{"id":"W4385573577","doi":"10.18653/v1/2022.emnlp-main.680","title":"Specializing Multi-domain NMT via Penalizing Low Mutual Information","year":2022,"lang":"en","type":"article","venue":"","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Kootenay Association for Science & Technology","funders":"Institute for Information and Communications Technology Promotion; Ministry of Science and ICT, South Korea; Korea Advanced Institute of Science and Technology; National Research Foundation of Korea; National Research Foundation","keywords":"Computer science; Domain (mathematical analysis); Mutual information; Artificial intelligence; Machine translation; Domain model; Task (project management); Machine learning; Domain knowledge; Mathematics; Engineering","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002391979,0.001387671,0.0011874,0.0007350064,0.000678691,0.001039571,0.001365073,0.001921271,0.001809096],"category_scores_gemma":[0.006378352,0.0004750592,0.000880764,0.0007810393,0.001105224,0.002369503,0.00237811,0.002177311,0.0007234488],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008691985,"about_ca_system_score_gemma":0.001218838,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001983356,"about_ca_topic_score_gemma":0.003382891,"domain_scores_codex":[0.9989091,0.0004377667,0.00006981014,0.0002665597,0.0002278517,0.00008897199],"domain_scores_gemma":[0.997429,0.001439652,0.000295996,0.0003695416,0.0003337387,0.0001321053],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0002642021,0.0001756029,0.001639385,0.000253643,0.0002110755,0.0003562318,0.0001831831,0.7820624,0.02021522,0.02576036,0.005510614,0.1633682],"study_design_scores_gemma":[0.000007575179,0.00003150788,0.0001149236,0.000005663238,0.00001219048,0.00005066528,0.000009400646,0.9926549,0.001922111,0.004513068,0.0006707695,0.000007186516],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03219521,0.0004667734,0.9636688,0.0003836107,0.00008474762,0.00004795982,0.00005311418,0.0007966167,0.002303284],"genre_scores_gemma":[0.6951377,0.0003473779,0.2974278,0.000624242,0.0002465619,0.0002050357,0.000525053,0.0005270366,0.004959267],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.002391979,"threshold_uncertainty_score":0.01265013,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0109811535054097,"score_gpt":0.2491207537971779,"score_spread":0.2381396002917681,"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."}}