{"id":"W4406404608","doi":"10.1007/s12665-024-12063-6","title":"Soft causal constraints in groundwater machine learning: a new way to balance accuracy and physical consistency","year":2025,"lang":"en","type":"article","venue":"Environmental Earth Sciences","topic":"Model Reduction and Neural Networks","field":"Physics and Astronomy","cited_by":2,"is_retracted":false,"has_abstract":false,"ca_institutions":"Université du Québec à Chicoutimi","funders":"","keywords":"Biogeosciences; Groundwater; Consistency (knowledge bases); Environmental engineering science; Balance (ability); Computer science; Econometrics; Environmental science; Hydrology (agriculture); Geology; Artificial intelligence; Mathematics; Geotechnical engineering; Psychology; Earth science","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00007583753,0.0001044136,0.0001208976,0.00004705424,0.0001617805,0.00005900673,0.00009053278,0.00001457637,0.0005633416],"category_scores_gemma":[0.000004023571,0.00008613159,0.00002823287,0.0001223101,0.0003906851,0.0001463677,0.00008053546,0.0001414852,0.00004016966],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000005953175,"about_ca_system_score_gemma":0.0000185677,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000159533,"about_ca_topic_score_gemma":0.00001025019,"domain_scores_codex":[0.9992418,0.00003260778,0.0001126054,0.0002891029,0.0001178734,0.0002060402],"domain_scores_gemma":[0.9997623,0.00005384534,0.00002820498,0.00006211585,0.000001016162,0.00009254505],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00002465389,0.000218002,0.7908149,0.000004295657,0.00001608263,0.000003429321,0.0009529992,0.003595802,0.009896813,0.005168228,0.0002428065,0.189062],"study_design_scores_gemma":[0.003371787,0.0008154855,0.7641551,0.0001968854,0.00004322871,0.000019427,0.003839066,0.1507999,0.01204381,0.01025321,0.05320149,0.001260541],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9937285,0.00006236658,0.001141054,0.000577006,0.00009507682,0.0001302364,0.000004254931,0.000009822067,0.004251722],"genre_scores_gemma":[0.9966779,0.000008213848,0.0002217231,0.0001694885,0.00006514057,0.000005802166,0.000004630994,0.000002813483,0.002844223],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1878015,"threshold_uncertainty_score":0.6168196,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01168865225692287,"score_gpt":0.2471802924468939,"score_spread":0.235491640189971,"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."}}