{"id":"W7046852729","doi":"","title":"Efficient fine-tuning of BERT-like models","year":2023,"lang":"en","type":"dissertation","venue":"eScholarship@McGill (McGill)","topic":"Magnetic confinement fusion research","field":"Physics and Astronomy","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"","keywords":"Inference; Deep learning; Transformer; Computational model; Task (project management); Artificial neural network; Language understanding; Computational learning theory","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":[0.00107615,0.0009761345,0.001009383,0.0004658177,0.0004905389,0.001349622,0.002230295,0.001364803,0.00665227],"category_scores_gemma":[0.007102886,0.00065627,0.0006026766,0.0003957505,0.0008752526,0.002062264,0.001504123,0.002782764,0.001304838],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001042374,"about_ca_system_score_gemma":0.001427562,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005364611,"about_ca_topic_score_gemma":0.01089524,"domain_scores_codex":[0.9995352,0.0001289211,0.00002184742,0.00009427214,0.0001317988,0.00008805551],"domain_scores_gemma":[0.9983293,0.001042111,0.00009521772,0.0002718882,0.0001802292,0.0000812342],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0001841303,0.0001559105,0.001169176,0.0001202522,0.00005444248,0.00009987906,0.00006978402,0.9020827,0.003384685,0.01569636,0.005120286,0.07186231],"study_design_scores_gemma":[0.00000756762,0.00001557549,0.00005043928,0.000005678761,0.000003633789,0.00001157798,0.000008188621,0.9952903,0.0004771701,0.003711911,0.0004150835,0.000002912884],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.1284602,0.001196324,0.8335905,0.001300482,0.0003094229,0.0001754536,0.0004342047,0.005749402,0.02878409],"genre_scores_gemma":[0.8675445,0.0003097669,0.1238788,0.0005951572,0.00006927962,0.0001453269,0.0006469483,0.0005403419,0.006269873],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.00665227,"threshold_uncertainty_score":0.02225405,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02246227878884424,"score_gpt":0.263509359038512,"score_spread":0.2410470802496677,"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."}}