{"id":"W4415139051","doi":"10.2118/228194-ms","title":"Stop Using Convolutional Neural Networks: Knowledge Distillation for an Interpretable and Lightweight Decision Tree in Rod Pump Working Condition Diagnosis","year":2025,"lang":"en","type":"article","venue":"SPE Annual Technical Conference and Exhibition","topic":"Oil and Gas Production Techniques","field":"Engineering","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Calgary; University of British Columbia","funders":"","keywords":"Decision tree; Leverage (statistics); Convolutional neural network; Tree (set theory); Field (mathematics); Decision tree learning; Feature (linguistics); Incremental decision tree; Pruning","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":[0.0009537398,0.001434166,0.001070312,0.0007090014,0.0005989284,0.001003703,0.00147015,0.001553624,0.002420045],"category_scores_gemma":[0.002874059,0.0004434329,0.001030863,0.0005925106,0.000551107,0.001617614,0.001098188,0.002135415,0.0007527221],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009585826,"about_ca_system_score_gemma":0.001804215,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01074675,"about_ca_topic_score_gemma":0.01182907,"domain_scores_codex":[0.9995151,0.0001057301,0.00003176123,0.0001473436,0.0001017542,0.00009836022],"domain_scores_gemma":[0.9989034,0.0006693556,0.0000757482,0.00008420854,0.0002073865,0.00005983639],"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":[0.0003619565,0.0002274629,0.002590134,0.0002088895,0.00009414644,0.0003068847,0.0001193595,0.6498866,0.006929131,0.006255329,0.007856356,0.3251638],"study_design_scores_gemma":[0.000009100119,0.00002547635,0.0001013397,0.00001208042,0.00001180757,0.00001318266,0.000009450092,0.9953708,0.001328151,0.002757552,0.0003560056,0.000005146851],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1254306,0.001881889,0.8599998,0.001770139,0.0002412165,0.0001749424,0.0007116877,0.005590002,0.004199688],"genre_scores_gemma":[0.8469406,0.0003783916,0.1472195,0.0008006892,0.0001104005,0.00015284,0.001325872,0.0001622998,0.002909329],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01074675,"threshold_uncertainty_score":0.02136838,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.024183795717615,"score_gpt":0.2941158875659894,"score_spread":0.2699320918483745,"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."}}