJFWZcmIlGDkGgmuGRd5CjYIFWMxKFg/XB2X/r9OVOaJ/IJFykLYjKRPOKUoJFGtu1PCeZ+THAaRvmgKEZ2w226SzibxKtIAyp0RvaXP05oFjOJVBCth56bYpAThZwKVtT9TLOU0BmZsKGhksRMB/kyeeFcG2XsRIkyT6KzVH9v5CTWehGHZrKMqNe9UvzPG2YY3QU5l2mGTNLVoSgTDiZOWYMz5opRFAtDCFXcZHXolChC0ZRVNyV461/eJL1W0zP80W20W1UdNbiEK7gBD26hDQ/QgS5QmMMzvMKblVsv1rv1sRrdsqqdC/gD6/MHHzeT5g==\nAAAB+XicbVDLSsNAFL2pr1pfUZduBovgqiRF0GXBjcsK9gFNKJPppB06eTBzUyihf+LGhSJu/RN3/o2TNgttPTBwOOde7pkTpFJodJxvq7K1vbO7V92vHRweHZ/Yp2ddnWSK8Q5LZKL6AdVciph3UKDk/VRxGgWS94LpfeH3ZlxpkcRPOE+5H9FxLELBKBppaNvehGLuRRQnQZj3F4uhXXcazhJkk7glqUOJ9tD+8kYJyyIeI5NU64HrpOjnVKFgki9qXqZ5StmUjvnA0JhGXPv5MvmCXBllRMJEmRcjWaq/N3IaaT2PAjNZRNTrXiH+5w0yDO/8XMRphjxmq0NhJgkmpKiBjITiDOXcEMqUMFkJm1BFGZqyaqYEd/3Lm6TbbLiGP97UW82yjipcwCVcgwu30IIHaEMHGMzgGV7hzcqtF+vd+liNVqxy5xz+wPr8ASB3k+o=\nAAAB+XicbVDLSsNAFL2pr1pfUZduBovgqiRF0GXBjcsK9gFNKJPppB06eTBzUyihf+LGhSJu/RN3/o2TNgttPTBwOOde7pkTpFJodJxvq7K1vbO7V92vHRweHZ/Yp2ddnWSK8Q5LZKL6AdVciph3UKDk/VRxGgWS94LpfeH3ZlxpkcRPOE+5H9FxLELBKBppaNvehGLuRRQnQZj3F4uhXXcazhJkk7glqUOJ9tD+8kYJyyIeI5NU64HrpOjnVKFgki9qXqZ5StmUjvnA0JhGXPv5MvmCXBllRMJEmRcjWaq/N3IaaT2PAjNZRNTrXiH+5w0yDO/8XMRphjxmq0NhJgkmpKiBjITiDOXcEMqUMFkJm1BFGZqyaqYEd/3Lm6TbbLiGP97UW82yjipcwCVcgwu30IIHaEMHGMzgGV7hzcqtF+vd+liNVqxy5xz+wPr8ASB3k+o=\nAAAB+XicbVDLSsNAFL2pr1pfUZduBovgqiRF0GXBjcsK9gFNKJPppB06eTBzUyihf+LGhSJu/RN3/o2TNgttPTBwOOde7pkTpFJodJxvq7K1vbO7V92vHRweHZ/Yp2ddnWSK8Q5LZKL6AdVciph3UKDk/VRxGgWS94LpfeH3ZlxpkcRPOE+5H9FxLELBKBppaNvehGLuRRQnQZj3F4uhXXcazhJkk7glqUOJ9tD+8kYJyyIeI5NU64HrpOjnVKFgki9qXqZ5StmUjvnA0JhGXPv5MvmCXBllRMJEmRcjWaq/N3IaaT2PAjNZRNTrXiH+5w0yDO/8XMRphjxmq0NhJgkmpKiBjITiDOXcEMqUMFkJm1BFGZqyaqYEd/3Lm6TbbLiGP97UW82yjipcwCVcgwu30IIHaEMHGMzgGV7hzcqtF+vd+liNVqxy5xz+wPr8ASB3k+o=\nAAAB+XicbVDLSsNAFL2pr1pfUZduBovgqiRF0GXBjcsK9gFNKJPppB06eTBzUyihf+LGhSJu/RN3/o2TNgttPTBwOOde7pkTpFJodJxvq7K1vbO7V92vHRweHZ/Yp2ddnWSK8Q5LZKL6AdVciph3UKDk/VRxGgWS94LpfeH3ZlxpkcRPOE+5H9FxLELBKBppaNvehGLuRRQnQZj3F4uhXXcazhJkk7glqUOJ9tD+8kYJyyIeI5NU64HrpOjnVKFgki9qXqZ5StmUjvnA0JhGXPv5MvmCXBllRMJEmRcjWaq/N3IaaT2PAjNZRNTrXiH+5w0yDO/8XMRphjxmq0NhJgkmpKiBjITiDOXcEMqUMFkJm1BFGZqyaqYEd/3Lm6TbbLiGP97UW82yjipcwCVcgwu30IIHaEMHGMzgGV7hzcqtF+vd+liNVqxy5xz+wPr8ASB3k+o=\nAAAB+XicbVDLSsNAFL2pr1pfUZduBovgqiRF0GXBjcsK9gFNKJPppB06eTBzUyihf+LGhSJu/RN3/o2TNgttPTBwOOde7pkTpFJodJxvq7K1vbO7V92vHRweHZ/Yp2ddnWSK8Q5LZKL6AdVciph3UKDk/VRxGgWS94LpfeH3ZlxpkcRPOE+5H9FxLELBKBppaNvehGLuRRQnQZj3F4uhXXcazhJkk7glqUOJ9tD+8kYJyyIeI5NU64HrpOjnVKFgki9qXqZ5StmUjvnA0JhGXPv5MvmCXBllRMJEmRcjWaq/N3IaaT2PAjNZRNTrXiH+5w0yDO/8XMRphjxmq0NhJgkmpKiBjITiDOXcEMqUMFkJm1BFGZqyaqYEd/3Lm6TbbLiGP97UW82yjipcwCVcgwu30IIHaEMHGMzgGV7hzcqtF+vd+liNVqxy5xz+wPr8ASB3k+o=\nAAAB+XicbVDLSsNAFL2pr1pfUZduBovgqiRF0GXBjcsK9gFNKJPppB06eTBzUyihf+LGhSJu/RN3/o2TNgttPTBwOOde7pkTpFJodJxvq7K1vbO7V92vHRweHZ/Yp2ddnWSK8Q5LZKL6AdVciph3UKDk/VRxGgWS94LpfeH3ZlxpkcRPOE+5H9FxLELBKBppaNvehGLuRRQnQZj3F4uhXXcazhJkk7glqUOJ9tD+8kYJyyIeI5NU64HrpOjnVKFgki9qXqZ5StmUjvnA0JhGXPv5MvmCXBllRMJEmRcjWaq/N3IaaT2PAjNZRNTrXiH+5w0yDO/8XMRphjxmq0NhJgkmpKiBjITiDOXcEMqUMFkJm1BFGZqyaqYEd/3Lm6TbbLiGP97UW82yjipcwCVcgwu30IIHaEMHGMzgGV7hzcqtF+vd+liNVqxy5xz+wPr8ASB3k+o=\n\nCase 5: Oua\n\nia\n\nAAACDXicbVC7TsMwFHV4lvIKMLJYFCSmKkEgEFOlLmwUiT6kJlSO67RWnYfsG0QV5QdY+BUWBhBiZWfjb3DaDNByJFtH59xr33u8WHAFlvVtLCwuLa+sltbK6xubW9vmzm5LRYmkrEkjEcmORxQTPGRN4CBYJ5aMBJ5gbW9Uz/32PZOKR+EtjGPmBmQQcp9TAlrqmYcOsAdI6/oJfHaJMycgMKREpNfZXZqQrJdyfZkVq2pNgOeJXZAKKtDomV9OP6JJwEKggijVta0Y3JRI4FSwrOwkisWEjsiAdTUNScCUm062yfCRVvrYj6Q+IeCJ+rsjJYFS48DTlfmwatbLxf+8bgL+hZvyME6AhXT6kZ8IDBHOo8F9LhkFMdaEUMn1rJgOiSQUdIBlHYI9u/I8aZ1Ubc1vTis1q4ijhPbRATpGNjpHNXSFGqiJKHpEz+gVvRlPxovxbnxMSxeMomcP/YHx+QPxS5wP\nAAACDXicbVC7TsMwFHV4lvIKMLJYFCSmKkEgEFOlLmwUiT6kJlSO67RWnYfsG0QV5QdY+BUWBhBiZWfjb3DaDNByJFtH59xr33u8WHAFlvVtLCwuLa+sltbK6xubW9vmzm5LRYmkrEkjEcmORxQTPGRN4CBYJ5aMBJ5gbW9Uz/32PZOKR+EtjGPmBmQQcp9TAlrqmYcOsAdI6/oJfHaJMycgMKREpNfZXZqQrJdyfZkVq2pNgOeJXZAKKtDomV9OP6JJwEKggijVta0Y3JRI4FSwrOwkisWEjsiAdTUNScCUm062yfCRVvrYj6Q+IeCJ+rsjJYFS48DTlfmwatbLxf+8bgL+hZvyME6AhXT6kZ8IDBHOo8F9LhkFMdaEUMn1rJgOiSQUdIBlHYI9u/I8aZ1Ubc1vTis1q4ijhPbRATpGNjpHNXSFGqiJKHpEz+gVvRlPxovxbnxMSxeMomcP/YHx+QPxS5wP\nAAACDXicbVC7TsMwFHV4lvIKMLJYFCSmKkEgEFOlLmwUiT6kJlSO67RWnYfsG0QV5QdY+BUWBhBiZWfjb3DaDNByJFtH59xr33u8WHAFlvVtLCwuLa+sltbK6xubW9vmzm5LRYmkrEkjEcmORxQTPGRN4CBYJ5aMBJ5gbW9Uz/32PZOKR+EtjGPmBmQQcp9TAlrqmYcOsAdI6/oJfHaJMycgMKREpNfZXZqQrJdyfZkVq2pNgOeJXZAKKtDomV9OP6JJwEKggijVta0Y3JRI4FSwrOwkisWEjsiAdTUNScCUm062yfCRVvrYj6Q+IeCJ+rsjJYFS48DTlfmwatbLxf+8bgL+hZvyME6AhXT6kZ8IDBHOo8F9LhkFMdaEUMn1rJgOiSQUdIBlHYI9u/I8aZ1Ubc1vTis1q4ijhPbRATpGNjpHNXSFGqiJKHpEz+gVvRlPxovxbnxMSxeMomcP/YHx+QPxS5wP\nAAACDXicbVC7TsMwFHV4lvIKMLJYFCSmKkEgEFOlLmwUiT6kJlSO67RWnYfsG0QV5QdY+BUWBhBiZWfjb3DaDNByJFtH59xr33u8WHAFlvVtLCwuLa+sltbK6xubW9vmzm5LRYmkrEkjEcmORxQTPGRN4CBYJ5aMBJ5gbW9Uz/32PZOKR+EtjGPmBmQQcp9TAlrqmYcOsAdI6/oJfHaJMycgMKREpNfZXZqQrJdyfZkVq2pNgOeJXZAKKtDomV9OP6JJwEKggijVta0Y3JRI4FSwrOwkisWEjsiAdTUNScCUm062yfCRVvrYj6Q+IeCJ+rsjJYFS48DTlfmwatbLxf+8bgL+hZvyME6AhXT6kZ8IDBHOo8F9LhkFMdaEUMn1rJgOiSQUdIBlHYI9u/I8aZ1Ubc1vTis1q4ijhPbRATpGNjpHNXSFGqiJKHpEz+gVvRlPxovxbnxMSxeMomcP/YHx+QPxS5wP\n\nInference network\n\nGenerative network\n\nFigure 1: Our model takes the adjacency matrix (A), the attribute matrix (X) and the partial node\nlabels (Yl) as input, and outputs Gaussian distributions as latent embeddings for all nodes and\nattributes (Zn and Za), as well as the latent labels of the unlabelled nodes (Yu). The two neural\nnetwork models, i.e. the inference network and the generative network, are trained by optimizing the\nELBO on the log marginal likelihood of the \ufb01ve cases of observations.\n\nThe partially labelled attributed networks are obviously heterogeneous data, as the nodes and at-\ntributes can be considered as two types of heterogeneous entities, and the adjacency matrix and\nthe attribute matrix can be regarded as the relationships between these two entities. To address the\nproblem, we propose the SCAN, a Semi-supervised Co-embedding model for Attributed Network\nthat semi-supervisedly co-embeds both attributes and nodes in the same semantic space, i.e., learns\nlatent variables/embeddings for both nodes and attributes, allowing for effective generalization of\nclassi\ufb01cation from a small number of labelled data sets to a large number of unlabelled ones. In what\nfollows, we \ufb01rst follow the principle of modelling semi-supervised learning for heterogeneous data\n(Subsection 3.2) to derive an overall lower bound for this problem by splitting the given attributed\nnetwork observations into \ufb01ve types of atomic data points, then provide a solution for addressing the\nbottleneck of the marginalization over all K classes. Fig. 1 provides an overview of the framework\nof our SCAN, and Fig. 2 provides probabilistic graphical perspective of our model.\n\nn\n\nAAAB/HicbVDLSsNAFL2pr1pf1S7dBIvgqiRS0GXBjcsK9gFNLJPJpB06mQkzEyGE+ituXCji1g9x5984abPQ1gPDHM65lzlzgoRRpR3n26psbG5t71R3a3v7B4dH9eOTvhKpxKSHBRNyGCBFGOWkp6lmZJhIguKAkUEwuyn8wSORigp+r7OE+DGacBpRjLSRxvWGFwgWqiw2V+4lUzp/MGrTaTkL2OvELUkTSnTH9S8vFDiNCdeYIaVGrpNoP0dSU8zIvOaliiQIz9CEjAzlKCbKzxfh5/a5UUI7EtIcru2F+nsjR7Eq8pnJGOmpWvUK8T9vlOro2s8pT1JNOF4+FKXM1sIumrBDKgnWLDMEYUlNVhtPkURYm75qpgR39cvrpH/Zcg2/azc77bKOKpzCGVyAC1fQgVvoQg8wZPAMr/BmPVkv1rv1sRytWOVOA/7A+vwBkyyVTw==\nAAAB/HicbVDLSsNAFL2pr1pf1S7dBIvgqiRS0GXBjcsK9gFNLJPJpB06mQkzEyGE+ituXCji1g9x5984abPQ1gPDHM65lzlzgoRRpR3n26psbG5t71R3a3v7B4dH9eOTvhKpxKSHBRNyGCBFGOWkp6lmZJhIguKAkUEwuyn8wSORigp+r7OE+DGacBpRjLSRxvWGFwgWqiw2V+4lUzp/MGrTaTkL2OvELUkTSnTH9S8vFDiNCdeYIaVGrpNoP0dSU8zIvOaliiQIz9CEjAzlKCbKzxfh5/a5UUI7EtIcru2F+nsjR7Eq8pnJGOmpWvUK8T9vlOro2s8pT1JNOF4+FKXM1sIumrBDKgnWLDMEYUlNVhtPkURYm75qpgR39cvrpH/Zcg2/azc77bKOKpzCGVyAC1fQgVvoQg8wZPAMr/BmPVkv1rv1sRytWOVOA/7A+vwBkyyVTw==\nAAAB/HicbVDLSsNAFL2pr1pf1S7dBIvgqiRS0GXBjcsK9gFNLJPJpB06mQkzEyGE+ituXCji1g9x5984abPQ1gPDHM65lzlzgoRRpR3n26psbG5t71R3a3v7B4dH9eOTvhKpxKSHBRNyGCBFGOWkp6lmZJhIguKAkUEwuyn8wSORigp+r7OE+DGacBpRjLSRxvWGFwgWqiw2V+4lUzp/MGrTaTkL2OvELUkTSnTH9S8vFDiNCdeYIaVGrpNoP0dSU8zIvOaliiQIz9CEjAzlKCbKzxfh5/a5UUI7EtIcru2F+nsjR7Eq8pnJGOmpWvUK8T9vlOro2s8pT1JNOF4+FKXM1sIumrBDKgnWLDMEYUlNVhtPkURYm75qpgR39cvrpH/Zcg2/azc77bKOKpzCGVyAC1fQgVvoQg8wZPAMr/BmPVkv1rv1sRytWOVOA/7A+vwBkyyVTw==\nAAAB/HicbVDLSsNAFL2pr1pf1S7dBIvgqiRS0GXBjcsK9gFNLJPJpB06mQkzEyGE+ituXCji1g9x5984abPQ1gPDHM65lzlzgoRRpR3n26psbG5t71R3a3v7B4dH9eOTvhKpxKSHBRNyGCBFGOWkp6lmZJhIguKAkUEwuyn8wSORigp+r7OE+DGacBpRjLSRxvWGFwgWqiw2V+4lUzp/MGrTaTkL2OvELUkTSnTH9S8vFDiNCdeYIaVGrpNoP0dSU8zIvOaliiQIz9CEjAzlKCbKzxfh5/a5UUI7EtIcru2F+nsjR7Eq8pnJGOmpWvUK8T9vlOro2s8pT1JNOF4+FKXM1sIumrBDKgnWLDMEYUlNVhtPkURYm75qpgR39cvrpH/Zcg2/azc77bKOKpzCGVyAC1fQgVvoQg8wZPAMr/BmPVkv1rv1sRytWOVOA/7A+vwBkyyVTw==\n\nZn\nj\n\nAAAB9XicbVDLSgMxFL1TX7W+qi7dBIvgqsxIQZcFNy4r2Ae205JJM21sJhmSjFKG/ocbF4q49V/c+Tdm2llo64HA4Zx7uScniDnTxnW/ncLa+sbmVnG7tLO7t39QPjxqaZkoQptEcqk6AdaUM0GbhhlOO7GiOAo4bQeT68xvP1KlmRR3ZhpTP8IjwUJGsLFSvxdhMw7C9H7WF4OHQbniVt050CrxclKBHI1B+as3lCSJqDCEY627nhsbP8XKMMLprNRLNI0xmeAR7VoqcES1n85Tz9CZVYYolMo+YdBc/b2R4kjraRTYySylXvYy8T+vm5jwyk+ZiBNDBVkcChOOjERZBWjIFCWGTy3BRDGbFZExVpgYW1TJluAtf3mVtC6qnuW3tUq9ltdRhBM4hXPw4BLqcAMNaAIBBc/wCm/Ok/PivDsfi9GCk+8cwx84nz/YKZKt\nAAAB9XicbVDLSgMxFL1TX7W+qi7dBIvgqsxIQZcFNy4r2Ae205JJM21sJhmSjFKG/ocbF4q49V/c+Tdm2llo64HA4Zx7uScniDnTxnW/ncLa+sbmVnG7tLO7t39QPjxqaZkoQptEcqk6AdaUM0GbhhlOO7GiOAo4bQeT68xvP1KlmRR3ZhpTP8IjwUJGsLFSvxdhMw7C9H7WF4OHQbniVt050CrxclKBHI1B+as3lCSJqDCEY627nhsbP8XKMMLprNRLNI0xmeAR7VoqcES1n85Tz9CZVYYolMo+YdBc/b2R4kjraRTYySylXvYy8T+vm5jwyk+ZiBNDBVkcChOOjERZBWjIFCWGTy3BRDGbFZExVpgYW1TJluAtf3mVtC6qnuW3tUq9ltdRhBM4hXPw4BLqcAMNaAIBBc/wCm/Ok/PivDsfi9GCk+8cwx84nz/YKZKt\nAAAB9XicbVDLSgMxFL1TX7W+qi7dBIvgqsxIQZcFNy4r2Ae205JJM21sJhmSjFKG/ocbF4q49V/c+Tdm2llo64HA4Zx7uScniDnTxnW/ncLa+sbmVnG7tLO7t39QPjxqaZkoQptEcqk6AdaUM0GbhhlOO7GiOAo4bQeT68xvP1KlmRR3ZhpTP8IjwUJGsLFSvxdhMw7C9H7WF4OHQbniVt050CrxclKBHI1B+as3lCSJqDCEY627nhsbP8XKMMLprNRLNI0xmeAR7VoqcES1n85Tz9CZVYYolMo+YdBc/b2R4kjraRTYySylXvYy8T+vm5jwyk+ZiBNDBVkcChOOjERZBWjIFCWGTy3BRDGbFZExVpgYW1TJluAtf3mVtC6qnuW3tUq9ltdRhBM4hXPw4BLqcAMNaAIBBc/wCm/Ok/PivDsfi9GCk+8cwx84nz/YKZKt\nAAAB9XicbVDLSgMxFL1TX7W+qi7dBIvgqsxIQZcFNy4r2Ae205JJM21sJhmSjFKG/ocbF4q49V/c+Tdm2llo64HA4Zx7uScniDnTxnW/ncLa+sbmVnG7tLO7t39QPjxqaZkoQptEcqk6AdaUM0GbhhlOO7GiOAo4bQeT68xvP1KlmRR3ZhpTP8IjwUJGsLFSvxdhMw7C9H7WF4OHQbniVt050CrxclKBHI1B+as3lCSJqDCEY627nhsbP8XKMMLprNRLNI0xmeAR7VoqcES1n85Tz9CZVYYolMo+YdBc/b2R4kjraRTYySylXvYy8T+vm5jwyk+ZiBNDBVkcChOOjERZBWjIFCWGTy3BRDGbFZExVpgYW1TJluAtf3mVtC6qnuW3tUq9ltdRhBM4hXPw4BLqcAMNaAIBBc/wCm/Ok/PivDsfi9GCk+8cwx84nz/YKZKt\n\nYl\n\nAAAB83icbVDLSsNAFL3xWeur6tLNYBFclaQIuiy4cVnBPqSJZTKdtEMnkzBzI5TQ33DjQhG3/ow7/8Zpm4W2Hhg4nHMv98wJUykMuu63s7a+sbm1Xdop7+7tHxxWjo7bJsk04y2WyER3Q2q4FIq3UKDk3VRzGoeSd8LxzczvPHFtRKLucZLyIKZDJSLBKFrJ92OKozDKH6aPsl+pujV3DrJKvIJUoUCzX/nyBwnLYq6QSWpMz3NTDHKqUTDJp2U/MzylbEyHvGepojE3QT7PPCXnVhmQKNH2KSRz9fdGTmNjJnFoJ2cZzbI3E//zehlG10EuVJohV2xxKMokwYTMCiADoTlDObGEMi1sVsJGVFOGtqayLcFb/vIqaddrnuV3l9VGvaijBKdwBhfgwRU04Baa0AIGKTzDK7w5mfPivDsfi9E1p9g5gT9wPn8ATiCRyw==\nAAAB83icbVDLSsNAFL3xWeur6tLNYBFclaQIuiy4cVnBPqSJZTKdtEMnkzBzI5TQ33DjQhG3/ow7/8Zpm4W2Hhg4nHMv98wJUykMuu63s7a+sbm1Xdop7+7tHxxWjo7bJsk04y2WyER3Q2q4FIq3UKDk3VRzGoeSd8LxzczvPHFtRKLucZLyIKZDJSLBKFrJ92OKozDKH6aPsl+pujV3DrJKvIJUoUCzX/nyBwnLYq6QSWpMz3NTDHKqUTDJp2U/MzylbEyHvGepojE3QT7PPCXnVhmQKNH2KSRz9fdGTmNjJnFoJ2cZzbI3E//zehlG10EuVJohV2xxKMokwYTMCiADoTlDObGEMi1sVsJGVFOGtqayLcFb/vIqaddrnuV3l9VGvaijBKdwBhfgwRU04Baa0AIGKTzDK7w5mfPivDsfi9E1p9g5gT9wPn8ATiCRyw==\nAAAB83icbVDLSsNAFL3xWeur6tLNYBFclaQIuiy4cVnBPqSJZTKdtEMnkzBzI5TQ33DjQhG3/ow7/8Zpm4W2Hhg4nHMv98wJUykMuu63s7a+sbm1Xdop7+7tHxxWjo7bJsk04y2WyER3Q2q4FIq3UKDk3VRzGoeSd8LxzczvPHFtRKLucZLyIKZDJSLBKFrJ92OKozDKH6aPsl+pujV3DrJKvIJUoUCzX/nyBwnLYq6QSWpMz3NTDHKqUTDJp2U/MzylbEyHvGepojE3QT7PPCXnVhmQKNH2KSRz9fdGTmNjJnFoJ2cZzbI3E//zehlG10EuVJohV2xxKMokwYTMCiADoTlDObGEMi1sVsJGVFOGtqayLcFb/vIqaddrnuV3l9VGvaijBKdwBhfgwRU04Baa0AIGKTzDK7w5mfPivDsfi9E1p9g5gT9wPn8ATiCRyw==\nAAAB83icbVDLSsNAFL3xWeur6tLNYBFclaQIuiy4cVnBPqSJZTKdtEMnkzBzI5TQ33DjQhG3/ow7/8Zpm4W2Hhg4nHMv98wJUykMuu63s7a+sbm1Xdop7+7tHxxWjo7bJsk04y2WyER3Q2q4FIq3UKDk3VRzGoeSd8LxzczvPHFtRKLucZLyIKZDJSLBKFrJ92OKozDKH6aPsl+pujV3DrJKvIJUoUCzX/nyBwnLYq6QSWpMz3NTDHKqUTDJp2U/MzylbEyHvGepojE3QT7PPCXnVhmQKNH2KSRz9fdGTmNjJnFoJ2cZzbI3E//zehlG10EuVJohV2xxKMokwYTMCiADoTlDObGEMi1sVsJGVFOGtqayLcFb/vIqaddrnuV3l9VGvaijBKdwBhfgwRU04Baa0AIGKTzDK7w5mfPivDsfi9E1p9g5gT9wPn8ATiCRyw==\n\nYj\n\nAAAB83icbVDLSsNAFL2pr1pfVZduBovgqiQi6LLgxmUF+5AmlMn0ph07mYSZiVBCf8ONC0Xc+jPu/BunbRbaemDgcM693DMnTAXXxnW/ndLa+sbmVnm7srO7t39QPTxq6yRTDFssEYnqhlSj4BJbhhuB3VQhjUOBnXB8M/M7T6g0T+S9maQYxHQoecQZNVby/ZiaURjlD9P+Y79ac+vuHGSVeAWpQYFmv/rlDxKWxSgNE1TrnuemJsipMpwJnFb8TGNK2ZgOsWeppDHqIJ9nnpIzqwxIlCj7pCFz9fdGTmOtJ3FoJ2cZ9bI3E//zepmJroOcyzQzKNniUJQJYhIyK4AMuEJmxMQSyhS3WQkbUUWZsTVVbAne8pdXSfui7ll+d1lrXBZ1lOEETuEcPLiCBtxCE1rAIIVneIU3J3NenHfnYzFacoqdY/gD5/MHTTeRzA==\nAAAB83icbVDLSsNAFL2pr1pfVZduBovgqiQi6LLgxmUF+5AmlMn0ph07mYSZiVBCf8ONC0Xc+jPu/BunbRbaemDgcM693DMnTAXXxnW/ndLa+sbmVnm7srO7t39QPTxq6yRTDFssEYnqhlSj4BJbhhuB3VQhjUOBnXB8M/M7T6g0T+S9maQYxHQoecQZNVby/ZiaURjlD9P+Y79ac+vuHGSVeAWpQYFmv/rlDxKWxSgNE1TrnuemJsipMpwJnFb8TGNK2ZgOsWeppDHqIJ9nnpIzqwxIlCj7pCFz9fdGTmOtJ3FoJ2cZ9bI3E//zepmJroOcyzQzKNniUJQJYhIyK4AMuEJmxMQSyhS3WQkbUUWZsTVVbAne8pdXSfui7ll+d1lrXBZ1lOEETuEcPLiCBtxCE1rAIIVneIU3J3NenHfnYzFacoqdY/gD5/MHTTeRzA==\nAAAB83icbVDLSsNAFL2pr1pfVZduBovgqiQi6LLgxmUF+5AmlMn0ph07mYSZiVBCf8ONC0Xc+jPu/BunbRbaemDgcM693DMnTAXXxnW/ndLa+sbmVnm7srO7t39QPTxq6yRTDFssEYnqhlSj4BJbhhuB3VQhjUOBnXB8M/M7T6g0T+S9maQYxHQoecQZNVby/ZiaURjlD9P+Y79ac+vuHGSVeAWpQYFmv/rlDxKWxSgNE1TrnuemJsipMpwJnFb8TGNK2ZgOsWeppDHqIJ9nnpIzqwxIlCj7pCFz9fdGTmOtJ3FoJ2cZ9bI3E//zepmJroOcyzQzKNniUJQJYhIyK4AMuEJmxMQSyhS3WQkbUUWZsTVVbAne8pdXSfui7ll+d1lrXBZ1lOEETuEcPLiCBtxCE1rAIIVneIU3J3NenHfnYzFacoqdY/gD5/MHTTeRzA==\nAAAB2XicbZDNSgMxFIXv1L86Vq1rN8EiuCozbnQpuHFZwbZCO5RM5k4bmskMyR2hDH0BF25EfC93vo3pz0JbDwQ+zknIvSculLQUBN9ebWd3b/+gfugfNfzjk9Nmo2fz0gjsilzl5jnmFpXU2CVJCp8LgzyLFfbj6f0i77+gsTLXTzQrMMr4WMtUCk7O6oyaraAdLMW2IVxDC9YaNb+GSS7KDDUJxa0dhEFBUcUNSaFw7g9LiwUXUz7GgUPNM7RRtRxzzi6dk7A0N+5oYkv394uKZ9bOstjdzDhN7Ga2MP/LBiWlt1EldVESarH6KC0Vo5wtdmaJNChIzRxwYaSblYkJN1yQa8Z3HYSbG29D77odOn4MoA7ncAFXEMIN3MEDdKALAhJ4hXdv4r15H6uuat66tDP4I+/zBzjGijg=\nAAAB6HicbZBLSwMxFIXv+Ky1anXrJlgEV2XGjS4FNy4r2Id0hpJJ77SxmcyQh1CG/g03LhTxF7nz35hpu9DWC4GPcxLuyYlzwbXx/W9vY3Nre2e3slfdrx0cHtWPax2dWcWwzTKRqV5MNQousW24EdjLFdI0FtiNJ7el331GpXkmH8w0xyilI8kTzqhxUhim1IzjpHicDZ4G9Ybf9OdD1iFYQgOW0xrUv8JhxmyK0jBBte4Hfm6igirDmcBZNbQac8omdIR9h5KmqKNinnlGzp0yJEmm3JGGzNXfLwqaaj1NY3ezzKhXvVL8z+tbk1xHBZe5NSjZYlFiBTEZKQsgQ66QGTF1QJniLithY6ooM66mqishWP3yOnQum4Hjex8qcApncAEBXMEN3EEL2sAghxd4g3fPeq/ex6KuDW/Z2wn8Ge/zBwk3kHw=\nAAAB6HicbZBLSwMxFIXv+Ky1anXrJlgEV2XGjS4FNy4r2Id0hpJJ77SxmcyQh1CG/g03LhTxF7nz35hpu9DWC4GPcxLuyYlzwbXx/W9vY3Nre2e3slfdrx0cHtWPax2dWcWwzTKRqV5MNQousW24EdjLFdI0FtiNJ7el331GpXkmH8w0xyilI8kTzqhxUhim1IzjpHicDZ4G9Ybf9OdD1iFYQgOW0xrUv8JhxmyK0jBBte4Hfm6igirDmcBZNbQac8omdIR9h5KmqKNinnlGzp0yJEmm3JGGzNXfLwqaaj1NY3ezzKhXvVL8z+tbk1xHBZe5NSjZYlFiBTEZKQsgQ66QGTF1QJniLithY6ooM66mqishWP3yOnQum4Hjex8qcApncAEBXMEN3EEL2sAghxd4g3fPeq/ex6KuDW/Z2wn8Ge/zBwk3kHw=\nAAAB83icbVDLSsNAFL2pr1pfVZduBovgqiQi6LLgxmUF+5AmlMl00o6dTMLMjVBCf8ONC0Xc+jPu/BsnbRfaemDgcM693DMnTKUw6LrfTmltfWNzq7xd2dnd2z+oHh61TZJpxlsskYnuhtRwKRRvoUDJu6nmNA4l74Tjm8LvPHFtRKLucZLyIKZDJSLBKFrJ92OKozDKH6b9x3615tbdGcgq8RakBgs0+9Uvf5CwLOYKmaTG9Dw3xSCnGgWTfFrxM8NTysZ0yHuWKhpzE+SzzFNyZpUBiRJtn0IyU39v5DQ2ZhKHdrLIaJa9QvzP62UYXQe5UGmGXLH5oSiTBBNSFEAGQnOGcmIJZVrYrISNqKYMbU0VW4K3/OVV0r6oe5bfubXG5aKOMpzAKZyDB1fQgFtoQgsYpPAMr/DmZM6L8+58zEdLzmLnGP7A+fwBS/eRyA==\nAAAB83icbVDLSsNAFL2pr1pfVZduBovgqiQi6LLgxmUF+5AmlMn0ph07mYSZiVBCf8ONC0Xc+jPu/BunbRbaemDgcM693DMnTAXXxnW/ndLa+sbmVnm7srO7t39QPTxq6yRTDFssEYnqhlSj4BJbhhuB3VQhjUOBnXB8M/M7T6g0T+S9maQYxHQoecQZNVby/ZiaURjlD9P+Y79ac+vuHGSVeAWpQYFmv/rlDxKWxSgNE1TrnuemJsipMpwJnFb8TGNK2ZgOsWeppDHqIJ9nnpIzqwxIlCj7pCFz9fdGTmOtJ3FoJ2cZ9bI3E//zepmJroOcyzQzKNniUJQJYhIyK4AMuEJmxMQSyhS3WQkbUUWZsTVVbAne8pdXSfui7ll+d1lrXBZ1lOEETuEcPLiCBtxCE1rAIIVneIU3J3NenHfnYzFacoqdY/gD5/MHTTeRzA==\nAAAB83icbVDLSsNAFL2pr1pfVZduBovgqiQi6LLgxmUF+5AmlMn0ph07mYSZiVBCf8ONC0Xc+jPu/BunbRbaemDgcM693DMnTAXXxnW/ndLa+sbmVnm7srO7t39QPTxq6yRTDFssEYnqhlSj4BJbhhuB3VQhjUOBnXB8M/M7T6g0T+S9maQYxHQoecQZNVby/ZiaURjlD9P+Y79ac+vuHGSVeAWpQYFmv/rlDxKWxSgNE1TrnuemJsipMpwJnFb8TGNK2ZgOsWeppDHqIJ9nnpIzqwxIlCj7pCFz9fdGTmOtJ3FoJ2cZ9bI3E//zepmJroOcyzQzKNniUJQJYhIyK4AMuEJmxMQSyhS3WQkbUUWZsTVVbAne8pdXSfui7ll+d1lrXBZ1lOEETuEcPLiCBtxCE1rAIIVneIU3J3NenHfnYzFacoqdY/gD5/MHTTeRzA==\nAAAB83icbVDLSsNAFL2pr1pfVZduBovgqiQi6LLgxmUF+5AmlMn0ph07mYSZiVBCf8ONC0Xc+jPu/BunbRbaemDgcM693DMnTAXXxnW/ndLa+sbmVnm7srO7t39QPTxq6yRTDFssEYnqhlSj4BJbhhuB3VQhjUOBnXB8M/M7T6g0T+S9maQYxHQoecQZNVby/ZiaURjlD9P+Y79ac+vuHGSVeAWpQYFmv/rlDxKWxSgNE1TrnuemJsipMpwJnFb8TGNK2ZgOsWeppDHqIJ9nnpIzqwxIlCj7pCFz9fdGTmOtJ3FoJ2cZ9bI3E//zepmJroOcyzQzKNniUJQJYhIyK4AMuEJmxMQSyhS3WQkbUUWZsTVVbAne8pdXSfui7ll+d1lrXBZ1lOEETuEcPLiCBtxCE1rAIIVneIU3J3NenHfnYzFacoqdY/gD5/MHTTeRzA==\nAAAB83icbVDLSsNAFL2pr1pfVZduBovgqiQi6LLgxmUF+5AmlMn0ph07mYSZiVBCf8ONC0Xc+jPu/BunbRbaemDgcM693DMnTAXXxnW/ndLa+sbmVnm7srO7t39QPTxq6yRTDFssEYnqhlSj4BJbhhuB3VQhjUOBnXB8M/M7T6g0T+S9maQYxHQoecQZNVby/ZiaURjlD9P+Y79ac+vuHGSVeAWpQYFmv/rlDxKWxSgNE1TrnuemJsipMpwJnFb8TGNK2ZgOsWeppDHqIJ9nnpIzqwxIlCj7pCFz9fdGTmOtJ3FoJ2cZ9bI3E//zepmJroOcyzQzKNniUJQJYhIyK4AMuEJmxMQSyhS3WQkbUUWZsTVVbAne8pdXSfui7ll+d1lrXBZ1lOEETuEcPLiCBtxCE1rAIIVneIU3J3NenHfnYzFacoqdY/gD5/MHTTeRzA==\nAAAB83icbVDLSsNAFL2pr1pfVZduBovgqiQi6LLgxmUF+5AmlMn0ph07mYSZiVBCf8ONC0Xc+jPu/BunbRbaemDgcM693DMnTAXXxnW/ndLa+sbmVnm7srO7t39QPTxq6yRTDFssEYnqhlSj4BJbhhuB3VQhjUOBnXB8M/M7T6g0T+S9maQYxHQoecQZNVby/ZiaURjlD9P+Y79ac+vuHGSVeAWpQYFmv/rlDxKWxSgNE1TrnuemJsipMpwJnFb8TGNK2ZgOsWeppDHqIJ9nnpIzqwxIlCj7pCFz9fdGTmOtJ3FoJ2cZ9bI3E//zepmJroOcyzQzKNniUJQJYhIyK4AMuEJmxMQSyhS3WQkbUUWZsTVVbAne8pdXSfui7ll+d1lrXBZ1lOEETuEcPLiCBtxCE1rAIIVneIU3J3NenHfnYzFacoqdY/gD5/MHTTeRzA==\nAAAB83icbVDLSsNAFL2pr1pfVZduBovgqiQi6LLgxmUF+5AmlMn0ph07mYSZiVBCf8ONC0Xc+jPu/BunbRbaemDgcM693DMnTAXXxnW/ndLa+sbmVnm7srO7t39QPTxq6yRTDFssEYnqhlSj4BJbhhuB3VQhjUOBnXB8M/M7T6g0T+S9maQYxHQoecQZNVby/ZiaURjlD9P+Y79ac+vuHGSVeAWpQYFmv/rlDxKWxSgNE1TrnuemJsipMpwJnFb8TGNK2ZgOsWeppDHqIJ9nnpIzqwxIlCj7pCFz9fdGTmOtJ3FoJ2cZ9bI3E//zepmJroOcyzQzKNniUJQJYhIyK4AMuEJmxMQSyhS3WQkbUUWZsTVVbAne8pdXSfui7ll+d1lrXBZ1lOEETuEcPLiCBtxCE1rAIIVneIU3J3NenHfnYzFacoqdY/gD5/MHTTeRzA==\n\nYu\n\nAAAB83icbVDLSsNAFL2pr1pfVZduBovgqiRFqMuCG5cV7EOaWCbTSTt0MgnzEErob7hxoYhbf8adf+OkzUJbDwwczrmXe+aEKWdKu+63U9rY3NreKe9W9vYPDo+qxyddlRhJaIckPJH9ECvKmaAdzTSn/VRSHIec9sLpTe73nqhULBH3epbSIMZjwSJGsLaS78dYT8Ioe5g/mmG15tbdBdA68QpSgwLtYfXLHyXExFRowrFSA89NdZBhqRnhdF7xjaIpJlM8pgNLBY6pCrJF5jm6sMoIRYm0T2i0UH9vZDhWahaHdjLPqFa9XPzPGxgdXQcZE6nRVJDlochwpBOUF4BGTFKi+cwSTCSzWRGZYImJtjVVbAne6pfXSbdR9yy/u6q1GkUdZTiDc7gED5rQgltoQwcIpPAMr/DmGOfFeXc+lqMlp9g5hT9wPn8AW8SR1A==\nAAAB83icbVDLSsNAFL2pr1pfVZduBovgqiRFqMuCG5cV7EOaWCbTSTt0MgnzEErob7hxoYhbf8adf+OkzUJbDwwczrmXe+aEKWdKu+63U9rY3NreKe9W9vYPDo+qxyddlRhJaIckPJH9ECvKmaAdzTSn/VRSHIec9sLpTe73nqhULBH3epbSIMZjwSJGsLaS78dYT8Ioe5g/mmG15tbdBdA68QpSgwLtYfXLHyXExFRowrFSA89NdZBhqRnhdF7xjaIpJlM8pgNLBY6pCrJF5jm6sMoIRYm0T2i0UH9vZDhWahaHdjLPqFa9XPzPGxgdXQcZE6nRVJDlochwpBOUF4BGTFKi+cwSTCSzWRGZYImJtjVVbAne6pfXSbdR9yy/u6q1GkUdZTiDc7gED5rQgltoQwcIpPAMr/DmGOfFeXc+lqMlp9g5hT9wPn8AW8SR1A==\nAAAB83icbVDLSsNAFL2pr1pfVZduBovgqiRFqMuCG5cV7EOaWCbTSTt0MgnzEErob7hxoYhbf8adf+OkzUJbDwwczrmXe+aEKWdKu+63U9rY3NreKe9W9vYPDo+qxyddlRhJaIckPJH9ECvKmaAdzTSn/VRSHIec9sLpTe73nqhULBH3epbSIMZjwSJGsLaS78dYT8Ioe5g/mmG15tbdBdA68QpSgwLtYfXLHyXExFRowrFSA89NdZBhqRnhdF7xjaIpJlM8pgNLBY6pCrJF5jm6sMoIRYm0T2i0UH9vZDhWahaHdjLPqFa9XPzPGxgdXQcZE6nRVJDlochwpBOUF4BGTFKi+cwSTCSzWRGZYImJtjVVbAne6pfXSbdR9yy/u6q1GkUdZTiDc7gED5rQgltoQwcIpPAMr/DmGOfFeXc+lqMlp9g5hT9wPn8AW8SR1A==\nAAAB83icbVDLSsNAFL2pr1pfVZduBovgqiRFqMuCG5cV7EOaWCbTSTt0MgnzEErob7hxoYhbf8adf+OkzUJbDwwczrmXe+aEKWdKu+63U9rY3NreKe9W9vYPDo+qxyddlRhJaIckPJH9ECvKmaAdzTSn/VRSHIec9sLpTe73nqhULBH3epbSIMZjwSJGsLaS78dYT8Ioe5g/mmG15tbdBdA68QpSgwLtYfXLHyXExFRowrFSA89NdZBhqRnhdF7xjaIpJlM8pgNLBY6pCrJF5jm6sMoIRYm0T2i0UH9vZDhWahaHdjLPqFa9XPzPGxgdXQcZE6nRVJDlochwpBOUF4BGTFKi+cwSTCSzWRGZYImJtjVVbAne6pfXSbdR9yy/u6q1GkUdZTiDc7gED5rQgltoQwcIpPAMr/DmGOfFeXc+lqMlp9g5hT9wPn8AW8SR1A==\n\n\u2713e\n\nAAAB/nicbVDLSsNAFJ3UV62vqLhyEyyCq5JIQZcFNy4r2Ac0sUwmN+3QyYOZG6GEgr/ixoUibv0Od/6NkzYLbT0wzOGce5kzx08FV2jb30ZlbX1jc6u6XdvZ3ds/MA+PuirJJIMOS0Qi+z5VIHgMHeQooJ9KoJEvoOdPbgq/9whS8SS+x2kKXkRHMQ85o6iloXni+okI1DTSV+7iGJDOHmBo1u2GPYe1SpyS1EmJ9tD8coOEZRHEyARVauDYKXo5lciZgFnNzRSklE3oCAaaxjQC5eXz+DPrXCuBFSZSnxitufp7I6eRKhLqyYjiWC17hfifN8gwvPZyHqcZQswWD4WZsDCxii6sgEtgKKaaUCa5zmqxMZWUoW6spktwlr+8SrqXDUfzu2a91SzrqJJTckYuiEOuSIvckjbpEEZy8kxeyZvxZLwY78bHYrRilDvH5A+Mzx8jG5Yv\nAAAB/nicbVDLSsNAFJ3UV62vqLhyEyyCq5JIQZcFNy4r2Ac0sUwmN+3QyYOZG6GEgr/ixoUibv0Od/6NkzYLbT0wzOGce5kzx08FV2jb30ZlbX1jc6u6XdvZ3ds/MA+PuirJJIMOS0Qi+z5VIHgMHeQooJ9KoJEvoOdPbgq/9whS8SS+x2kKXkRHMQ85o6iloXni+okI1DTSV+7iGJDOHmBo1u2GPYe1SpyS1EmJ9tD8coOEZRHEyARVauDYKXo5lciZgFnNzRSklE3oCAaaxjQC5eXz+DPrXCuBFSZSnxitufp7I6eRKhLqyYjiWC17hfifN8gwvPZyHqcZQswWD4WZsDCxii6sgEtgKKaaUCa5zmqxMZWUoW6spktwlr+8SrqXDUfzu2a91SzrqJJTckYuiEOuSIvckjbpEEZy8kxeyZvxZLwY78bHYrRilDvH5A+Mzx8jG5Yv\nAAAB/nicbVDLSsNAFJ3UV62vqLhyEyyCq5JIQZcFNy4r2Ac0sUwmN+3QyYOZG6GEgr/ixoUibv0Od/6NkzYLbT0wzOGce5kzx08FV2jb30ZlbX1jc6u6XdvZ3ds/MA+PuirJJIMOS0Qi+z5VIHgMHeQooJ9KoJEvoOdPbgq/9whS8SS+x2kKXkRHMQ85o6iloXni+okI1DTSV+7iGJDOHmBo1u2GPYe1SpyS1EmJ9tD8coOEZRHEyARVauDYKXo5lciZgFnNzRSklE3oCAaaxjQC5eXz+DPrXCuBFSZSnxitufp7I6eRKhLqyYjiWC17hfifN8gwvPZyHqcZQswWD4WZsDCxii6sgEtgKKaaUCa5zmqxMZWUoW6spktwlr+8SrqXDUfzu2a91SzrqJJTckYuiEOuSIvckjbpEEZy8kxeyZvxZLwY78bHYrRilDvH5A+Mzx8jG5Yv\nAAAB/nicbVDLSsNAFJ3UV62vqLhyEyyCq5JIQZcFNy4r2Ac0sUwmN+3QyYOZG6GEgr/ixoUibv0Od/6NkzYLbT0wzOGce5kzx08FV2jb30ZlbX1jc6u6XdvZ3ds/MA+PuirJJIMOS0Qi+z5VIHgMHeQooJ9KoJEvoOdPbgq/9whS8SS+x2kKXkRHMQ85o6iloXni+okI1DTSV+7iGJDOHmBo1u2GPYe1SpyS1EmJ9tD8coOEZRHEyARVauDYKXo5lciZgFnNzRSklE3oCAaaxjQC5eXz+DPrXCuBFSZSnxitufp7I6eRKhLqyYjiWC17hfifN8gwvPZyHqcZQswWD4WZsDCxii6sgEtgKKaaUCa5zmqxMZWUoW6spktwlr+8SrqXDUfzu2a91SzrqJJTckYuiEOuSIvckjbpEEZy8kxeyZvxZLwY78bHYrRilDvH5A+Mzx8jG5Yv\n\nAij\n\nAAAB+HicbVBNS8NAFHypX7V+NOrRy2IRPJVECnqsePFYwbZCG8Jmu2nXbjZhdyPUkF/ixYMiXv0p3vw3btoctHVgYZh5jzc7QcKZ0o7zbVXW1jc2t6rbtZ3dvf26fXDYU3EqCe2SmMfyPsCKciZoVzPN6X0iKY4CTvvB9Lrw+49UKhaLOz1LqBfhsWAhI1gbybfrwwjrSRBmV7mfsYfctxtO05kDrRK3JA0o0fHtr+EoJmlEhSYcKzVwnUR7GZaaEU7z2jBVNMFkisd0YKjAEVVeNg+eo1OjjFAYS/OERnP190aGI6VmUWAmi5hq2SvE/7xBqsNLL2MiSTUVZHEoTDnSMSpaQCMmKdF8ZggmkpmsiEywxESbrmqmBHf5y6ukd950Db9tNdqtso4qHMMJnIELF9CGG+hAFwik8Ayv8GY9WS/Wu/WxGK1Y5c4R/IH1+QM2SJNk\nAAAB+HicbVBNS8NAFHypX7V+NOrRy2IRPJVECnqsePFYwbZCG8Jmu2nXbjZhdyPUkF/ixYMiXv0p3vw3btoctHVgYZh5jzc7QcKZ0o7zbVXW1jc2t6rbtZ3dvf26fXDYU3EqCe2SmMfyPsCKciZoVzPN6X0iKY4CTvvB9Lrw+49UKhaLOz1LqBfhsWAhI1gbybfrwwjrSRBmV7mfsYfctxtO05kDrRK3JA0o0fHtr+EoJmlEhSYcKzVwnUR7GZaaEU7z2jBVNMFkisd0YKjAEVVeNg+eo1OjjFAYS/OERnP190aGI6VmUWAmi5hq2SvE/7xBqsNLL2MiSTUVZHEoTDnSMSpaQCMmKdF8ZggmkpmsiEywxESbrmqmBHf5y6ukd950Db9tNdqtso4qHMMJnIELF9CGG+hAFwik8Ayv8GY9WS/Wu/WxGK1Y5c4R/IH1+QM2SJNk\nAAAB+HicbVBNS8NAFHypX7V+NOrRy2IRPJVECnqsePFYwbZCG8Jmu2nXbjZhdyPUkF/ixYMiXv0p3vw3btoctHVgYZh5jzc7QcKZ0o7zbVXW1jc2t6rbtZ3dvf26fXDYU3EqCe2SmMfyPsCKciZoVzPN6X0iKY4CTvvB9Lrw+49UKhaLOz1LqBfhsWAhI1gbybfrwwjrSRBmV7mfsYfctxtO05kDrRK3JA0o0fHtr+EoJmlEhSYcKzVwnUR7GZaaEU7z2jBVNMFkisd0YKjAEVVeNg+eo1OjjFAYS/OERnP190aGI6VmUWAmi5hq2SvE/7xBqsNLL2MiSTUVZHEoTDnSMSpaQCMmKdF8ZggmkpmsiEywxESbrmqmBHf5y6ukd950Db9tNdqtso4qHMMJnIELF9CGG+hAFwik8Ayv8GY9WS/Wu/WxGK1Y5c4R/IH1+QM2SJNk\nAAAB+HicbVBNS8NAFHypX7V+NOrRy2IRPJVECnqsePFYwbZCG8Jmu2nXbjZhdyPUkF/ixYMiXv0p3vw3btoctHVgYZh5jzc7QcKZ0o7zbVXW1jc2t6rbtZ3dvf26fXDYU3EqCe2SmMfyPsCKciZoVzPN6X0iKY4CTvvB9Lrw+49UKhaLOz1LqBfhsWAhI1gbybfrwwjrSRBmV7mfsYfctxtO05kDrRK3JA0o0fHtr+EoJmlEhSYcKzVwnUR7GZaaEU7z2jBVNMFkisd0YKjAEVVeNg+eo1OjjFAYS/OERnP190aGI6VmUWAmi5hq2SvE/7xBqsNLL2MiSTUVZHEoTDnSMSpaQCMmKdF8ZggmkpmsiEywxESbrmqmBHf5y6ukd950Db9tNdqtso4qHMMJnIELF9CGG+hAFwik8Ayv8GY9WS/Wu/WxGK1Y5c4R/IH1+QM2SJNk\n\nc\n\nAAAB/HicbVDLSsNAFL2pr1pf0S7dBIvgqiRS0GXBjcsK9gFNLJPJtB06mQkzEyGE+ituXCji1g9x5984abPQ1gPDHM65lzlzwoRRpV3326psbG5t71R3a3v7B4dH9vFJT4lUYtLFggk5CJEijHLS1VQzMkgkQXHISD+c3RR+/5FIRQW/11lCghhNOB1TjLSRRnbdDwWLVBabK/eTKZ0/4JHdcJvuAs468UrSgBKdkf3lRwKnMeEaM6TU0HMTHeRIaooZmdf8VJEE4RmakKGhHMVEBfki/Nw5N0rkjIU0h2tnof7eyFGsinxmMkZ6qla9QvzPG6Z6fB3klCepJhwvHxqnzNHCKZpwIioJ1iwzBGFJTVYHT5FEWJu+aqYEb/XL66R32fQMv2s12q2yjiqcwhlcgAdX0IZb6EAXMGTwDK/wZj1ZL9a79bEcrVjlTh3+wPr8AYKAlUQ=\nAAAB/HicbVDLSsNAFL2pr1pf0S7dBIvgqiRS0GXBjcsK9gFNLJPJtB06mQkzEyGE+ituXCji1g9x5984abPQ1gPDHM65lzlzwoRRpV3326psbG5t71R3a3v7B4dH9vFJT4lUYtLFggk5CJEijHLS1VQzMkgkQXHISD+c3RR+/5FIRQW/11lCghhNOB1TjLSRRnbdDwWLVBabK/eTKZ0/4JHdcJvuAs468UrSgBKdkf3lRwKnMeEaM6TU0HMTHeRIaooZmdf8VJEE4RmakKGhHMVEBfki/Nw5N0rkjIU0h2tnof7eyFGsinxmMkZ6qla9QvzPG6Z6fB3klCepJhwvHxqnzNHCKZpwIioJ1iwzBGFJTVYHT5FEWJu+aqYEb/XL66R32fQMv2s12q2yjiqcwhlcgAdX0IZb6EAXMGTwDK/wZj1ZL9a79bEcrVjlTh3+wPr8AYKAlUQ=\nAAAB/HicbVDLSsNAFL2pr1pf0S7dBIvgqiRS0GXBjcsK9gFNLJPJtB06mQkzEyGE+ituXCji1g9x5984abPQ1gPDHM65lzlzwoRRpV3326psbG5t71R3a3v7B4dH9vFJT4lUYtLFggk5CJEijHLS1VQzMkgkQXHISD+c3RR+/5FIRQW/11lCghhNOB1TjLSRRnbdDwWLVBabK/eTKZ0/4JHdcJvuAs468UrSgBKdkf3lRwKnMeEaM6TU0HMTHeRIaooZmdf8VJEE4RmakKGhHMVEBfki/Nw5N0rkjIU0h2tnof7eyFGsinxmMkZ6qla9QvzPG6Z6fB3klCepJhwvHxqnzNHCKZpwIioJ1iwzBGFJTVYHT5FEWJu+aqYEb/XL66R32fQMv2s12q2yjiqcwhlcgAdX0IZb6EAXMGTwDK/wZj1ZL9a79bEcrVjlTh3+wPr8AYKAlUQ=\nAAAB/HicbVDLSsNAFL2pr1pf0S7dBIvgqiRS0GXBjcsK9gFNLJPJtB06mQkzEyGE+ituXCji1g9x5984abPQ1gPDHM65lzlzwoRRpV3326psbG5t71R3a3v7B4dH9vFJT4lUYtLFggk5CJEijHLS1VQzMkgkQXHISD+c3RR+/5FIRQW/11lCghhNOB1TjLSRRnbdDwWLVBabK/eTKZ0/4JHdcJvuAs468UrSgBKdkf3lRwKnMeEaM6TU0HMTHeRIaooZmdf8VJEE4RmakKGhHMVEBfki/Nw5N0rkjIU0h2tnof7eyFGsinxmMkZ6qla9QvzPG6Z6fB3klCepJhwvHxqnzNHCKZpwIioJ1iwzBGFJTVYHT5FEWJu+aqYEb/XL66R32fQMv2s12q2yjiqcwhlcgAdX0IZb6EAXMGTwDK/wZj1ZL9a79bEcrVjlTh3+wPr8AYKAlUQ=\n\nZn\ni\n\nAAAB9XicbVDLSgMxFL1TX7W+qi7dBIvgqsyIoMuCG5cV7AP7IpNm2tBMZkjuKGXof7hxoYhb/8Wdf2OmnYW2HggczrmXe3L8WAqDrvvtFNbWNza3itulnd29/YPy4VHTRIlmvMEiGem2Tw2XQvEGCpS8HWtOQ1/ylj+5yfzWI9dGROoepzHvhXSkRCAYRSv1uyHFsR+kD7O+GohBueJW3TnIKvFyUoEc9UH5qzuMWBJyhUxSYzqeG2MvpRoFk3xW6iaGx5RN6