{"id":"W4396849416","doi":"10.1088/2632-2153/ad4ae5","title":"Reducing training data needs with minimal multilevel machine learning (M3L)","year":2024,"lang":"en","type":"article","venue":"Machine Learning Science and Technology","topic":"Machine Learning in Materials Science","field":"Materials Science","cited_by":18,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"H2020 European Research Council; Canada First Research Excellence Fund; Vetenskapsrådet; University of Toronto; European Commission; Canadian Institute for Advanced Research","keywords":"Training (meteorology); Computer science; Machine learning; Artificial intelligence; Training set; Geography","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":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","sts","scholarly_communication"],"consensus_categories":["sts"],"category_scores_codex":[0.006692573,0.0005243569,0.0005794838,0.001953907,0.001927124,0.001379263,0.002732683,0.0002236224,0.0004667375],"category_scores_gemma":[0.003691751,0.0004070601,0.00003183815,0.004126371,0.003314028,0.001601555,0.002096079,0.001765246,0.0002043997],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001200354,"about_ca_system_score_gemma":0.0006588875,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0006148388,"about_ca_topic_score_gemma":0.00005234312,"domain_scores_codex":[0.994541,0.0002349846,0.0005824038,0.002036498,0.00120394,0.001401209],"domain_scores_gemma":[0.9977946,0.0003643193,0.0002567433,0.001085228,0.0002415989,0.0002574623],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00005100241,0.00004023367,0.01202868,0.0001466818,0.00001683277,0.0001851734,0.003173636,0.005049903,0.8854761,0.004790348,0.00006234392,0.08897913],"study_design_scores_gemma":[0.0005112992,0.0008123909,0.0008590545,0.0004004686,0.00005773732,0.001303427,0.001334077,0.9360831,0.02092002,0.0004145695,0.0364716,0.0008322855],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9824846,0.001906153,0.005998946,0.004226999,0.000783839,0.0003193802,0.0000391333,0.002621775,0.001619212],"genre_scores_gemma":[0.9720285,0.00006868524,0.02629,0.00009847913,0.0001517325,0.00002883261,0.00003646739,0.00007567998,0.001221621],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9310332,"threshold_uncertainty_score":0.9998381,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02659985087734975,"score_gpt":0.2837103781007253,"score_spread":0.2571105272233756,"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."}}