{"id":"W3092446983","doi":"10.1609/aaai.v36i6.20611","title":"Gradient Flow in Sparse Neural Networks and How Lottery Tickets Win","year":2022,"lang":"en","type":"article","venue":"Proceedings of the AAAI Conference on Artificial Intelligence","topic":"Advanced Neural Network Applications","field":"Computer Science","cited_by":37,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Calgary","funders":"","keywords":"Initialization; Computer science; Pruning; Generalization; Artificial intelligence; Artificial neural network; Inference; Machine learning; Flow (mathematics); Sparse matrix; Lottery; Early stopping; Work flow; Mathematics","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":[0.002083234,0.0004825847,0.0008004679,0.0007445403,0.001011953,0.001835323,0.00119302,0.001731137,0.006288683],"category_scores_gemma":[0.01320195,0.0006276377,0.0005050316,0.0004908389,0.002187296,0.004927706,0.001462152,0.002147075,0.0006791455],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00113161,"about_ca_system_score_gemma":0.0008540802,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003797818,"about_ca_topic_score_gemma":0.003743926,"domain_scores_codex":[0.9994811,0.0002219531,0.0000220491,0.0000965933,0.00009010539,0.00008819161],"domain_scores_gemma":[0.99753,0.001665736,0.000225247,0.0001826921,0.0002245592,0.0001717737],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002863585,0.00008918748,0.002875344,0.0001539127,0.00007751122,0.0001832145,0.0003081055,0.431939,0.001475761,0.4607673,0.01280606,0.08903825],"study_design_scores_gemma":[0.00002890747,0.00004355347,0.0004124572,0.00004146542,0.000012985,0.00005330957,0.00006433956,0.6520996,0.0006211806,0.344404,0.002201123,0.00001703395],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.170423,0.003179569,0.7865999,0.01097771,0.0006487388,0.00009945819,0.0002816172,0.001060648,0.02672933],"genre_scores_gemma":[0.9258141,0.001070685,0.05987611,0.0008122965,0.0002467575,0.0001223936,0.0001682644,0.0003550212,0.01153435],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.006288683,"threshold_uncertainty_score":0.0210377,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06280376046387758,"score_gpt":0.2699841959839114,"score_spread":0.2071804355200338,"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."}}