Ih3LFU05KaXzlPPyJlVhiSItH0KyVz9vZHS0Jhp6NvJLKVZ9jLxP6+TYHDdS4WKE+SKLQ4FiSQYkawCMhSaM5RTSyjTwmYlbEw1ZWiLKtkSvOUvr5LmRdWz/O6yUrvM6yjCCZzCOXhwBTW4hTo0gIGGZ3iFN+fJeXHenY/FaMHJd47hD5zPH9alkqw=\nAAAB9XicbVDLSgMxFL1TX7W+qi7dBIvgqsyIoMuCG5cV7AP7IpNm2tBMZkjuKGXof7hxoYhb/8Wdf2OmnYW2HggczrmXe3L8WAqDrvvtFNbWNza3itulnd29/YPy4VHTRIlmvMEiGem2Tw2XQvEGCpS8HWtOQ1/ylj+5yfzWI9dGROoepzHvhXSkRCAYRSv1uyHFsR+kD7O+GohBueJW3TnIKvFyUoEc9UH5qzuMWBJyhUxSYzqeG2MvpRoFk3xW6iaGx5RN6Ih3LFU05KaXzlPPyJlVhiSItH0KyVz9vZHS0Jhp6NvJLKVZ9jLxP6+TYHDdS4WKE+SKLQ4FiSQYkawCMhSaM5RTSyjTwmYlbEw1ZWiLKtkSvOUvr5LmRdWz/O6yUrvM6yjCCZzCOXhwBTW4hTo0gIGGZ3iFN+fJeXHenY/FaMHJd47hD5zPH9alkqw=\nAAAB9XicbVDLSgMxFL1TX7W+qi7dBIvgqsyIoMuCG5cV7AP7IpNm2tBMZkjuKGXof7hxoYhb/8Wdf2OmnYW2HggczrmXe3L8WAqDrvvtFNbWNza3itulnd29/YPy4VHTRIlmvMEiGem2Tw2XQvEGCpS8HWtOQ1/ylj+5yfzWI9dGROoepzHvhXSkRCAYRSv1uyHFsR+kD7O+GohBueJW3TnIKvFyUoEc9UH5qzuMWBJyhUxSYzqeG2MvpRoFk3xW6iaGx5RN6Ih3LFU05KaXzlPPyJlVhiSItH0KyVz9vZHS0Jhp6NvJLKVZ9jLxP6+TYHDdS4WKE+SKLQ4FiSQYkawCMhSaM5RTSyjTwmYlbEw1ZWiLKtkSvOUvr5LmRdWz/O6yUrvM6yjCCZzCOXhwBTW4hTo0gIGGZ3iFN+fJeXHenY/FaMHJd47hD5zPH9alkqw=\nAAAB9XicbVDLSgMxFL1TX7W+qi7dBIvgqsyIoMuCG5cV7AP7IpNm2tBMZkjuKGXof7hxoYhb/8Wdf2OmnYW2HggczrmXe3L8WAqDrvvtFNbWNza3itulnd29/YPy4VHTRIlmvMEiGem2Tw2XQvEGCpS8HWtOQ1/ylj+5yfzWI9dGROoepzHvhXSkRCAYRSv1uyHFsR+kD7O+GohBueJW3TnIKvFyUoEc9UH5qzuMWBJyhUxSYzqeG2MvpRoFk3xW6iaGx5RN6Ih3LFU05KaXzlPPyJlVhiSItH0KyVz9vZHS0Jhp6NvJLKVZ9jLxP6+TYHDdS4WKE+SKLQ4FiSQYkawCMhSaM5RTSyjTwmYlbEw1ZWiLKtkSvOUvr5LmRdWz/O6yUrvM6yjCCZzCOXhwBTW4hTo0gIGGZ3iFN+fJeXHenY/FaMHJd47hD5zPH9alkqw=\n\nYl\n\nAAAB83icbVDLSsNAFL3xWeur6tLNYBFclaQIuiy4cVnBPqSJZTKdtEMnkzBzI5TQ33DjQhG3/ow7/8Zpm4W2Hhg4nHMv98wJUykMuu63s7a+sbm1Xdop7+7tHxxWjo7bJsk04y2WyER3Q2q4FIq3UKDk3VRzGoeSd8LxzczvPHFtRKLucZLyIKZDJSLBKFrJ92OKozDKH6aPsl+pujV3DrJKvIJUoUCzX/nyBwnLYq6QSWpMz3NTDHKqUTDJp2U/MzylbEyHvGepojE3QT7PPCXnVhmQKNH2KSRz9fdGTmNjJnFoJ2cZzbI3E//zehlG10EuVJohV2xxKMokwYTMCiADoTlDObGEMi1sVsJGVFOGtqayLcFb/vIqaddrnuV3l9VGvaijBKdwBhfgwRU04Baa0AIGKTzDK7w5mfPivDsfi9E1p9g5gT9wPn8ATiCRyw==\nAAAB83icbVDLSsNAFL3xWeur6tLNYBFclaQIuiy4cVnBPqSJZTKdtEMnkzBzI5TQ33DjQhG3/ow7/8Zpm4W2Hhg4nHMv98wJUykMuu63s7a+sbm1Xdop7+7tHxxWjo7bJsk04y2WyER3Q2q4FIq3UKDk3VRzGoeSd8LxzczvPHFtRKLucZLyIKZDJSLBKFrJ92OKozDKH6aPsl+pujV3DrJKvIJUoUCzX/nyBwnLYq6QSWpMz3NTDHKqUTDJp2U/MzylbEyHvGepojE3QT7PPCXnVhmQKNH2KSRz9fdGTmNjJnFoJ2cZzbI3E//zehlG10EuVJohV2xxKMokwYTMCiADoTlDObGEMi1sVsJGVFOGtqayLcFb/vIqaddrnuV3l9VGvaijBKdwBhfgwRU04Baa0AIGKTzDK7w5mfPivDsfi9E1p9g5gT9wPn8ATiCRyw==\nAAAB83icbVDLSsNAFL3xWeur6tLNYBFclaQIuiy4cVnBPqSJZTKdtEMnkzBzI5TQ33DjQhG3/ow7/8Zpm4W2Hhg4nHMv98wJUykMuu63s7a+sbm1Xdop7+7tHxxWjo7bJsk04y2WyER3Q2q4FIq3UKDk3VRzGoeSd8LxzczvPHFtRKLucZLyIKZDJSLBKFrJ92OKozDKH6aPsl+pujV3DrJKvIJUoUCzX/nyBwnLYq6QSWpMz3NTDHKqUTDJp2U/MzylbEyHvGepojE3QT7PPCXnVhmQKNH2KSRz9fdGTmNjJnFoJ2cZzbI3E//zehlG10EuVJohV2xxKMokwYTMCiADoTlDObGEMi1sVsJGVFOGtqayLcFb/vIqaddrnuV3l9VGvaijBKdwBhfgwRU04Baa0AIGKTzDK7w5mfPivDsfi9E1p9g5gT9wPn8ATiCRyw==\nAAAB83icbVDLSsNAFL3xWeur6tLNYBFclaQIuiy4cVnBPqSJZTKdtEMnkzBzI5TQ33DjQhG3/ow7/8Zpm4W2Hhg4nHMv98wJUykMuu63s7a+sbm1Xdop7+7tHxxWjo7bJsk04y2WyER3Q2q4FIq3UKDk3VRzGoeSd8LxzczvPHFtRKLucZLyIKZDJSLBKFrJ92OKozDKH6aPsl+pujV3DrJKvIJUoUCzX/nyBwnLYq6QSWpMz3NTDHKqUTDJp2U/MzylbEyHvGepojE3QT7PPCXnVhmQKNH2KSRz9fdGTmNjJnFoJ2cZzbI3E//zehlG10EuVJohV2xxKMokwYTMCiADoTlDObGEMi1sVsJGVFOGtqayLcFb/vIqaddrnuV3l9VGvaijBKdwBhfgwRU04Baa0AIGKTzDK7w5mfPivDsfi9E1p9g5gT9wPn8ATiCRyw==\n\nYi\n\nAAAB83icbVDLSsNAFL2pr1pfVZduBovgqiRS0GXBjcsK9iFNKZPpTTt0MgkzE6GE/oYbF4q49Wfc+TdO2iy09cDA4Zx7uWdOkAiujet+O6WNza3tnfJuZW//4PCoenzS0XGqGLZZLGLVC6hGwSW2DTcCe4lCGgUCu8H0Nve7T6g0j+WDmSU4iOhY8pAzaqzk+xE1kyDMHudDPqzW3Lq7AFknXkFqUKA1rH75o5ilEUrDBNW677mJGWRUGc4Ezit+qjGhbErH2LdU0gj1IFtknpMLq4xIGCv7pCEL9fdGRiOtZ1FgJ/OMetXLxf+8fmrCm0HGZZIalGx5KEwFMTHJCyAjrpAZMbOEMsVtVsImVFFmbE0VW4K3+uV10rmqe5bfN2rNRlFHGc7gHC7Bg2towh20oA0MEniGV3hzUufFeXc+lqMlp9g5hT9wPn8AS7ORyw==\nAAAB83icbVDLSsNAFL2pr1pfVZduBovgqiRS0GXBjcsK9iFNKZPpTTt0MgkzE6GE/oYbF4q49Wfc+TdO2iy09cDA4Zx7uWdOkAiujet+O6WNza3tnfJuZW//4PCoenzS0XGqGLZZLGLVC6hGwSW2DTcCe4lCGgUCu8H0Nve7T6g0j+WDmSU4iOhY8pAzaqzk+xE1kyDMHudDPqzW3Lq7AFknXkFqUKA1rH75o5ilEUrDBNW677mJGWRUGc4Ezit+qjGhbErH2LdU0gj1IFtknpMLq4xIGCv7pCEL9fdGRiOtZ1FgJ/OMetXLxf+8fmrCm0HGZZIalGx5KEwFMTHJCyAjrpAZMbOEMsVtVsImVFFmbE0VW4K3+uV10rmqe5bfN2rNRlFHGc7gHC7Bg2towh20oA0MEniGV3hzUufFeXc+lqMlp9g5hT9wPn8AS7ORyw==\nAAAB83icbVDLSsNAFL2pr1pfVZduBovgqiRS0GXBjcsK9iFNKZPpTTt0MgkzE6GE/oYbF4q49Wfc+TdO2iy09cDA4Zx7uWdOkAiujet+O6WNza3tnfJuZW//4PCoenzS0XGqGLZZLGLVC6hGwSW2DTcCe4lCGgUCu8H0Nve7T6g0j+WDmSU4iOhY8pAzaqzk+xE1kyDMHudDPqzW3Lq7AFknXkFqUKA1rH75o5ilEUrDBNW677mJGWRUGc4Ezit+qjGhbErH2LdU0gj1IFtknpMLq4xIGCv7pCEL9fdGRiOtZ1FgJ/OMetXLxf+8fmrCm0HGZZIalGx5KEwFMTHJCyAjrpAZMbOEMsVtVsImVFFmbE0VW4K3+uV10rmqe5bfN2rNRlFHGc7gHC7Bg2towh20oA0MEniGV3hzUufFeXc+lqMlp9g5hT9wPn8AS7ORyw==\nAAAB83icbVDLSsNAFL2pr1pfVZduBovgqiRS0GXBjcsK9iFNKZPpTTt0MgkzE6GE/oYbF4q49Wfc+TdO2iy09cDA4Zx7uWdOkAiujet+O6WNza3tnfJuZW//4PCoenzS0XGqGLZZLGLVC6hGwSW2DTcCe4lCGgUCu8H0Nve7T6g0j+WDmSU4iOhY8pAzaqzk+xE1kyDMHudDPqzW3Lq7AFknXkFqUKA1rH75o5ilEUrDBNW677mJGWRUGc4Ezit+qjGhbErH2LdU0gj1IFtknpMLq4xIGCv7pCEL9fdGRiOtZ1FgJ/OMetXLxf+8fmrCm0HGZZIalGx5KEwFMTHJCyAjrpAZMbOEMsVtVsImVFFmbE0VW4K3+uV10rmqe5bfN2rNRlFHGc7gHC7Bg2towh20oA0MEniGV3hzUufFeXc+lqMlp9g5hT9wPn8AS7ORyw==\n\nZa\naAAAB9XicbVDLSgMxFL1TX7W+qi7dBIvgqsyIoMuCG5cV7AP74k6aaUMzmSHJKGXof7hxoYhb/8Wdf2OmnYW2HggczrmXe3L8WHBtXPfbKaytb2xuFbdLO7t7+wflw6OmjhJFWYNGIlJtHzUTXLKG4UawdqwYhr5gLX9yk/mtR6Y0j+S9mcasF+JI8oBTNFbqd0M0Yz9IH2Z9HOCgXHGr7hxklXg5qUCO+qD81R1GNAmZNFSg1h3PjU0vRWU4FWxW6iaaxUgnOGIdSyWGTPfSeeoZObPKkASRsk8aMld/b6QYaj0NfTuZpdTLXib+53USE1z3Ui7jxDBJF4eCRBATkawCMuSKUSOmliBV3GYldIwKqbFFlWwJ3vKXV0nzoupZfndZqV3mdRThBE7hHDy4ghrcQh0aQEHBM7zCm/PkvDjvzsditODkO8fwB87nD7a3kpc=\n\nAAAB9XicbVDLSgMxFL1TX7W+qi7dBIvgqsyIoMuCG5cV7AP74k6aaUMzmSHJKGXof7hxoYhb/8Wdf2OmnYW2HggczrmXe3L8WHBtXPfbKaytb2xuFbdLO7t7+wflw6OmjhJFWYNGIlJtHzUTXLKG4UawdqwYhr5gLX9yk/mtR6Y0j+S9mcasF+JI8oBTNFbqd0M0Yz9IH2Z9HOCgXHGr7hxklXg5qUCO+qD81R1GNAmZNFSg1h3PjU0vRWU4FWxW6iaaxUgnOGIdSyWGTPfSeeoZObPKkASRsk8aMld/b6QYaj0NfTuZpdTLXib+53USE1z3Ui7jxDBJF4eCRBATkawCMuSKUSOmliBV3GYldIwKqbFFlWwJ3vKXV0nzoupZfndZqV3mdRThBE7hHDy4ghrcQh0aQEHBM7zCm/PkvDjvzsditODkO8fwB87nD7a3kpc=\nAAAB9XicbVDLSgMxFL1TX7W+qi7dBIvgqsyIoMuCG5cV7AP74k6aaUMzmSHJKGXof7hxoYhb/8Wdf2OmnYW2HggczrmXe3L8WHBtXPfbKaytb2xuFbdLO7t7+wflw6OmjhJFWYNGIlJtHzUTXLKG4UawdqwYhr5gLX9yk/mtR6Y0j+S9mcasF+JI8oBTNFbqd0M0Yz9IH2Z9HOCgXHGr7hxklXg5qUCO+qD81R1GNAmZNFSg1h3PjU0vRWU4FWxW6iaaxUgnOGIdSyWGTPfSeeoZObPKkASRsk8aMld/b6QYaj0NfTuZpdTLXib+53USE1z3Ui7jxDBJF4eCRBATkawCMuSKUSOmliBV3GYldIwKqbFFlWwJ3vKXV0nzoupZfndZqV3mdRThBE7hHDy4ghrcQh0aQEHBM7zCm/PkvDjvzsditODkO8fwB87nD7a3kpc=\nAAAB2XicbZDNSgMxFIXv1L86Vq1rN8EiuCozbnQpuHFZwbZCO5RM5k4bmskMyR2hDH0BF25EfC93vo3pz0JbDwQ+zknIvSculLQUBN9ebWd3b/+gfugfNfzjk9Nmo2fz0gjsilzl5jnmFpXU2CVJCp8LgzyLFfbj6f0i77+gsTLXTzQrMMr4WMtUCk7O6oyaraAdLMW2IVxDC9YaNb+GSS7KDDUJxa0dhEFBUcUNSaFw7g9LiwUXUz7GgUPNM7RRtRxzzi6dk7A0N+5oYkv394uKZ9bOstjdzDhN7Ga2MP/LBiWlt1EldVESarH6KC0Vo5wtdmaJNChIzRxwYaSblYkJN1yQa8Z3HYSbG29D77odOn4MoA7ncAFXEMIN3MEDdKALAhJ4hXdv4r15H6uuat66tDP4I+/zBzjGijg=\nAAAB6nicbZDNSgMxFIXv+Ftr1erWTbAIrsqMG10KblxWsD/YTsudNNOGJpkhyShl6Hu4caGID+TOtzHTdqGtFwIf5yTckxOlghvr+9/exubW9s5uaa+8Xzk4PKoeV1omyTRlTZqIRHciNExwxZqWW8E6qWYoI8Ha0eS28NtPTBueqAc7TVkocaR4zClaJ/V7Eu04ivPHWR8HOKjW/Lo/H7IOwRJqsJzGoPrVGyY0k0xZKtCYbuCnNsxRW04Fm5V7mWEp0gmOWNehQslMmM9Tz8i5U4YkTrQ7ypK5+vtFjtKYqYzczSKlWfUK8T+vm9n4Osy5SjPLFF0sijNBbEKKCsiQa0atmDpAqrnLSugYNVLriiq7EoLVL69D67IeOL73oQSncAYXEMAV3MAdNKAJFDS8wBu8e8/eq/exqGvDW/Z2An/G+/wBalqRRQ==\nAAAB6nicbZDNSgMxFIXv+Ftr1erWTbAIrsqMG10KblxWsD/YTsudNNOGJpkhyShl6Hu4caGID+TOtzHTdqGtFwIf5yTckxOlghvr+9/exubW9s5uaa+8Xzk4PKoeV1omyTRlTZqIRHciNExwxZqWW8E6qWYoI8Ha0eS28NtPTBueqAc7TVkocaR4zClaJ/V7Eu04ivPHWR8HOKjW/Lo/H7IOwRJqsJzGoPrVGyY0k0xZKtCYbuCnNsxRW04Fm5V7mWEp0gmOWNehQslMmM9Tz8i5U4YkTrQ7ypK5+vtFjtKYqYzczSKlWfUK8T+vm9n4Osy5SjPLFF0sijNBbEKKCsiQa0atmDpAqrnLSugYNVLriiq7EoLVL69D67IeOL73oQSncAYXEMAV3MAdNKAJFDS8wBu8e8/eq/exqGvDW/Z2An/G+/wBalqRRQ==\nAAAB9XicbVDLSgMxFL1TX7W+qi7dBIvgqsyIoMuCG5cV7APbabmTZtrQTGZIMkoZ+h9uXCji1n9x59+YtrPQ1gOBwzn3ck9OkAiujet+O4W19Y3NreJ2aWd3b/+gfHjU1HGqKGvQWMSqHaBmgkvWMNwI1k4UwygQrBWMb2Z+65EpzWN5byYJ8yMcSh5yisZKvW6EZhSE2cO0h33slytu1Z2DrBIvJxXIUe+Xv7qDmKYRk4YK1LrjuYnxM1SGU8GmpW6qWYJ0jEPWsVRixLSfzVNPyZlVBiSMlX3SkLn6eyPDSOtJFNjJWUq97M3E/7xOasJrP+MySQ2TdHEoTAUxMZlVQAZcMWrExBKkitushI5QITW2qJItwVv+8ippXlQ9y+/cSu0yr6MIJ3AK5+DBFdTgFurQAAoKnuEV3pwn58V5dz4WowUn3zmGP3A+fwC1d5KT\nAAAB9XicbVDLSgMxFL1TX7W+qi7dBIvgqsyIoMuCG5cV7AP74k6aaUMzmSHJKGXof7hxoYhb/8Wdf2OmnYW2HggczrmXe3L8WHBtXPfbKaytb2xuFbdLO7t7+wflw6OmjhJFWYNGIlJtHzUTXLKG4UawdqwYhr5gLX9yk/mtR6Y0j+S9mcasF+JI8oBTNFbqd0M0Yz9IH2Z9HOCgXHGr7hxklXg5qUCO+qD81R1GNAmZNFSg1h3PjU