{"id":"W4416978430","doi":"10.1038/s41467-025-66983-3","title":"Sufficient is better than optimal for training neural networks","year":2025,"lang":"en","type":"article","venue":"Nature Communications","topic":"Stochastic Gradient Optimization Techniques","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Queen's University","funders":"Natural Sciences and Engineering Research Council of Canada; Government of Canada; Universities Space Research Association","keywords":"Overfitting; Leverage (statistics); Spurious relationship; Artificial neural network; Deep neural networks; Convolutional neural network; Feedforward neural network; Training (meteorology)","routes":{"ca_aff":true,"ca_fund":true,"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":[0.003647399,0.0009732002,0.001074577,0.0008575644,0.0006592596,0.001427344,0.001038757,0.001440701,0.002918703],"category_scores_gemma":[0.01716929,0.0006548099,0.0009023348,0.0006333704,0.002389655,0.004146681,0.001525046,0.003265399,0.000700434],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001061833,"about_ca_system_score_gemma":0.001485424,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001536616,"about_ca_topic_score_gemma":0.002301723,"domain_scores_codex":[0.9985937,0.0005650042,0.00009698455,0.0003373015,0.0003083371,0.00009859868],"domain_scores_gemma":[0.9961942,0.002423669,0.0002131067,0.0007910606,0.0002873846,0.00009054098],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002253053,0.0001316533,0.002712888,0.0003379854,0.000152577,0.00008677029,0.0002808576,0.4977442,0.00570029,0.3630269,0.005999505,0.123601],"study_design_scores_gemma":[0.00002097921,0.00007712277,0.0002452404,0.00005709189,0.00001873637,0.00004204552,0.0000282926,0.814009,0.00310737,0.1790422,0.003336971,0.000015054],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02029557,0.0005979357,0.9724352,0.001361651,0.0000875394,0.0000338659,0.00007021052,0.0009754876,0.004142524],"genre_scores_gemma":[0.5414159,0.0007625789,0.4507665,0.001472992,0.0001887389,0.0002457245,0.0003202621,0.000923305,0.003904005],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003647399,"threshold_uncertainty_score":0.01928955,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03103649933153931,"score_gpt":0.3159515656064312,"score_spread":0.2849150662748919,"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."}}