{"id":"W3171727251","doi":"10.1088/2632-2153/ac4f3f","title":"Probing transfer learning with a model of synthetic correlated datasets","year":2022,"lang":"en","type":"article","venue":"Machine Learning Science and Technology","topic":"Domain Adaptation and Few-Shot Learning","field":"Computer Science","cited_by":22,"is_retracted":false,"has_abstract":true,"ca_institutions":"Canadian Institute for Advanced Research","funders":"European Commission; Royal Society; Wellcome Trust; Wellcome","keywords":"Computer science; Generalization; Transfer of learning; Artificial intelligence; Salient; Machine learning; Task (project management); Synthetic data; Parametric statistics; Feature (linguistics); Artificial neural network; Binary classification; Sample (material); Data mining; Mathematics","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":[0.001294087,0.0001416459,0.0002032079,0.0008152953,0.001273551,0.0000802463,0.0009776605,0.0000461683,0.00002240142],"category_scores_gemma":[0.0002371005,0.0001257579,0.00001782949,0.002825569,0.0006475226,0.0003682585,0.0005945161,0.001053113,0.000003189091],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000486442,"about_ca_system_score_gemma":0.0002285707,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002478956,"about_ca_topic_score_gemma":0.000003580305,"domain_scores_codex":[0.9981459,0.000103078,0.0002166164,0.0005739194,0.0005751801,0.0003852565],"domain_scores_gemma":[0.9993297,0.00006469884,0.0001090701,0.0003215966,0.0001076089,0.00006732742],"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":[0.00003612339,0.00009885513,0.01000051,0.00002595134,0.00001944685,0.0000343151,0.003858187,0.725479,0.04722899,0.1227252,0.000009992998,0.0904834],"study_design_scores_gemma":[0.0004168568,0.0005005091,0.00008991741,0.00001710591,0.000008020866,0.0001538383,0.000572784,0.9939442,0.001036515,0.0007212357,0.002367769,0.0001712909],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5537141,0.0002167973,0.441323,0.002558519,0.0000681694,0.0002638841,0.000005583763,0.0007062277,0.001143754],"genre_scores_gemma":[0.9918245,0.00001036075,0.007822018,0.00007470909,0.000001879432,0.00002729239,0.000007207062,0.00001150612,0.0002205377],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4381104,"threshold_uncertainty_score":0.979525,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.009529413654964517,"score_gpt":0.2171074155043629,"score_spread":0.2075780018493983,"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."}}