0vRWU4FWxW6iaaxUgnOGIdSyWGTPfSeeoZObPKkASRsk8aMld/b6QYaj0NfTuZpdTLXib+53USE1z3Ui7jxDBJF4eCRBATkawCMuSKUSOmliBV3GYldIwKqbFFlWwJ3vKXV0nzoupZfndZqV3mdRThBE7hHDy4ghrcQh0aQEHBM7zCm/PkvDjvzsditODkO8fwB87nD7a3kpc=\nAAAB9XicbVDLSgMxFL1TX7W+qi7dBIvgqsyIoMuCG5cV7AP74k6aaUMzmSHJKGXof7hxoYhb/8Wdf2OmnYW2HggczrmXe3L8WHBtXPfbKaytb2xuFbdLO7t7+wflw6OmjhJFWYNGIlJtHzUTXLKG4UawdqwYhr5gLX9yk/mtR6Y0j+S9mcasF+JI8oBTNFbqd0M0Yz9IH2Z9HOCgXHGr7hxklXg5qUCO+qD81R1GNAmZNFSg1h3PjU0vRWU4FWxW6iaaxUgnOGIdSyWGTPfSeeoZObPKkASRsk8aMld/b6QYaj0NfTuZpdTLXib+53USE1z3Ui7jxDBJF4eCRBATkawCMuSKUSOmliBV3GYldIwKqbFFlWwJ3vKXV0nzoupZfndZqV3mdRThBE7hHDy4ghrcQh0aQEHBM7zCm/PkvDjvzsditODkO8fwB87nD7a3kpc=\nAAAB9XicbVDLSgMxFL1TX7W+qi7dBIvgqsyIoMuCG5cV7AP74k6aaUMzmSHJKGXof7hxoYhb/8Wdf2OmnYW2HggczrmXe3L8WHBtXPfbKaytb2xuFbdLO7t7+wflw6OmjhJFWYNGIlJtHzUTXLKG4UawdqwYhr5gLX9yk/mtR6Y0j+S9mcasF+JI8oBTNFbqd0M0Yz9IH2Z9HOCgXHGr7hxklXg5qUCO+qD81R1GNAmZNFSg1h3PjU0vRWU4FWxW6iaaxUgnOGIdSyWGTPfSeeoZObPKkASRsk8aMld/b6QYaj0NfTuZpdTLXib+53USE1z3Ui7jxDBJF4eCRBATkawCMuSKUSOmliBV3GYldIwKqbFFlWwJ3vKXV0nzoupZfndZqV3mdRThBE7hHDy4ghrcQh0aQEHBM7zCm/PkvDjvzsditODkO8fwB87nD7a3kpc=\nAAAB9XicbVDLSgMxFL1TX7W+qi7dBIvgqsyIoMuCG5cV7AP74k6aaUMzmSHJKGXof7hxoYhb/8Wdf2OmnYW2HggczrmXe3L8WHBtXPfbKaytb2xuFbdLO7t7+wflw6OmjhJFWYNGIlJtHzUTXLKG4UawdqwYhr5gLX9yk/mtR6Y0j+S9mcasF+JI8oBTNFbqd0M0Yz9IH2Z9HOCgXHGr7hxklXg5qUCO+qD81R1GNAmZNFSg1h3PjU0vRWU4FWxW6iaaxUgnOGIdSyWGTPfSeeoZObPKkASRsk8aMld/b6QYaj0NfTuZpdTLXib+53USE1z3Ui7jxDBJF4eCRBATkawCMuSKUSOmliBV3GYldIwKqbFFlWwJ3vKXV0nzoupZfndZqV3mdRThBE7hHDy4ghrcQh0aQEHBM7zCm/PkvDjvzsditODkO8fwB87nD7a3kpc=\nAAAB9XicbVDLSgMxFL1TX7W+qi7dBIvgqsyIoMuCG5cV7AP74k6aaUMzmSHJKGXof7hxoYhb/8Wdf2OmnYW2HggczrmXe3L8WHBtXPfbKaytb2xuFbdLO7t7+wflw6OmjhJFWYNGIlJtHzUTXLKG4UawdqwYhr5gLX9yk/mtR6Y0j+S9mcasF+JI8oBTNFbqd0M0Yz9IH2Z9HOCgXHGr7hxklXg5qUCO+qD81R1GNAmZNFSg1h3PjU0vRWU4FWxW6iaaxUgnOGIdSyWGTPfSeeoZObPKkASRsk8aMld/b6QYaj0NfTuZpdTLXib+53USE1z3Ui7jxDBJF4eCRBATkawCMuSKUSOmliBV3GYldIwKqbFFlWwJ3vKXV0nzoupZfndZqV3mdRThBE7hHDy4ghrcQh0aQEHBM7zCm/PkvDjvzsditODkO8fwB87nD7a3kpc=\nAAAB9XicbVDLSgMxFL1TX7W+qi7dBIvgqsyIoMuCG5cV7AP74k6aaUMzmSHJKGXof7hxoYhb/8Wdf2OmnYW2HggczrmXe3L8WHBtXPfbKaytb2xuFbdLO7t7+wflw6OmjhJFWYNGIlJtHzUTXLKG4UawdqwYhr5gLX9yk/mtR6Y0j+S9mcasF+JI8oBTNFbqd0M0Yz9IH2Z9HOCgXHGr7hxklXg5qUCO+qD81R1GNAmZNFSg1h3PjU0vRWU4FWxW6iaaxUgnOGIdSyWGTPfSeeoZObPKkASRsk8aMld/b6QYaj0NfTuZpdTLXib+53USE1z3Ui7jxDBJF4eCRBATkawCMuSKUSOmliBV3GYldIwKqbFFlWwJ3vKXV0nzoupZfndZqV3mdRThBE7hHDy4ghrcQh0aQEHBM7zCm/PkvDjvzsditODkO8fwB87nD7a3kpc=\n\nYu\n\nAAAB83icbVDLSsNAFL2pr1pfVZduBovgqiRFqMuCG5cV7EOaWCbTSTt0MgnzEErob7hxoYhbf8adf+OkzUJbDwwczrmXe+aEKWdKu+63U9rY3NreKe9W9vYPDo+qxyddlRhJaIckPJH9ECvKmaAdzTSn/VRSHIec9sLpTe73nqhULBH3epbSIMZjwSJGsLaS78dYT8Ioe5g/mmG15tbdBdA68QpSgwLtYfXLHyXExFRowrFSA89NdZBhqRnhdF7xjaIpJlM8pgNLBY6pCrJF5jm6sMoIRYm0T2i0UH9vZDhWahaHdjLPqFa9XPzPGxgdXQcZE6nRVJDlochwpBOUF4BGTFKi+cwSTCSzWRGZYImJtjVVbAne6pfXSbdR9yy/u6q1GkUdZTiDc7gED5rQgltoQwcIpPAMr/DmGOfFeXc+lqMlp9g5hT9wPn8AW8SR1A==\nAAAB83icbVDLSsNAFL2pr1pfVZduBovgqiRFqMuCG5cV7EOaWCbTSTt0MgnzEErob7hxoYhbf8adf+OkzUJbDwwczrmXe+aEKWdKu+63U9rY3NreKe9W9vYPDo+qxyddlRhJaIckPJH9ECvKmaAdzTSn/VRSHIec9sLpTe73nqhULBH3epbSIMZjwSJGsLaS78dYT8Ioe5g/mmG15tbdBdA68QpSgwLtYfXLHyXExFRowrFSA89NdZBhqRnhdF7xjaIpJlM8pgNLBY6pCrJF5jm6sMoIRYm0T2i0UH9vZDhWahaHdjLPqFa9XPzPGxgdXQcZE6nRVJDlochwpBOUF4BGTFKi+cwSTCSzWRGZYImJtjVVbAne6pfXSbdR9yy/u6q1GkUdZTiDc7gED5rQgltoQwcIpPAMr/DmGOfFeXc+lqMlp9g5hT9wPn8AW8SR1A==\nAAAB83icbVDLSsNAFL2pr1pfVZduBovgqiRFqMuCG5cV7EOaWCbTSTt0MgnzEErob7hxoYhbf8adf+OkzUJbDwwczrmXe+aEKWdKu+63U9rY3NreKe9W9vYPDo+qxyddlRhJaIckPJH9ECvKmaAdzTSn/VRSHIec9sLpTe73nqhULBH3epbSIMZjwSJGsLaS78dYT8Ioe5g/mmG15tbdBdA68QpSgwLtYfXLHyXExFRowrFSA89NdZBhqRnhdF7xjaIpJlM8pgNLBY6pCrJF5jm6sMoIRYm0T2i0UH9vZDhWahaHdjLPqFa9XPzPGxgdXQcZE6nRVJDlochwpBOUF4BGTFKi+cwSTCSzWRGZYImJtjVVbAne6pfXSbdR9yy/u6q1GkUdZTiDc7gED5rQgltoQwcIpPAMr/DmGOfFeXc+lqMlp9g5hT9wPn8AW8SR1A==\nAAAB83icbVDLSsNAFL2pr1pfVZduBovgqiRFqMuCG5cV7EOaWCbTSTt0MgnzEErob7hxoYhbf8adf+OkzUJbDwwczrmXe+aEKWdKu+63U9rY3NreKe9W9vYPDo+qxyddlRhJaIckPJH9ECvKmaAdzTSn/VRSHIec9sLpTe73nqhULBH3epbSIMZjwSJGsLaS78dYT8Ioe5g/mmG15tbdBdA68QpSgwLtYfXLHyXExFRowrFSA89NdZBhqRnhdF7xjaIpJlM8pgNLBY6pCrJF5jm6sMoIRYm0T2i0UH9vZDhWahaHdjLPqFa9XPzPGxgdXQcZE6nRVJDlochwpBOUF4BGTFKi+cwSTCSzWRGZYImJtjVVbAne6pfXSbdR9yy/u6q1GkUdZTiDc7gED5rQgltoQwcIpPAMr/DmGOfFeXc+lqMlp9g5hT9wPn8AW8SR1A==\n\n\u2713a\n\nAAAB/nicbVDLSsNAFJ3UV62vqLhyM1gEVyWRgi4LblxWsA9oYplMJu3QySTM3AglFPwVNy4Ucet3uPNvnLRZaOuBYQ7n3MucOUEquAbH+bYqa+sbm1vV7drO7t7+gX141NVJpijr0EQkqh8QzQSXrAMcBOunipE4EKwXTG4Kv/fIlOaJvIdpyvyYjCSPOCVgpKF94gWJCPU0NlfuwZgBmT2QoV13Gs4ceJW4JamjEu2h/eWFCc1iJoEKovXAdVLwc6KAU8FmNS/TLCV0QkZsYKgkMdN+Po8/w+dGCXGUKHMk4Ln6eyMnsS4SmsmYwFgve4X4nzfIILr2cy7TDJiki4eiTGBIcNEFDrliFMTUEEIVN1kxHRNFKJjGaqYEd/nLq6R72XANv2vWW82yjio6RWfoArnoCrXQLWqjDqIoR8/oFb1ZT9aL9W59LEYrVrlzjP7A+vwBHQuWKw==\nAAAB/nicbVDLSsNAFJ3UV62vqLhyM1gEVyWRgi4LblxWsA9oYplMJu3QySTM3AglFPwVNy4Ucet3uPNvnLRZaOuBYQ7n3MucOUEquAbH+bYqa+sbm1vV7drO7t7+gX141NVJpijr0EQkqh8QzQSXrAMcBOunipE4EKwXTG4Kv/fIlOaJvIdpyvyYjCSPOCVgpKF94gWJCPU0NlfuwZgBmT2QoV13Gs4ceJW4JamjEu2h/eWFCc1iJoEKovXAdVLwc6KAU8FmNS/TLCV0QkZsYKgkMdN+Po8/w+dGCXGUKHMk4Ln6eyMnsS4SmsmYwFgve4X4nzfIILr2cy7TDJiki4eiTGBIcNEFDrliFMTUEEIVN1kxHRNFKJjGaqYEd/nLq6R72XANv2vWW82yjio6RWfoArnoCrXQLWqjDqIoR8/oFb1ZT9aL9W59LEYrVrlzjP7A+vwBHQuWKw==\nAAAB/nicbVDLSsNAFJ3UV62vqLhyM1gEVyWRgi4LblxWsA9oYplMJu3QySTM3AglFPwVNy4Ucet3uPNvnLRZaOuBYQ7n3MucOUEquAbH+bYqa+sbm1vV7drO7t7+gX141NVJpijr0EQkqh8QzQSXrAMcBOunipE4EKwXTG4Kv/fIlOaJvIdpyvyYjCSPOCVgpKF94gWJCPU0NlfuwZgBmT2QoV13Gs4ceJW4JamjEu2h/eWFCc1iJoEKovXAdVLwc6KAU8FmNS/TLCV0QkZsYKgkMdN+Po8/w+dGCXGUKHMk4Ln6eyMnsS4SmsmYwFgve4X4nzfIILr2cy7TDJiki4eiTGBIcNEFDrliFMTUEEIVN1kxHRNFKJjGaqYEd/nLq6R72XANv2vWW82yjio6RWfoArnoCrXQLWqjDqIoR8/oFb1ZT9aL9W59LEYrVrlzjP7A+vwBHQuWKw==\nAAAB/nicbVDLSsNAFJ3UV62vqLhyM1gEVyWRgi4LblxWsA9oYplMJu3QySTM3AglFPwVNy4Ucet3uPNvnLRZaOuBYQ7n3MucOUEquAbH+bYqa+sbm1vV7drO7t7+gX141NVJpijr0EQkqh8QzQSXrAMcBOunipE4EKwXTG4Kv/fIlOaJvIdpyvyYjCSPOCVgpKF94gWJCPU0NlfuwZgBmT2QoV13Gs4ceJW4JamjEu2h/eWFCc1iJoEKovXAdVLwc6KAU8FmNS/TLCV0QkZsYKgkMdN+Po8/w+dGCXGUKHMk4Ln6eyMnsS4SmsmYwFgve4X4nzfIILr2cy7TDJiki4eiTGBIcNEFDrliFMTUEEIVN1kxHRNFKJjGaqYEd/nLq6R72XANv2vWW82yjio6RWfoArnoCrXQLWqjDqIoR8/oFb1ZT9aL9W59LEYrVrlzjP7A+vwBHQuWKw==\n\nXia\n\nAAAB+HicbVDLSsNAFL2pr1ofjbp0M1gEVyWRQl0W3LisYB/QhjCZTtqhk0mYmQg15EvcuFDErZ/izr9x0mahrQcGDufcyz1zgoQzpR3n26psbe/s7lX3aweHR8d1++S0r+JUEtojMY/lMMCKciZoTzPN6TCRFEcBp4Ngflv4g0cqFYvFg14k1IvwVLCQEayN5Nv1cYT1LAizYe5nDOe+3XCazhJok7glaUCJrm9/jScxSSMqNOFYqZHrJNrLsNSMcJrXxqmiCSZzPKUjQwWOqPKyZfAcXRplgsJYmic0Wqq/NzIcKbWIAjNZxFTrXiH+541SHd54GRNJqqkgq0NhypGOUdECmjBJieYLQzCRzGRFZIYlJtp0VTMluOtf3iT966Zr+H2r0WmVdVThHC7gClxoQwfuoAs9IJDCM7zCm/VkvVjv1sdqtGKVO2fwB9bnD0wBk3I=\nAAAB+HicbVDLSsNAFL2pr1ofjbp0M1gEVyWRQl0W3LisYB/QhjCZTtqhk0mYmQg15EvcuFDErZ/izr9x0mahrQcGDufcyz1zgoQzpR3n26psbe/s7lX3aweHR8d1++S0r+JUEtojMY/lMMCKciZoTzPN6TCRFEcBp4Ngflv4g0cqFYvFg14k1IvwVLCQEayN5Nv1cYT1LAizYe5nDOe+3XCazhJok7glaUCJrm9/jScxSSMqNOFYqZHrJNrLsNSMcJrXxqmiCSZzPKUjQwWOqPKyZfAcXRplgsJYmic0Wqq/NzIcKbWIAjNZxFTrXiH+541SHd54GRNJqqkgq0NhypGOUdECmjBJieYLQzCRzGRFZIYlJtp0VTMluOtf3iT966Zr+H2r0WmVdVThHC7gClxoQwfuoAs9IJDCM7zCm/VkvVjv1sdqtGKVO2fwB9bnD0wBk3I=\nAAAB+HicbVDLSsNAFL2pr1ofjbp0M1gEVyWRQl0W3LisYB/QhjCZTtqhk0mYmQg15EvcuFDErZ/izr9x0mahrQcGDufcyz1zgoQzpR3n26psbe/s7lX3aweHR8d1++S0r+JUEtojMY/lMMCKciZoTzPN6TCRFEcBp4Ngflv4g0cqFYvFg14k1IvwVLCQEayN5Nv1cYT1LAizYe5nDOe+3XCazhJok7glaUCJrm9/jScxSSMqNOFYqZHrJNrLsNSMcJrXxqmiCSZzPKUjQwWOqPKyZfAcXRplgsJYmic0Wqq/NzIcKbWIAjNZxFTrXiH+541SHd54GRNJqqkgq0NhypGOUdECmjBJieYLQzCRzGRFZIYlJtp0VTMluOtf3iT966Zr+H2r0WmVdVThHC7gClxoQwfuoAs9IJDCM7zCm/VkvVjv1sdqtGKVO2fwB9bnD0wBk3I=\nAAAB+HicbVDLSsNAFL2pr1ofjbp0M1gEVyWRQl0W3LisYB/QhjCZTtqhk0mYmQg15EvcuFDErZ/izr9x0mahrQcGDufcyz1zgoQzpR3n26psbe/s7lX3aweHR8d1++S0r+JUEtojMY/lMMCKciZoTzPN6TCRFEcBp4Ngflv4g0cqFYvFg14k1IvwVLCQEayN5Nv1cYT1LAizYe5nDOe+3XCazhJok7glaUCJrm9/jScxSSMqNOFYqZHrJNrLsNSMcJrXxqmiCSZzPKUjQwWOqPKyZfAcXRplgsJYmic0Wqq/NzIcKbWIAjNZxFTrXiH+541SHd54GRNJqqkgq0NhypGOUdECmjBJieYLQzCRzGRFZIYlJtp0VTMluOtf3iT966Zr+H2r0WmVdVThHC7gClxoQwfuoAs9IJDCM7zCm/VkvVjv1sdqtGKVO2fwB9bnD0wBk3I=\n\na\n\nAAAB/HicbVDLSsNAFL2pr1pf0S7dBIvgqiRS0GXBjcsK9gFNLJPJtB06mQkzEyGE+ituXCji1g9x5984abPQ1gPDHM65lzlzwoRRpV3326psbG5t71R3a3v7B4dH9vFJT4lUYtLFggk5CJEijHLS1VQzMkgkQXHISD+c3RR+/5FIRQW/11lCghhNOB1TjLSRRnbdDwWLVBabK/eTKZ0/oJHdcJvuAs468UrSgBKdkf3lRwKnMeEaM6TU0HMTHeRIaooZmdf8VJEE4RmakKGhHMVEBfki/Nw5N0rkjIU0h2tnof7eyFGsinxmMkZ6qla9QvzPG6Z6fB3klCepJhwvHxqnzNHCKZpwIioJ1iwzBGFJTVYHT5FEWJu+aqYEb/XL66R32fQMv2s12q2yjiqcwhlcgAdX0IZb6EAXMGTwDK/wZj1ZL9a79bEcrVjlTh3+wPr8AX94lUI=\nAAAB/HicbVDLSsNAFL2pr1pf0S7dBIvgqiRS0GXBjcsK9gFNLJPJtB06mQkzEyGE+ituXCji1g9x5984abPQ1gPDHM65lzlzwoRRpV3326psbG5t71R3a3v7B4dH9vFJT4lUYtLFggk5CJEijHLS1VQzMkgkQXHISD+c3RR+/5FIRQW/11lCghhNOB1TjLSRRnbdDwWLVBabK/eTKZ0/oJHdcJvuAs468UrSgBKdkf3lRwKnMeEaM6TU0HMTHeRIaooZmdf8VJEE4RmakKGhHMVEBfki/Nw5N0rkjIU0h2tnof7eyFGsinxmMkZ6qla9QvzPG6Z6fB3klCepJhwvHxqnzNHCKZpwIioJ1iwzBGFJTVYHT5FEWJu+aqYEb/XL66R32fQMv2s12q2yjiqcwhlcgAdX0IZb6EAXMGTwDK/wZj1ZL9a79bEcrVjlTh3+wPr8AX94lUI=\nAAAB/HicbVDLSsNAFL2pr1pf0S7dBIvgqiRS0GXBjcsK9gFNLJPJtB06mQkzEyGE+ituXCji1g9x5984abPQ1gPDHM65lzlzwoRRpV3326psbG5t71R3a3v7B4dH9vFJT4lUYtLFggk5CJEijHLS1VQzMkgkQXHISD+c3RR+/5FIRQW/11lCghhNOB1TjLSRRnbdDwWLVBabK/eTKZ0/oJHdcJvuAs468UrSgBKdkf3lRwKnMeEaM6TU0HMTHeRIaooZmdf8VJEE4RmakKGhHMVEBfki/Nw5N0rkjIU0h2tnof7eyFGsinxmMkZ6qla9QvzPG6Z6fB3klCepJhwvHxqnzNHCKZpwIioJ1iwzBGFJTVYHT5FEWJu+aqYEb/XL66R32fQMv2s12q2yjiqcwhlcgAdX0IZb6EAXMGTwDK/wZj1ZL9a79bEcrVjlTh3+wPr8AX94lUI=\nAAAB2XicbZDNSgMxFIXv1L86Vq1rN8EiuCozbnQpuHFZwbZCO5RM5k4bmskMyR2hDH0BF25EfC93vo3pz0JbDwQ+zknIvSculLQUBN9ebWd3b/+gfugfNfzjk9Nmo2fz0gjsilzl5jnmFpXU2CVJCp8LgzyLFfbj6f0i77+gsTLXTzQrMMr4WMtUCk7O6oyaraAdLMW2IVxDC9YaNb+GSS7KDDUJxa0dhEFBUcUNSaFw7g9LiwUXUz7GgUPNM7RRtRxzzi6dk7A0N+5oYkv394uKZ9bOstjdzDhN7Ga2MP/LBiWlt1EldVESarH6KC0Vo5wtdmaJNChIzRxwYaSblYkJN1yQa8Z3HYSbG29D77odOn4MoA7ncAFXEMIN3MEDdKALAhJ4hXdv4r15H6uuat66tDP4I+/zBzjGijg=\nAAAB8XicbZBNS8MwHMb/nW9zTq169BIcgqfRetGj4MXjBPcCax1pmm5haVKSVChlfhUvHhTxy3jz25huO+jmH0J+PE9CnjxRxpk2nvft1DY2t7Z36ruNveb+waF71OxpmStCu0RyqQYR1pQzQbuGGU4HmaI4jTjtR9Pbyu8/UaWZFA+myGiY4rFgCSPYWGnkngSR5LEuUruVQTZhs0c8clte25sPWgd/CS1YTmfkfgWxJHlKhSEcaz30vcyEJVaGEU5njSDXNMNkisd0aFHglOqwnIefoXOrxCiRyi5h0Fz9faPEqa7y2ZMpNhO96lXif94wN8l1WDKR5YYKsngoyTkyElVNoJgpSgwvLGCimM2KyAQrTIztq2FL8Fe/vA69y7Zv+d6DOpzCGVyAD1dwA3fQgS4QKOAF3uDdeXZenY9FXTVn2dsx/Bnn8wcXBJPp\nAAAB8XicbZBNS8MwHMb/nW9zTq169BIcgqfRetGj4MXjBPcCax1pmm5haVKSVChlfhUvHhTxy3jz25huO+jmH0J+PE9CnjxRxpk2nvft1DY2t7Z36ruNveb+waF71OxpmStCu0RyqQYR1pQzQbuGGU4HmaI4jTjtR9Pbyu8/UaWZFA+myGiY4rFgCSPYWGnkngSR5LEuUruVQTZhs0c8clte25sPWgd/CS1YTmfkfgWxJHlKhSEcaz30vcyEJVaGEU5njSDXNMNkisd0aFHglOqwnIefoXOrxCiRyi5h0Fz9faPEqa7y2ZMpNhO96lXif94wN8l1WDKR5YYKsngoyTkyElVNoJgpSgwvLGCimM2KyAQrTIztq2FL8Fe/vA69y7Zv+d6DOpzCGVyAD1dwA3fQgS4QKOAF3uDdeXZenY9FXTVn2dsx/Bnn8wcXBJPp\nAAAB/HicbVDLSsNAFL3xWesr2qWbwSK4KokIuiy4cVnBPqCJZTKZtEMnkzAzEUKov+LGhSJu/RB3/o2TNgttPTDM4Zx7mTMnSDlT2nG+rbX1jc2t7dpOfXdv/+DQPjruqSSThHZJwhM5CLCinAna1UxzOkglxXHAaT+Y3pR+/5FKxRJxr/OU+jEeCxYxgrWRRnbDCxIeqjw2V+GlEzZ7wCO76bScOdAqcSvShAqdkf3lhQnJYio04Vipoeuk2i+w1IxwOqt7maIpJlM8pkNDBY6p8ot5+Bk6M0qIokSaIzSaq783ChyrMp+ZjLGeqGWvFP/zhpmOrv2CiTTTVJDFQ1HGkU5Q2QQKmaRE89wQTCQzWRGZYImJNn3VTQnu8pdXSe+i5Rp+5zTbl1UdNTiBUzgHF66gDbfQgS4QyOEZXuHNerJerHfrYzG6ZlU7DfgD6/MHfjiVPg==\nAAAB/HicbVDLSsNAFL2pr1pf0S7dBIvgqiRS0GXBjcsK9gFNLJPJtB06mQkzEyGE+ituXCji1g9x5984abPQ1gPDHM65lzlzwoRRpV3326psbG5t71R3a3v7B4dH9vFJT4lUYtLFggk5CJEijHLS1VQzMkgkQXHISD+c3RR+/5FIRQW/11lCghhNOB1TjLSRRnbdDwWLVBabK/eTKZ0/oJHdcJvuAs468UrSgBKdkf3lRwKnMeEaM6TU0HMTHeRIaooZmdf8VJEE4RmakKGhHMVEBfki/Nw5N0rkjIU0h2tnof7eyFGsinxmMkZ6qla9QvzPG6Z6fB3klCepJhwvHxqnzNHCKZpwIioJ1iwzBGFJTVYHT5FEWJu+aqYEb/XL66R32fQMv2s12q2yjiqcwhlcgAdX0IZb6EAXMGTwDK/wZj1ZL9a79bEcrVjlTh3+wPr8AX94lUI=\nAAAB/HicbVDLSsNAFL2pr1pf0S7dBIvgqiRS0GXBjcsK9gFNLJPJtB06mQkzEyGE+ituXCji1g9x5984abPQ1gPDHM65lzlzwoRRpV3326psbG5t71R3a3v7B4dH9vFJT4lUYtLFggk5CJEijHLS1VQzMkgkQXHISD+c3RR+/5FIRQW/11lCghhNOB1TjLSRRnbdDwWLVBabK/eTKZ0/oJHdcJvuAs468UrSgBKdkf3lRwKnMeEaM6TU0HMTHeRIaooZmdf8VJEE4RmakKGhHMVEBfki/Nw5N0rkjIU0h2tnof7eyFGsinxmMkZ6qla9QvzPG6Z6fB3klCepJhwvHxqnzNHCKZpwIioJ1iwzBGFJTVYHT5FEWJu+aqYEb/XL66R32fQMv2s12q2yjiqcwhlcgAdX0IZb6EAXMGTwDK/wZj1ZL9a79bEcrVjlTh3+wPr8AX94lUI=\nAAAB/HicbVDLSsNAFL2pr1pf0S7dBIvgqiRS0GXBjcsK9gFNLJPJtB06mQkzEyGE+ituXCji1g9x5984abPQ1gPDHM65lzlzwoRRpV3326psbG5t71R3a3v7B4dH9vFJT4lUYtLFggk5CJEijHLS1VQzMkgkQXHISD+c3RR+/5FIRQW/11lCghhNOB1TjLSRRnbdDwWLVBabK/eTKZ0/oJHdcJvuAs468UrSgBKdkf3lRwKnMeEaM6TU0HMTHeRIaooZmdf8VJEE4RmakKGhHMVEBfki/Nw5N0rkjIU0h2tnof7eyFGsinxmMkZ6qla9QvzPG6Z6fB3klCepJhwvHxqnzNHCKZpwIioJ1iwzBGFJTVYHT5FEWJu+aqYEb/XL66R32fQMv2s12q2yjiqcwhlcgAdX0IZb6EAXMGTwDK/wZj1ZL9a79bEcrVjlTh3+wPr8AX94lUI=\nAAAB/HicbVDLSsNAFL2pr1pf0S7dBIvgqiRS0GXBjcsK9gFNLJPJtB06mQkzEyGE+ituXCji1g9x5984abPQ1gPDHM65lzlzwoRRpV3326psbG5t71R3a3v7B4dH9vFJT4lUYtLFggk5CJEijHLS1VQzMkgkQXHISD+c3RR+/5FIRQW/11lCghhNOB1TjLSRRnbdDwWLVBabK/eTKZ0/oJHdcJvuAs468UrSgBKdkf3lRwKnMeEaM6TU0HMTHeRIaooZmdf8VJEE4RmakKGhHMVEBfki/Nw5N0rkjIU0h2tnof7eyFGsinxmMkZ6qla9QvzPG6Z6fB3klCepJhwvHxqnzNHCKZpwIioJ1iwzBGFJTVYHT5FEWJu+aqYEb/XL66R32fQMv2s12q2yjiqcwhlcgAdX0IZb6EAXMGTwDK/wZj1ZL9a79bEcrVjlTh3+wPr8AX94lUI=\nAAAB/HicbVDLSsNAFL2pr1pf0S7dBIvgqiRS0GXBjcsK9gFNLJPJtB06mQkzEyGE+ituXCji1g9x5984abPQ1gPDHM65lzlzwoRRpV3326psbG5t71R3a3v7B4dH9vFJT4lUYtLFggk5CJEijHLS1VQzMkgkQXHISD+c3RR+/5FIRQW/11lCghhNOB1TjLSRRnbdDwWLVBabK/eTKZ0/oJHdcJvuAs468UrSgBKdkf3lRwKnMeEaM6TU0HMTHeRIaooZmdf8VJEE4RmakKGhHMVEBfki/Nw5N0rkjIU0h2tnof7eyFGsinxmMkZ6qla9QvzPG6Z6fB3klCepJhwvHxqnzNHCKZpwIioJ1iwzBGFJTVYHT5FEWJu+aqYEb/XL66R32fQMv2s12q2yjiqcwhlcgAdX0IZb6EAXMGTwDK/wZj1ZL9a79bEcrVjlTh3+wPr8AX94lUI=\nAAAB/HicbVDLSsNAFL2pr1pf0S7dBIvgqiRS0GXBjcsK9gFNLJPJtB06mQkzEyGE+ituXCji1g9x5984abPQ1gPDHM65lzlzwoRRpV3326psbG5t71R3a3v7B4dH9vFJT4lUYtLFggk5CJEijHLS1VQzMkgkQXHISD+c3RR+/5FIRQW/11lCghhNOB1TjLSRRnbdDwWLVBabK/eTKZ0/oJHdcJvuAs468UrSgBKdkf3lRwKnMeEaM6TU0HMTHeRIaooZmdf8VJEE4RmakKGhHMVEBfki/Nw5N0rkjIU0h2tnof7eyFGsinxmMkZ6qla9QvzPG6Z6fB3klCepJhwvHxqnzNHCKZpwIioJ1iwzBGFJTVYHT5FEWJu+aqYEb/XL66R32fQMv2s12q2yjiqcwhlcgAdX0IZb6EAXMGTwDK/wZj1ZL9a79bEcrVjlTh3+wPr8AX94lUI=\n\nFigure 2: Probabilistic graphical\nmodel of our SCAN. Generative\nmodel dependencies are shown\nin solid arrows, while inference\nmodel dependencies are shown\nin dashed arrows.\n\nThe Variational Evidence Lower Bound. As shown in Fig. 1, we have two relationship observa-\ntions: the adjacency matrix A and the node attribute matrix X. The elements in both of matrices (i.e.\nthe edges between nodes and the attribute values of nodes) can be categorized into \ufb01ve cases: (case 1)\nan edge connecting two labelled nodes; (case 2) an edge connecting two unlabelled nodes; (case 3) an\nedge connecting a labelled node and an unlabelled node; (case 4) an attribute value associating with a\nlabelled node and an attribute; (case 5) an attribute value associating with an unlabelled node and\n\n5\n\n\fij) for case 1, U(Ouu\n\nia ) for case 5, respectively 1.\n\nan attribute. We can easily obtain the ELBOs for these \ufb01ve types of atomic observations according\nto Eq. 5, namely, L(Oll\nia) for case 4 and\nC(Oua\nOnce all the ELBOs for the above \ufb01ve cases are obtained, we can obtain the variational bound on the\nmarginal likelihood for the entire adjacency matrix and node attribute matrix as follows:\nij )\n\nij ) for case 2, M(Olu\n\nij ) for case 3, B(Ola\n\nJ (A, X, Yl) = Xi,j2V l L(Oll\n+ Xi2V l,a2A\n\nij ) + Xi2V l,j2V u M(Olu\nij) + Xi2V u,j2V u U(Ouu\nia) + Xi2V u,a2A\nC(Oua\nB(Ola\nia ).\n\nIn the objective function of Eq. 8, the \ufb01rst three terms on right-hand side are the ELBOs on the\nmarginal likelihood of edges, while the other two terms are the bound loss for attributes. To make\nour model more \ufb02exible to govern the loss between the edge data points and the attribute data points,\nwe introduce an adjustable hyper-parameter that balances reconstruction accuracy between edges\nand attributes. In addition, similar to [19], we wish the parameters of our predictive distribution, i.e.\nv ; therefore,\nqc(Yv | (Fn\nwe add a classi\ufb01cation loss to Eq. 8 and introduce a hyper-parameter \u21b5 to govern the relative weight\nbetween generative and purely discriminative models, which results in the following loss:\n\nv )), can also be trained within the labelled nodes based on their feature Fn\n\n(8)\n\nJ (A, X, Yl) = \u00b7 Xi,j2V l L(Oll\n+ (1 ) \u00b7 Xi2V l,a2A\n\nij) + Xi2V u,j2V u U(Ouu\nia) + Xi2V u,a2A\nB(Ola\n\nij ) + Xi2V l,j2V u M(Olu\nij )\n\nia ) + \u21b5 \u00b7 Ev\u21e0V l[ log qc(Yv | (Fn\nC(Oua\n\nv ))] .\n\n(9)\n\ni = \u00b5n + \u270fn (or Za\n\nOptimization. The parameters, i.e. \u2713 = (\u2713e, \u2713a) and = (n, a, c) (Fig. 2), of the generative\nand inference networks are jointly trained by optimizing Eq. 9 using gradient descent2.