{"id":"W2080021477","doi":"10.1016/j.neucom.2008.12.025","title":"Improving a statistical language model through non-linear prediction","year":2009,"lang":"en","type":"article","venue":"Neurocomputing","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":18,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Toronto","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Perplexity; Computer science; Feature (linguistics); Context (archaeology); Artificial intelligence; Language model; Word (group theory); Feature vector; Linear model; Term (time); Pattern recognition (psychology); Machine learning; Mathematics","routes":{"ca_aff":true,"ca_fund":true,"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":[0.001544956,0.001041921,0.001041457,0.0009710112,0.0006238593,0.00123201,0.001372369,0.001070307,0.002559405],"category_scores_gemma":[0.006320151,0.0006108979,0.001035394,0.001058607,0.000443333,0.003107836,0.001156146,0.002890675,0.002984259],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005678347,"about_ca_system_score_gemma":0.001300682,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006461081,"about_ca_topic_score_gemma":0.008669,"domain_scores_codex":[0.9989997,0.0003634437,0.00007176143,0.0002891335,0.0001901519,0.0000857456],"domain_scores_gemma":[0.9955349,0.003152374,0.0001647183,0.00036682,0.0006940379,0.00008717742],"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":[0.0005108361,0.0005441851,0.002991183,0.0002130482,0.0003066881,0.0003200889,0.0001736729,0.3562962,0.02221906,0.007187629,0.01049285,0.5987445],"study_design_scores_gemma":[0.000006280524,0.00002540066,0.0001270482,0.000002995052,0.00002303586,0.00002118697,0.000006979308,0.9952191,0.001802307,0.002503775,0.0002556612,0.000006274609],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04121412,0.0005327755,0.9506783,0.0008548487,0.0002453648,0.00004893825,0.0003694814,0.004950363,0.001105822],"genre_scores_gemma":[0.6208225,0.0007055994,0.3681593,0.0007382094,0.000352712,0.0002038728,0.00184776,0.0005697071,0.006600217],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.006461081,"threshold_uncertainty_score":0.01284695,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01265086074894661,"score_gpt":0.2861839819035844,"score_spread":0.2735331211546377,"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."}}