{"id":"W6920901177","doi":"10.60692/rz8hd-c0104","title":"Pragmatic Inference with a CLIP Listener for Contrastive Captioning","year":2023,"lang":"en","type":"article","venue":"Greater South Information System","topic":"Multimodal Machine Learning Applications","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"","keywords":"Closed captioning; Discriminative model; Inference; Leverage (statistics); Fluency; Prosody; Hyperparameter","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":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0004415665,0.0001489761,0.0001783985,0.0002667346,0.0002701342,0.0004055679,0.0003910824,0.00005787914,0.000002261749],"category_scores_gemma":[0.00009101177,0.0001145902,0.00004370241,0.0006369408,0.00002461917,0.0009957686,0.00008505966,0.00009688609,0.0008279618],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00007366374,"about_ca_system_score_gemma":0.00006589347,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002663459,"about_ca_topic_score_gemma":4.103629e-7,"domain_scores_codex":[0.9988205,0.0000512818,0.0004031513,0.0001958374,0.0002590884,0.0002700863],"domain_scores_gemma":[0.9987748,0.00008165298,0.0003149713,0.0004406522,0.0003058927,0.00008200496],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001251536,0.00001085371,0.2967274,0.001976926,0.0002349374,0.000007906634,0.3829944,0.1127645,0.00002275793,0.1921276,0.000394096,0.01261343],"study_design_scores_gemma":[0.000956552,0.00006666255,0.1148384,0.0001645545,0.00001333408,0.00002697808,0.00237452,0.8810558,0.00004080696,0.00002109782,0.0002308221,0.0002105036],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1146959,3.619918e-7,0.8816448,0.0002334624,0.0001050291,0.0007346519,0.0000357825,0.0009719498,0.001578054],"genre_scores_gemma":[0.9780795,2.247105e-8,0.02079345,0.0001011388,0.00003717845,0.0008423926,0.00003621474,0.000008724997,0.0001014072],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8633835,"threshold_uncertainty_score":0.99995,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03101776588450244,"score_gpt":0.2535954053786056,"score_spread":0.2225776394941031,"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."}}