\nWe assume the priors and the variational posteriors of Zn and Za to be Gaussian distributions, then\nthe KL-divergence terms in Eq. 9 have analytical forms. However, analytical solutions of expectations\nw.r.t. these two variational posteriors are still intractable in the general case. To address this problem,\nwe can reduce the problem of estimating the gradient w.r.t. parameters of the posterior distribution to\nthe simpler problem of estimating the gradient w.r.t. parameters of a deterministic function, which\nis called the reparameterization trick [18, 32]. Speci\ufb01cally, we sample noise \u270f \u21e0N (0, ID) and\nreparameterize Zn\ni = \u00b5a + \u270fa for attributes). By doing so, the value of\ni (or Za\ni ) is deterministic given both the parameters of variational posterior distributions, so that\nZn\nthe stochasticity in the sampling process is isolated and the gradient with respect to \u00b5n and n (or\n\u00b5a and a) can be back-propagated through the sampled Zn\nHowever, this trick is unavailable for optimizing the parameters of the latent label distribution\ni ), because the categorical distribution is not reparameterizable. Kingma et.al [19]\nqc(Yu\ni\napproach this by marginalizing out Yu\ni over all classes, so that for unlabelled data, inference is still\non qc(Yu\ni . As we have mentioned in subsection 3.1, this simple solution is\n| Fn\ni\nprohibitively expensive for a large number of classes, especially in our case where we have double\nexpectation over qc(Yu\nj ) of\ni\nij ), please refer to Eq. 18 in the supplementary material). In this paper, we alleviate this by\nU(Ouu\napplying the Gumbel-Softmax trick.\nThe Gumbel-Softmax trick [33, 34] provides a continuous and differentiable approximation to draw\ncategorical samples Yu\n\ni from a categorical distribution with class probabilities \u21e1c:\n\ni ) for each Yu\n| Fn\n\nj ) (factorized from qc(Yu\n\ni ) and qc(Yu\nj\n\ni (or Za\ni ).\n\n| Fn\n\n| Fn\n\n| Fn\n\ni , Yu\nj\n\ni , Fn\n\nYu\n\nik =\n\nexp(log \u21e1c,k + gk)/\u2327\nk=1 exp(log \u21e1c,k + gk)/\u2327\n\nPK\n\n,\n\n(10)\n\nwhere {gk}K\nk=1 are i.i.d. samples drawn from the Gumbel(0, 1) distribution, and \u2327 is the softmax\ntemperature which is set to be 0.2 in our experiments. Samples from the Gumbel-Softmax distribution\nbecome one-hot when \u2327 ! 0 and smooth when \u2327> 0. With this trick, Gumbel-Softmax allows us to\nbackpropagate through Yu\n\ni \u21e0 qc(Yu | Fn) for single sample gradient estimation.\n1The detail derivations of the \ufb01ve ELBOs can be found in the supplementary material.\n2The implementation details of the inference and generative networks are given in our supplementary\n\nmaterial.\n\n6\n\n\f5 Experiments\n\nThe research questions guiding the remainder of this paper are: (RQ1) How is performance of our\nSCAN in semi-supervised node classi\ufb01cation task? (RQ2) Can our SCAN perform better than other\nmodels in the attribute inference task, where capturing the af\ufb01nities between nodes and attributes is\ncrucial? (RQ3) Can our SCAN learn meaningful embeddings for the task of network visualizations?\n\n5.1 Experimental Settings\nTo evaluate the performance of our SCAN model on semi-supervised node classi\ufb01cation task,\nfour state-of-the-art semi-supervised network embedding methods are included for comparisons:\nPlanetoid-T [11], GCN [12], SEANO [13] and GraphSGAN [14]. We also evaluate the learned\nembeddings of SCAN by the attribute inference task, comparing to four attribute inference baselines:\nSAN [16], EdgeExp [35], BLA [15] and CAN [24]. All experiments of this paper are conducted\nbase on three real-world attributed networks, i.e. Pubmed [12], BlogCatalog [36] and Flickr [36].\nFor our SCAN model, the inferred embeddings of nodes can be directly used as the input features\nof a classi\ufb01er to predict the class labels. Here we apply Support Vector Machine (SVM) as our\nclassi\ufb01er, which we refer to it as SCAN_SVM. In addition, the inference model of SCAN (i.e. the\ndiscriminative network) also infers the latent labels for all the unlabelled nodes during the inference\nprocess, which can be used as another classi\ufb01er, namely the SCAN_DIS. 1\n\nTable 1: Performance of semi-supervised node classi\ufb01cation. The best and the second best perfor-\nmance runs per metric per dataset are marked in boldface and underlined, respectively.\n\nMethod\n\nPubmed\n\nFlickr\n\nBlogCatalog\n\nMa_F1 Mi_F1 ACC Ma_F1 Mi_F1 ACC Ma_F1 Mi_F1 ACC\nSEANO\n.841\n.648\nPlanetoid-T\n.817\n.815\nGCN\n.538\n.838\nGraphSGAN\n.719\n.839\nSCAN_SVM .847\u2020\n.835\u2020\nSCAN_DIS\n.850\u2020\n.839\u2020\n\n.850\n.823\n.850\n.841\n.851\n.862\u2020\n\n.845\n.825\n.847\n.842\n.852\u2020\n.858\u2020\n\n.738\n.721\n.286\n.697\n.747\u2020\n.749\u2020\n\n.748\n.743\n.291\n.715\n.750\n.753\u2020\n\n.741\n.733\n.309\n.702\n.747\u2020\n.750\u2020\n\n.627\n.803\n.509\n.698\n.820\u2020\n.829\u2020\n\n.635\n.811\n.527\n.703\n.829\u2020\n.832\u2020\n\n5.2 Results\nSemi-supervised Node Classi\ufb01cation. To answer (RQ1), we conduct semi-supervised node classi-\n\ufb01cation experiments in the three networks. Speci\ufb01cally, we randomly select 10% of nodes as the\nlabelled nodes, train our SCAN model to obtain the embeddings of nodes and predict the labels for\nthe unlabelled nodes by the discriminator, and then feed the obtained embeddings of nodes into an\nSVM classi\ufb01er to predict the labels. We repeatedly this process 10 times and report the average\nperformance. As for evaluation metrics, we employ macro F1 (Ma_F1), micro F1 (Mi_F1) and\naccuracy (ACC) to measure the performance of semi-supervised node classi\ufb01cation. Tab. 1 shows the\nresult of our SCAN methods and other four baseline methods on our datasets2. As shown in the table,\nboth our SCAN_DIS and SCAN_SVM can outperform the baseline models, and SCAN_DIS always\nachieves the best performance in all the metrics with signi\ufb01cant improvement. It is worth noting that\nwhile some baselines like SEANO and GCN perform poorly on BlogCatalog dataset, our methods,\nboth SCAN_DIS and SCAN_SVM, can still achieve signi\ufb01cantly better performance. This result\nshows that our model is not only capable of accurately classifying the nodes by the discriminative\nnetwork, but also capable of learning effective representations of nodes for the attributed networks.\n\nAttribute Inference. Subsequently, we answer RQ3 and aim at understanding the performance of\nSCAN and the baselines over attribute inference task. Attribute inference aims at predicting the value\nof attributes of the nodes. In this task, we take four state-of-the-art attribute inference algorithms,\n\n1Other experimental setting details and parameter sensitivity analysis of our model can be found in supple-\n\nmentary material, due to the space limitation.\n\n2In each dataset, signi\ufb01cant improvements over the comparative methods (other than ours) are marked with\n\n\u2020 (paired t-test, p < .05)\n\n7\n\n\fnamely SAN [16], EdgeExp [35] BLA [15] and CAN [24], as our baselines for performance com-\nparison. We adopt the same experimental setting as in [37, 12], we randomly divide all edges into\nthree sets, i.e., the training (85%), validating (5%) and testing (10%) sets, and employ area under the\nROC curve (AUC) and average precision (AP) scores as evaluation metrics to measure the attribute\ninference performance. Tab. 2 presents the attribute inference performance of our SCAN and the\nbaseline models in the three attributed networks. We can observe that our model outperforms all the\nbaseline models in all the datasets, and the improvement is signi\ufb01cant in Pubmed and BlogCatalog\nnetworks. This can be explained by the fact that our SCAN optimizes a loss function consisting of\nthe reconstruction error of all the attributes. This again shows that our co-embedding model can\nlearn effective representations for both nodes and attributes where the in\ufb01nities between nodes and\nattributes can be effectively captured and measured.\n\nMethod\n\nEdgeExp\nSAN\nBLA\nCAN\nSCAN\n\nPubmed\nAP\n.576\n.572\n.602\n.652\n.682\u2020\n\nAUC\n.586\n.579\n.622\n.670\n.713\u2020\n\nFlickr\n\nAUC\n.678\n.653\n.730\n.867\n.874\u2020\n\nAP\n.685\n.660\n.769\n.868\n.871\n\nBlogCatalog\nAUC\nAP\n.744\n.684\n.710\n.694\n.792\n.787\n.867\n.865\n.895\u2020\n.893\u2020\n\nTable 2: Attribute inference\nperformance. The best runs\nper metric per dataset are\nmarked in boldface.\n\nNetwork Visualization. Finally, to further evaluate the qualities of embeddings in our approach, we\nmake a comparison on the visualization of node representation in Fig. 3. Speci\ufb01cally, we obtain all\nthe 64-dimensional embeddings of nodes for each comparison methods, then use the t-SNE tool [38]\nto transfer them into 2-D vectors and plot these vectors on 2-D planes. We can \ufb01nd that our approach\ncan achieve more compact and separated clusters compared with the baseline methods. This result\ncan also explain why our approach achieves better performance on node classi\ufb01cation task. We\nalso visualize the variances of each node representation by ellipsoids, where the 2-D variances are\nobtained by dividing all the dimension into two groups and then calculating average variances of\nthem. We see that some of these nodes have large variances as their features are relatively sparser,\nresulting in more uncertainty of their representations.