{"id":"W7132893767","doi":"","title":"BitChop: A Heuristic Approach to Memory Footprint Reduction in AI Training","year":2022,"lang":"","type":"dissertation","venue":"TSpace","topic":"Big Data and Digital Economy","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"University of Toronto","keywords":"Memory footprint; Reduction (mathematics); Footprint; Heuristics; Artificial neural network; Training (meteorology); Heuristic; Lossy compression","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008837247,0.001118516,0.0006820055,0.001148251,0.0006030403,0.001556931,0.003030022,0.001019823,0.00870489],"category_scores_gemma":[0.005329341,0.0004971528,0.0004663419,0.001546283,0.0008744164,0.002400334,0.001601281,0.001809977,0.002387613],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001034197,"about_ca_system_score_gemma":0.001883173,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004491786,"about_ca_topic_score_gemma":0.008487944,"domain_scores_codex":[0.9992593,0.0001620809,0.00005721836,0.0001032586,0.0003130131,0.0001051134],"domain_scores_gemma":[0.9980603,0.0008836215,0.0001442467,0.0004610344,0.0003872598,0.00006359479],"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":[0.000730005,0.0002604902,0.0016077,0.0004070913,0.0001032568,0.0001991304,0.0002841898,0.1754367,0.02439867,0.01464048,0.02460957,0.7573227],"study_design_scores_gemma":[0.000132,0.0002456501,0.0005876856,0.00007946567,0.00004159228,0.0001876563,0.0001271558,0.9337646,0.03289985,0.01958542,0.01231144,0.00003748727],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05289488,0.002106609,0.9120379,0.0009902762,0.0003318661,0.000342468,0.0007020838,0.01670069,0.01389317],"genre_scores_gemma":[0.2750931,0.0006807372,0.7123586,0.000747732,0.0001174246,0.0006586219,0.001397397,0.001618765,0.00732762],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.00870489,"threshold_uncertainty_score":0.0291208,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06713135526166314,"score_gpt":0.3354844691610666,"score_spread":0.2683531138994034,"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."}}