{"id":"W4283790576","doi":"10.1609/aaai.v36i10.21340","title":"From Fully Trained to Fully Random Embeddings: Improving Neural Machine Translation with Compact Word Embedding Tables","year":2022,"lang":"en","type":"article","venue":"Proceedings of the AAAI Conference on Artificial Intelligence","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University; Thomson Reuters (Canada)","funders":"","keywords":"Computer science; Embedding; Machine translation; Word (group theory); Word embedding; Natural language processing; Artificial intelligence; Context (archaeology); Artificial neural network; Task (project management); Key (lock); Language model; Speech recognition; Mathematics","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":[0.001331958,0.001297357,0.001110398,0.0006773446,0.0003126698,0.001277609,0.001284192,0.001031711,0.003829465],"category_scores_gemma":[0.007882078,0.0005212794,0.0006318484,0.001173715,0.0004968648,0.004846775,0.001430341,0.001405321,0.003698566],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004584374,"about_ca_system_score_gemma":0.0006869913,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00213402,"about_ca_topic_score_gemma":0.003352821,"domain_scores_codex":[0.999189,0.0003142889,0.00009158217,0.0001856324,0.0001630743,0.00005647581],"domain_scores_gemma":[0.99773,0.001117455,0.0001505517,0.0005784644,0.0003740689,0.00004954599],"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":[0.0005528746,0.0002729267,0.001076911,0.0002934248,0.000132431,0.0002821993,0.0001754379,0.3291103,0.01679335,0.009971699,0.01204964,0.6292888],"study_design_scores_gemma":[0.00003872982,0.0001302134,0.00018118,0.00001967001,0.00002420842,0.00007566519,0.0000282489,0.9816976,0.005947596,0.009344322,0.00249516,0.0000174393],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1300249,0.002354999,0.8464744,0.0006742985,0.0004358967,0.0001210198,0.000913185,0.0147108,0.004290365],"genre_scores_gemma":[0.5975263,0.001036682,0.3869179,0.0005391847,0.0002316138,0.0002804088,0.005252957,0.001008608,0.007206345],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003829465,"threshold_uncertainty_score":0.01281089,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03792903244196088,"score_gpt":0.2924550729584006,"score_spread":0.2545260405164397,"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."}}