{"id":"W2152808281","doi":"10.1109/tnn.2007.912312","title":"Adaptive Importance Sampling to Accelerate Training of a Neural Probabilistic Language Model","year":2008,"lang":"en","type":"article","venue":"IEEE Transactions on Neural Networks","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":233,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Montréal","funders":"","keywords":"Computer science; Language model; Artificial neural network; Vocabulary; Probabilistic neural network; Speedup; Artificial intelligence; Probabilistic logic; Computation; Machine learning; Feedforward neural network; Sampling (signal processing); Statistical model; Training (meteorology); Backpropagation; Importance sampling; Time delay neural network; Algorithm; Statistics; Mathematics; Monte Carlo method","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":[0.001585378,0.0005938052,0.0007230479,0.0004931991,0.0002870628,0.000416031,0.001440164,0.0007244152,0.001907883],"category_scores_gemma":[0.006871984,0.0005024433,0.0004509148,0.0006185902,0.0004797015,0.001571464,0.00104709,0.001811918,0.0005297738],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005440061,"about_ca_system_score_gemma":0.0009135033,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004947551,"about_ca_topic_score_gemma":0.007286832,"domain_scores_codex":[0.9993852,0.0002446672,0.00003064425,0.00009257408,0.0001869738,0.00005993163],"domain_scores_gemma":[0.9979785,0.001421674,0.00007275648,0.0002142424,0.0002505785,0.00006222086],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002780461,0.0001704341,0.001164615,0.00009962964,0.00005380576,0.0001457789,0.0001291847,0.6906656,0.01374676,0.02257713,0.002846483,0.2681226],"study_design_scores_gemma":[0.000003942233,0.000007090905,0.00003231261,8.368856e-7,0.000001557699,0.00000713231,0.000001417201,0.997331,0.0007044728,0.001771098,0.0001379753,0.000001223942],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01263745,0.0001067266,0.9860241,0.00009055774,0.00003229607,0.00002218457,0.00001599635,0.0006980363,0.0003725924],"genre_scores_gemma":[0.3584071,0.0001908351,0.6390409,0.0001295552,0.0001077151,0.0001352105,0.0002067091,0.0001903799,0.0015916],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.004947551,"threshold_uncertainty_score":0.009837508,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0677840225644307,"score_gpt":0.2919126299227441,"score_spread":0.2241286073583134,"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."}}