%PDF-1.3 1 0 obj << /Kids [ 4 0 R 5 0 R 6 0 R 7 0 R 8 0 R 9 0 R 10 0 R 11 0 R 12 0 R 13 0 R ] /Type /Pages /Count 10 >> endobj 2 0 obj << /Subject (Neural Information Processing Systems http\072\057\057nips\056cc\057) /Publisher (Curran Associates\054 Inc\056) /Language (en\055US) /Created (2018) /EventType (Poster) /Description-Abstract (Model\055free reinforcement learning aims to offer off\055the\055shelf solutions for controlling dynamical systems without requiring models of the system dynamics\056 We introduce a model\055free random search algorithm for training static\054 linear policies for continuous control problems\056 Common evaluation methodology shows that our method matches state\055of\055the\055art sample efficiency on the benchmark MuJoCo locomotion tasks\056 Nonetheless\054 more rigorous evaluation reveals that the assessment of performance on these benchmarks is optimistic\056 We evaluate the performance of our method over hundreds of random seeds and many different hyperparameter configurations for each benchmark task\056 This extensive evaluation is possible because of the small computational footprint of our method\056 Our simulations highlight a high variability in performance in these benchmark tasks\054 indicating that commonly used estimations of sample efficiency do not adequately evaluate the performance of RL algorithms\056 Our results stress the need for new baselines\054 benchmarks and evaluation methodology for RL algorithms\056) /Producer (PyPDF2) /Title (Simple random search of static linear policies is competitive for reinforcement learning) /Date (2018) /ModDate (D\07220190218213150\05508\04700\047) /Published (2018) /Type (Conference Proceedings) /firstpage (1800) /Book (Advances in Neural Information Processing Systems 31) /Description (Paper accepted and presented at the Neural Information Processing Systems Conference \050http\072\057\057nips\056cc\057\051) /Editors (S\056 Bengio and H\056 Wallach and H\056 Larochelle and K\056 Grauman and N\056 Cesa\055Bianchi and R\056 Garnett) /Author (Horia Mania\054 Aurelia Guy\054 Benjamin Recht) /lastpage (1809) >> endobj 3 0 obj << /Type /Catalog /Pages 1 0 R >> endobj 4 0 obj << /Parent 1 0 R /Contents 14 0 R /Resources 15 0 R /Rotate 0 /MediaBox [ 0 0 612 792 ] /Annots 40 0 R /Type /Page >> endobj 5 0 obj << /Parent 1 0 R /Contents 123 0 R /Resources 124 0 R /Rotate 0 /MediaBox [ 0 0 612 792 ] /Annots 176 0 R /Type /Page >> endobj 6 0 obj << /Parent 1 0 R /Contents 252 0 R /Resources 253 0 R /Rotate 0 /MediaBox [ 0 0 612 792 ] /Annots 273 0 R /Type /Page >> endobj 7 0 obj << /Parent 1 0 R /Contents 379 0 R /Resources 380 0 R /Rotate 0 /MediaBox [ 0 0 612 792 ] /Annots 458 0 R /Type /Page >> endobj 8 0 obj << /Parent 1 0 R /Contents 489 0 R /Resources 490 0 R /Rotate 0 /MediaBox [ 0 0 612 792 ] /Annots 491 0 R /Type /Page >> endobj 9 0 obj << /Parent 1 0 R /Contents 562 0 R /Resources 563 0 R /Rotate 0 /MediaBox [ 0 0 612 792 ] /Annots 571 0 R /Type /Page >> endobj 10 0 obj << /Parent 1 0 R /Contents 712 0 R /Resources 713 0 R /Rotate 0 /MediaBox [ 0 0 612 792 ] /Annots 817 0 R /Type /Page >> endobj 11 0 obj << /Parent 1 0 R /Contents 853 0 R /Resources 854 0 R /Rotate 0 /MediaBox [ 0 0 612 792 ] /Annots 855 0 R /Type /Page >> endobj 12 0 obj << /Parent 1 0 R /Contents 861 0 R /Resources 862 0 R /Rotate 0 /MediaBox [ 0 0 612 792 ] /Type /Page >> endobj 13 0 obj << /Parent 1 0 R /Contents 863 0 R /Resources 864 0 R /Rotate 0 /MediaBox [ 0 0 612 792 ] /Type /Page >> endobj 14 0 obj << /Length 4248 /Filter /FlateDecode >> stream xZI)PÅ-T$BQt (b|C%ѨC9_NN[8?:R'9nENNfi;K\ߋBy<{IIfQ8{Ϲ{p~ro˵*n[?:86Oꊼqfy0O]fެZV_צ*eљ%~vv{qAQtucs}ٗ8N3#gunvfJx).Eݛee8fSAA;^d"J8827XEnN|\?Nٴen&Ö=Pqv!X2 Dv_Uƹ1qc02}ܕ>qh'y}hۃ8>2=sYQ칿(_KY~_ϰ&?) bl3 &'E!9v8s0ԢF{qcߖǼ2}Y?Ƭ]>mf_->,AlE g<v,֯l:5.FğƠb:*VujO& y`+g? (0>o#cxsy@JBq0p8vBXNeS;Ms*ۇIc#i#~pOgA= U'=,E%Kj]S }Ԛn):{lۥ|?f}brпo ܃pS /̩{Bqen