\n\n(a) GCN\n\n(b) SEANO\n\n(c) Planetoid-T\n\n(d) SCAN\n\nFigure 3: 2-D visualization of the inferred embeddings on the BlogCatalog dataset. The same\ncolour indicates the same class label. Ellipsoids surrounding the nodes in our model are the averaged\nvariances indicating the uncertainty of their embeddings. Note that the ellipsoids of our model need\nto be zoomed in to be visible.\n6 Conclusion\nWe aim at solving the problem of embedding attributed networks in a semi-supervised way. We\nshowed how the SVAE can be generalized to the heterogeneous data, and proposed a semi-supervised\nco-embedding model (called SCAN) to solve the problem. Our SCAN learns low-dimensional\nGaussian embeddings for both nodes and attributes in the same semantic space in a semi-supervised\nway, such that the af\ufb01nities between nodes and their attributes and the similarities among nodes can\nbe effectively measured with the uncertainty can be preserved. Meanwhile, it is also able to learn an\neffective discriminative model to generalize from small labelled data to large unlabelled ones. Our\nexperiments showed that our SCAN model can yield excellent and better performance compared with\nthe state-of-the-art baselines in various applications, including semi-supervised node classi\ufb01cation\nand attribute inference, and leverage the expressive power to obtain high-quality representations of\nboth nodes and attributes. As to future work, we intend to extend our SCAN model to heterogeneous\nnetworks, and embed nodes and attributes for dynamic attributed networks.\nAcknowledgments. This research was partially supported by the National Natural Science Founda-\ntion of China (Grant No. 61906219).\n\n8\n\n\fReferences\n[1] Bryan Perozzi, Rami Al-Rfou, and Steven Skiena. Deepwalk: Online learning of social\n\nrepresentations. In SIGKDD, pages 701\u2013710. ACM, 2014.\n\n[2] Aditya Grover and Jure Leskovec. node2vec: Scalable feature learning for networks.\n\nSIGKDD, pages 855\u2013864. ACM, 2016.\n\nIn\n\n[3] Jian Tang, Meng Qu, Mingzhe Wang, Ming Zhang, Jun Yan, and Qiaozhu Mei. Line: Large-\n\nscale information network embedding. In WWW, pages 1067\u20131077. WWW, 2015.\n\n[4] Shaosheng Cao, Wei Lu, and Qiongkai Xu. Grarep: Learning graph representations with global\n\nstructural information. In CIKM, pages 891\u2013900. ACM, 2015.\n\n[5] Xiao Wang, Peng Cui, Jing Wang, Jian Pei, Wenwu Zhu, and Shiqiang Yang. Community\n\npreserving network embedding. In AAAI, 2017.\n\n[6] Aleksandar Bojchevski and Stephan G\u00fcnnemann. Deep gaussian embedding of graphs: Unsu-\n\npervised inductive learning via ranking. In ICLR, 2018.\n\n[7] Xiao Huang, Qingquan Song, Jundong Li, and Xia Hu. Exploring expert cognition for attributed\n\nnetwork embedding. In WSDM, pages 270\u2013278. ACM, 2018.\n\n[8] Xiao Huang, Jundong Li, and Xia Hu. Accelerated attributed network embedding. In SDM,\n\npages 633\u2013641. SIAM, 2017.\n\n[9] Zhen Zhang, Hongxia Yang, Jiajun Bu, Sheng Zhou, Pinggang Yu, Jianwei Zhang, Martin Ester,\nand Can Wang. Anrl: Attributed network representation learning via deep neural networks. In\nIJCAI, pages 3155\u20133161, 2018.\n\n[10] Will Hamilton, Zhitao Ying, and Jure Leskovec. Inductive representation learning on large\n\ngraphs. In NIPS, pages 1024\u20131034, 2017.\n\n[11] Zhilin Yang, William Cohen, and Ruslan Salakhutdinov. Revisiting semi-supervised learning\n\nwith graph embeddings. In ICML, pages 40\u201348, 2016.\n\n[12] Thomas N. Kipf and Max Welling. Semi-supervised classi\ufb01cation with graph convolutional\n\nnetworks. 2017.\n\n[13] Jiongqian Liang, Peter Jacobs, Jiankai Sun, and Srinivasan Parthasarathy. Semi-supervised\n\nembedding in attributed networks with outliers. In SDM, pages 153\u2013161. SIAM, 2018.\n\n[14] Ming Ding, Jie Tang, and Jie Zhang. Semi-supervised learning on graphs with generative\n\nadversarial nets. In CIKM, pages 913\u2013922. ACM, 2018.\n\n[15] Carl Yang, Lin Zhong, Li-Jia Li, and Luo Jie. Bi-directional joint inference for user links and\n\nattributes on large social graphs. In WWW, pages 564\u2013573. WWW, 2017.\n\n[16] Neil Zhenqiang Gong, Ameet Talwalkar, Lester Mackey, Ling Huang, Eui Chul Richard Shin,\nEmil Stefanov, Elaine Runting Shi, and Dawn Song. Joint link prediction and attribute inference\nusing a social-attribute network. TIST, 5(2):27, 2014.\n\n[17] Shangsong Liang, Xiangliang Zhang, Zhaochun Ren, and Evangelos Kanoulas. Dynamic\n\nembeddings for user pro\ufb01ling in twitter. In SIGKDD, pages 1764\u20131773. ACM, 2018.\n\n[18] Diederik P. Kingma and Max Welling. Auto-encoding variational bayes. In ICLR, 2014.\n\n[19] Diederik P Kingma, Shakir Mohamed, Danilo Jimenez Rezende, and Max Welling. Semi-\n\nsupervised learning with deep generative models. In NIPS, pages 3581\u20133589, 2014.\n\n[20] Daixin Wang, Peng Cui, and Wenwu Zhu. Structural deep network embedding. In SIGKDD,\n\npages 1225\u20131234. ACM, 2016.\n\n[21] Cheng Yang, Zhiyuan Liu, Deli Zhao, Maosong Sun, and Edward Y Chang. Network represen-\n\ntation learning with rich text information. In IJCAI, pages 2111\u20132117, 2015.\n\n9\n\n\f[22] Suhang Wang, Charu Aggarwal, Jiliang Tang, and Huan Liu. Attributed signed network\n\nembedding. In CIKM, pages 137\u2013146. ACM, 2017.\n\n[23] Hongchang Gao and Heng Huang. Deep attributed network embedding. In IJCAI, pages\n\n3364\u20133370, 2018.\n\n[24] Zaiqiao Meng, Shangsong Liang, Hongyan Bao, and Xiangliang Zhang. Co-embedding\n\nattributed networks. In WSDM, pages 393\u2013401. ACM, 2019.\n\n[25] Shizhu He, Kang Liu, Guoliang Ji, and Jun Zhao. Learning to represent knowledge graphs with\n\ngaussian embedding. In CIKM, pages 623\u2013632. ACM, 2015.\n\n[26] Ludovic Dos Santos, Benjamin Piwowarski, and Patrick Gallinari. Multilabel classi\ufb01cation on\nheterogeneous graphs with gaussian embeddings. In ECMLKDD, pages 606\u2013622. Springer,\n2016.\n\n[27] Weidi Xu, Haoze Sun, Chao Deng, and Ying Tan. Variational autoencoder for semi-supervised\n\ntext classi\ufb01cation. In AAAI, 2017.\n\n[28] Yong Cheng, Wei Xu, Zhongjun He, Wei He, Hua Wu, Maosong Sun, and Yang Liu. Semi-\nIn ACL, volume 1, pages 1965\u20131974,\n\nsupervised learning for neural machine translation.\n2016.\n\n[29] Caio Corro and Ivan Titov. Differentiable perturb-and-parse: Semi-supervised parsing with a\n\nstructured variational autoencoder. In ICLR, 2019.\n\n[30] Dawen Liang, Rahul G Krishnan, Matthew D Hoffman, and Tony Jebara. Variational autoen-\ncoders for collaborative \ufb01ltering. In WWW, pages 689\u2013698. International World Wide Web\nConferences Steering Committee, 2018.\n\n[31] Teng Xiao, Shangsong Liang, Weizhou Shen, and Zaiqiao Meng. Bayesian deep collaborative\n\nmatrix factorization. In AAAI, 2019.\n\n[32] Diederik Kingma and Max Welling. Ef\ufb01cient gradient-based inference through transformations\n\nbetween bayes nets and neural nets. In ICML, pages 1782\u20131790, 2014.\n\n[33] Eric Jang, Shixiang Gu, and Ben Poole. Categorical reparametrization with gumble-softmax. In\n\nICLR, 2017.\n\n[34] Chris J Maddison, Andriy Mnih, and Yee Whye Teh. The concrete distribution: A continuous\n\nrelaxation of discrete random variables. In ICLR, 2017.\n\n[35] Deepayan Chakrabarti, Stanislav Funiak, Jonathan Chang, and Sofus A Macskassy. Joint\ninference of multiple label types in large networks. In ICML, pages II\u2013874. JMLR. org, 2014.\n[36] Xiao Huang, Jundong Li, and Xia Hu. Label informed attributed network embedding. In WSDM,\n\npages 731\u2013739. ACM, 2017.\n\n[37] Thomas N Kipf and Max Welling. Variational graph auto-encoders. In NIPS Workshop on\n\nBayesian Deep Learning, 2016.\n\n[38] Laurens van der Maaten and Geoffrey Hinton. Visualizing data using t-sne. JMLR, 9(Nov):2579\u2013\n\n2605, 2008.\n\n10\n\n\f", "award": [], "sourceid": 3507, "authors": [{"given_name": "Zaiqiao", "family_name": "Meng", "institution": "University of Glasgow"}, {"given_name": "Shangsong", "family_name": "Liang", "institution": "Sun Yat-sen University"}, {"given_name": "Jinyuan", "family_name": "Fang", "institution": "Sun Yat-sen University"}, {"given_name": "Teng", "family_name": "Xiao", "institution": "Sun Yat-sen University"}]}