NeurIPS 2019
Sun Dec 8th through Sat the 14th, 2019 at Vancouver Convention Center
Paper ID:7426
Title:Recurrent Kernel Networks

This paper generalizes convolutional kernel networks to model gaps in sequences. It further introduces a new point of view on RNNs and proposes a new way to simulate max pooling in RKHS. The author rebuttal was useful to clarify the originality of the work compared to existing models, and the additional experiments shown in the rebuttal (which should be included in the final version if the paper is accepted) were important to support the relevance of the model. There were still concerns that the scope of the work is somewhat limited to biological sequence processing, but there is a healthy tradition at NeurIPS to accept application papers with novel machine learning ideas. Overall, we therefore consider this paper as a solid piece of work bringing new ML ideas and promising results on biological sequence classification.