{"id":"W4389115567","doi":"10.48550/arxiv.2311.15077","title":"Multilingual self-supervised speech representations improve the speech recognition of low-resource African languages with codeswitching","year":2023,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Speech Recognition and Synthesis","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Canadian Institute for Advanced Research; National Science Foundation","keywords":"Computer science; Bespoke; Code (set theory); Natural language processing; Artificial intelligence; Speech recognition; Language model; Scratch; Programming language","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0007415763,0.000965559,0.0003758477,0.0004137373,0.0002938927,0.0007851641,0.0005671173,0.0005200809,0.00251924],"category_scores_gemma":[0.00327123,0.0002522387,0.0005222094,0.0003028463,0.0004608277,0.0013071,0.001375466,0.001468966,0.002993131],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003087929,"about_ca_system_score_gemma":0.000737959,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004703604,"about_ca_topic_score_gemma":0.009775554,"domain_scores_codex":[0.9993941,0.000187222,0.00002847954,0.0002399468,0.00007569836,0.00007464406],"domain_scores_gemma":[0.9988073,0.0005197226,0.00006549775,0.0002682249,0.0002658537,0.00007343716],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0007941676,0.0004811493,0.009663893,0.0002624459,0.0002263058,0.0002390552,0.0007099372,0.08404378,0.1358885,0.00175913,0.01167491,0.7542567],"study_design_scores_gemma":[0.00004722838,0.0003507057,0.007496843,0.00005458713,0.00008105512,0.0002113896,0.0004692048,0.8767285,0.1020551,0.003596697,0.008836906,0.00007165625],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5214196,0.0009451097,0.4473115,0.0007872524,0.0005646436,0.0001448187,0.001787613,0.01782405,0.009215419],"genre_scores_gemma":[0.8844242,0.0001971262,0.1026631,0.000305623,0.00008479707,0.0001087756,0.00445056,0.0007580366,0.007007779],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.004703604,"threshold_uncertainty_score":0.009352446,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07033471131280217,"score_gpt":0.2221951529390186,"score_spread":0.1518604416262165,"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."}}