{"meta":{"page":1,"per_page":50,"max_per_page":100,"total":3,"total_is_capped":false,"direct_labels_cover":0,"predictions_cover":3,"direct_label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline (scores rank; they never assert a category)","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12","author_layer_release":"2026-06-26"},"query_hash":"50a8d0646b88","filters":{"venue":"Interspeech 2022"}},"results":[{"id":"W4297841867","doi":"10.21437/interspeech.2022-10761","title":"Daft-Exprt: Cross-Speaker Prosody Transfer on Any Text for Expressive Speech Synthesis","year":2022,"lang":"en","type":"article","venue":"Interspeech 2022","topic":"Speech Recognition and Synthesis","field":"Computer Science","cited_by":21,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Ubisoft (Canada)","funders":"","keywords":"Prosody; Computer science; Speech synthesis; Speech recognition; Transfer (computing); Natural language processing","authors":[{"name":"Julian Zaïdi","is_ca":true},{"name":"Hugo Seuté","is_ca":true},{"name":"Benjamin van Niekerk","is_ca":false},{"name":"Marc-André Carbonneau","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02503636111819494,"gpt":0.2806554597688282,"spread":0.2556190986506332,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009278741,0.001233376,0.0005582059,0.0002489874,0.0002666481,0.0006383726,0.001481478,0.0009852932,0.008162303],"category_scores_gemma":[0.002034144,0.0003246837,0.0009129012,0.0001227344,0.000462859,0.001068024,0.00211043,0.001949543,0.003424792],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002977506,"about_ca_system_score_gemma":0.0004375832,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000951498,"about_ca_topic_score_gemma":0.001431298,"domain_scores_codex":[0.9996499,0.00008412806,0.00001723451,0.0001222784,0.00009508589,0.00003131744],"domain_scores_gemma":[0.9996015,0.0002002753,0.00001910781,0.00009710409,0.00005237169,0.00002961468],"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":[0.0006150588,0.0002949603,0.0006591768,0.0003668607,0.0002239098,0.0005883347,0.000287843,0.3955076,0.09570426,0.008225786,0.01205413,0.4854721],"study_design_scores_gemma":[0.00004545154,0.0002125647,0.0001829201,0.00002293884,0.00002784206,0.000199974,0.00002253214,0.9639949,0.02413775,0.00497185,0.006154557,0.00002666806],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02470925,0.0006491448,0.9560316,0.0002326616,0.000404907,0.0001904366,0.0005085266,0.01143992,0.005833603],"genre_scores_gemma":[0.5836082,0.0006115011,0.3837751,0.0006375303,0.0002122818,0.0006794605,0.002953858,0.002201454,0.0253207],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.008162303,"threshold_uncertainty_score":0.0273056,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4297841830","doi":"10.21437/interspeech.2022-11066","title":"SoundChoice: Grapheme-to-Phoneme Models with Semantic Disambiguation","year":2022,"lang":"en","type":"article","venue":"Interspeech 2022","topic":"Speech Recognition and Synthesis","field":"Computer Science","cited_by":12,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Concordia University; Université de Montréal; Mila - Quebec Artificial Intelligence Institute","funders":"","keywords":"Grapheme; Computer science; Natural language processing; Artificial intelligence","authors":[{"name":"Artem Ploujnikov","is_ca":true},{"name":"Mirco Ravanelli","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01880117531733043,"gpt":0.232155006894166,"spread":0.2133538315768356,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005575834,0.001308712,0.0006085241,0.0004787887,0.0003677752,0.0008268433,0.001833068,0.00127374,0.004003125],"category_scores_gemma":[0.001447758,0.0004412702,0.0008079904,0.0006124548,0.0005325142,0.001855212,0.001327382,0.001916601,0.002283742],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000533464,"about_ca_system_score_gemma":0.001109724,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006591864,"about_ca_topic_score_gemma":0.01365571,"domain_scores_codex":[0.9997147,0.00005466382,0.0000132842,0.0001409102,0.00004734483,0.00002923458],"domain_scores_gemma":[0.9996916,0.0001473798,0.00001859935,0.00006208089,0.00005437696,0.00002597496],"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":[0.0006628533,0.0002638207,0.001635847,0.0002119101,0.0001702461,0.0003515829,0.0002343815,0.3623157,0.03156865,0.01081519,0.0176079,0.5741619],"study_design_scores_gemma":[0.00003786265,0.00007184809,0.0002491886,0.0000100298,0.00002272107,0.00004821633,0.00002671391,0.9771527,0.006528786,0.01276406,0.003067625,0.00002020129],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03268733,0.0008478563,0.9468369,0.0004124805,0.000312676,0.0001033529,0.000885305,0.01409129,0.003822862],"genre_scores_gemma":[0.6506605,0.0005174534,0.3298153,0.0006997507,0.0001209023,0.0002785399,0.003667812,0.0008715849,0.0133681],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.006591864,"threshold_uncertainty_score":0.01339185,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4297841482","doi":"10.21437/interspeech.2022-760","title":"Low-bit Shift Network for End-to-End Spoken Language Understanding","year":2022,"lang":"en","type":"article","venue":"Interspeech 2022","topic":"Neural Networks and Applications","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Huawei Technologies (Canada)","funders":"","keywords":"Computer science; End-to-end principle; Computer network","authors":[{"name":"Anderson R. Avila","is_ca":true},{"name":"Khalil Bibi","is_ca":true},{"name":"Rui Heng Yang","is_ca":true},{"name":"Xinlin Li","is_ca":true},{"name":"Chao Xing","is_ca":true},{"name":"Xiao Dong Chen","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02692692012041936,"gpt":0.2755905298264654,"spread":0.248663609706046,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004560396,0.0006837757,0.0003306044,0.0003470608,0.0002692783,0.0006331495,0.001060773,0.0007795544,0.008531108],"category_scores_gemma":[0.001233771,0.0002059548,0.0002296687,0.0003766178,0.000346823,0.001647309,0.000821456,0.001205746,0.002697292],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005505077,"about_ca_system_score_gemma":0.0008983724,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005031242,"about_ca_topic_score_gemma":0.01105824,"domain_scores_codex":[0.9997095,0.00004635135,0.00001982505,0.00008316289,0.0001082569,0.00003289113],"domain_scores_gemma":[0.9997552,0.00008038453,0.00002065056,0.00004524576,0.00008631579,0.00001221831],"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":[0.0005370138,0.0001819651,0.0005639637,0.0002196752,0.00005162003,0.000175772,0.0001657222,0.07001624,0.08528774,0.007320559,0.01279664,0.8226831],"study_design_scores_gemma":[0.00003213869,0.0001861089,0.0004558717,0.00003338171,0.00002593406,0.0001016628,0.00007040727,0.9307634,0.05141743,0.008600439,0.008285872,0.00002742356],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.06155954,0.002033676,0.9123696,0.0006290958,0.0002512204,0.000154814,0.0006523278,0.01441315,0.007936485],"genre_scores_gemma":[0.6684049,0.0007289307,0.3110766,0.0005919792,0.00008003786,0.0002217711,0.001847568,0.0003586963,0.0166896],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.008531108,"threshold_uncertainty_score":0.02853936,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null}]}