{"id":"W3043619075","doi":"10.1109/isca45697.2020.00092","title":"Echo: Compiler-based GPU Memory Footprint Reduction for LSTM RNN Training","year":2020,"lang":"en","type":"preprint","venue":"","topic":"Advanced Neural Network Applications","field":"Computer Science","cited_by":33,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"","keywords":"Computer science; Memory footprint; Recurrent neural network; Bottleneck; Compiler; Parallel computing; Deep learning; Artificial intelligence; Artificial neural network; Embedded system; Programming language","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009411729,0.001800171,0.0007305114,0.0006672102,0.0004986169,0.001018577,0.002774192,0.0006873561,0.00470364],"category_scores_gemma":[0.005085887,0.0005206038,0.0007075793,0.0009126232,0.0006626194,0.002400151,0.001436838,0.001913159,0.001463118],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009496363,"about_ca_system_score_gemma":0.002874414,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00498563,"about_ca_topic_score_gemma":0.0096914,"domain_scores_codex":[0.9989744,0.0002252413,0.00009732512,0.0002096367,0.0003282143,0.0001651935],"domain_scores_gemma":[0.9979784,0.0007884632,0.0001508612,0.0005714256,0.0004065634,0.0001044074],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001410897,0.0006115371,0.006858137,0.0006939861,0.0002168865,0.0003707017,0.0003824888,0.3710095,0.05686925,0.01359678,0.04284918,0.5051306],"study_design_scores_gemma":[0.0001155684,0.0001985364,0.0006660902,0.00002516382,0.00003474672,0.00005344065,0.00005855163,0.9656646,0.02307918,0.004092208,0.005983797,0.00002817075],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.1722498,0.003371349,0.7364523,0.001027501,0.0007332809,0.0003462187,0.001195527,0.07230087,0.01232316],"genre_scores_gemma":[0.5467672,0.0005110951,0.439287,0.0005938815,0.0001120216,0.0004793983,0.002284337,0.004712429,0.005252564],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.00498563,"threshold_uncertainty_score":0.01573527,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1184545215385083,"score_gpt":0.322659852216518,"score_spread":0.2042053306780097,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}