{"id":"W4409793719","doi":"10.61091/jcmcc127a-178","title":"The Role and Optimization of Multi-task Learning Models in Legal Decision Prediction","year":2025,"lang":"en","type":"article","venue":"Journal of Combinatorial Mathematics and Combinatorial Computing","topic":"Artificial Intelligence in Law","field":"Social Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"","keywords":"Task (project management); Computer science; Artificial intelligence; Machine learning; Multi-task learning; Management science; Psychology; Engineering; Systems engineering","routes":{"ca_aff":false,"ca_fund":false,"ca_venue":true,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003476548,0.0001180568,0.0003449756,0.0001655676,0.0006913767,0.0002216179,0.000253689,0.0001431622,0.000001240851],"category_scores_gemma":[0.00184882,0.00009769679,0.00006549664,0.0003811048,0.0002521755,0.0003331498,0.0001283566,0.0003816901,1.328885e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00009532756,"about_ca_system_score_gemma":0.0002013781,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001714158,"about_ca_topic_score_gemma":0.00002195699,"domain_scores_codex":[0.997981,0.0002371559,0.0009700045,0.0001248492,0.0004765339,0.0002105131],"domain_scores_gemma":[0.9969575,0.001695359,0.0006646801,0.00009449491,0.0005176784,0.00007021892],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.00008557014,0.0001702479,0.0007196238,0.00002308373,0.00002354019,0.000001606884,0.004507159,0.03968316,0.00007330889,0.9436275,0.000008657518,0.0110765],"study_design_scores_gemma":[0.0007648174,0.0001438249,0.00004079736,0.0003424234,0.00002731478,0.000002323287,0.004984807,0.4113073,0.0001636935,0.5814006,0.0007494146,0.0000727217],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7812086,0.001572099,0.1996801,0.000271651,0.01301207,0.0005147433,9.995535e-7,0.00003226095,0.003707509],"genre_scores_gemma":[0.9942493,0.0003641006,0.005087276,0.000004051786,0.0002732401,0.000001158397,2.033051e-7,0.000008958393,0.00001170613],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3716241,"threshold_uncertainty_score":0.531758,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02259824559106606,"score_gpt":0.3058734535768515,"score_spread":0.2832752079857855,"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."}}