{"id":"W191017350","doi":"","title":"Enlarging Paraphrase Collections through Generalization and Instantiation","year":2012,"lang":"en","type":"article","venue":"NPARC","topic":"Topic Modeling","field":"Computer Science","cited_by":17,"is_retracted":false,"has_abstract":true,"ca_institutions":"National Research Council Canada","funders":"Japan Society for the Promotion of Science; National Research Council Canada","keywords":"Paraphrase; Computer science; Natural language processing; Generalization; Artificial intelligence; Quality (philosophy); Parallel corpora; Exploit; Machine translation; Mathematics","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":true,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003371632,0.00148169,0.001621219,0.006173621,0.001015297,0.001915876,0.002647035,0.001123055,0.005016697],"category_scores_gemma":[0.01272536,0.001087732,0.00167096,0.004872291,0.001005238,0.005986513,0.004325228,0.002019504,0.002815678],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006244539,"about_ca_system_score_gemma":0.0009516444,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001371508,"about_ca_topic_score_gemma":0.003431099,"domain_scores_codex":[0.9970351,0.0009352652,0.0003383172,0.000944952,0.0005878348,0.0001585235],"domain_scores_gemma":[0.9903349,0.004332067,0.0004757884,0.003845125,0.0007947658,0.0002174018],"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":[0.0002585228,0.0005588956,0.006496619,0.0007464669,0.0002548295,0.0008150338,0.00184439,0.009824048,0.05569328,0.009741295,0.01224236,0.9015243],"study_design_scores_gemma":[0.0001933047,0.0007774126,0.02209344,0.0004706305,0.0008270806,0.005182701,0.002475752,0.6576498,0.1236408,0.08193064,0.1044821,0.000276312],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.09050915,0.001376669,0.881353,0.0005413157,0.00005658151,0.001120902,0.001611037,0.01465168,0.00877957],"genre_scores_gemma":[0.2611343,0.0008358562,0.7262029,0.0002315798,0.00009580914,0.0004939243,0.005624616,0.001206482,0.004174476],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.006173621,"threshold_uncertainty_score":0.01783109,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02717435825852078,"score_gpt":0.255400939592151,"score_spread":0.2282265813336302,"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."}}