{"id":"W4404782836","doi":"10.18653/v1/2024.emnlp-main.1231","title":"Is Child-Directed Speech Effective Training Data for Language Models?","year":2024,"lang":"en","type":"article","venue":"","topic":"Speech Recognition and Synthesis","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Training (meteorology); Speech recognition; Training set; Language model; Natural language processing; Artificial intelligence","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000365873,0.0001141333,0.0001358106,0.0001238604,0.00008067659,0.0003182164,0.0007952713,0.00005116022,0.0002030682],"category_scores_gemma":[0.00009418408,0.00009264739,0.00007109174,0.0003265959,0.00001468482,0.0007680138,0.0002022205,0.00008945641,0.0001426691],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001960835,"about_ca_system_score_gemma":0.00003873411,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003228388,"about_ca_topic_score_gemma":0.00002285519,"domain_scores_codex":[0.9989104,0.00003633472,0.0001229077,0.0005573815,0.000159521,0.0002134727],"domain_scores_gemma":[0.9988096,0.0004044198,0.00001576518,0.0006672831,0.00003666275,0.00006626773],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000002808052,0.00001665055,5.749361e-7,0.00002985987,0.00005065553,0.00002370147,0.002086553,0.000001275767,0.0001950523,0.007766537,0.009704616,0.9801217],"study_design_scores_gemma":[0.0001512647,0.00002548912,0.00002131917,0.00008441853,0.00001499411,0.00005325185,0.0002377329,0.9742091,0.01030798,0.005399851,0.009331608,0.0001630285],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.00134862,0.0003935061,0.9211935,0.001792623,0.000426925,0.0004395959,0.0001393021,0.001550585,0.07271539],"genre_scores_gemma":[0.3419594,0.00002625287,0.6517991,0.00221454,0.0003391183,0.0001084691,0.0001089486,0.00003686475,0.003407356],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.9799587,"threshold_uncertainty_score":0.3778051,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07857330905148112,"score_gpt":0.3169085696781402,"score_spread":0.2383352606266591,"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."}}