{"id":"W3108828338","doi":"","title":"Many-shot from low-shot: learning to annotate using mixed supervision for object detection","year":2020,"lang":"en","type":"book-chapter","venue":"Oxford University Research Archive (ORA) (University of Oxford)","topic":"Advanced Neural Network Applications","field":"Computer Science","cited_by":13,"is_retracted":false,"has_abstract":false,"ca_institutions":"Mila - Quebec Artificial Intelligence Institute","funders":"","keywords":"Computer science; Pascal (unit); Shot (pellet); Annotation; Artificial intelligence; Object detection; Single shot; One shot; Object (grammar); Training set; Set (abstract data type); Supervised learning; Classifier (UML); Pattern recognition (psychology); Labeled data; Machine learning; Computer vision; Programming language; Artificial neural network","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":["metaepi_narrow","sts"],"consensus_categories":[],"category_scores_codex":[0.0004289304,0.0005134105,0.000752621,0.001330994,0.001980785,0.00009695283,0.003271383,0.0003201864,0.00007746502],"category_scores_gemma":[0.00009166988,0.0007260677,0.0004998475,0.0008840376,0.0004972075,0.000988414,0.003650966,0.001589247,0.00003929801],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006203706,"about_ca_system_score_gemma":0.0003728965,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0003574536,"about_ca_topic_score_gemma":0.0009234525,"domain_scores_codex":[0.995859,0.0002809108,0.0002873656,0.001645589,0.001028494,0.0008986984],"domain_scores_gemma":[0.9961727,0.0009950724,0.0003657296,0.001062614,0.0007057433,0.0006981327],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0105179,0.0005837472,0.0004619659,0.001276544,0.001986164,0.002130231,0.0175558,0.04302295,0.1229773,0.3912818,0.009358242,0.3988474],"study_design_scores_gemma":[0.001922407,0.001408283,0.0004616018,0.0005303369,0.0001452656,0.0000128771,0.002230181,0.2089667,0.0008793774,0.03143873,0.7507899,0.00121435],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01136475,0.00004548713,0.9485449,0.001022482,0.0001614796,0.001888188,0.0006431336,0.0002835123,0.03604606],"genre_scores_gemma":[0.2834725,0.004401815,0.5759289,0.0002694657,0.0007573124,0.000005145863,0.001787903,0.0004221572,0.1329547],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.7414316,"threshold_uncertainty_score":0.9995191,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05895967697017019,"score_gpt":0.2762568271738079,"score_spread":0.2172971502036377,"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."}}