{"id":"W4280612849","doi":"10.18653/v1/2022.acl-short.45","title":"Problems with Cosine as a Measure of Embedding Similarity for High Frequency Words","year":2022,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":54,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada; National Science Foundation","keywords":"Cosine similarity; Similarity (geometry); Polysemy; Word (group theory); Natural language processing; Trigonometric functions; Embedding; Computer science; Argument (complex analysis); Artificial intelligence; Mathematics; Conjecture; Discrete cosine transform; TRACE (psycholinguistics); Speech recognition; Pattern recognition (psychology); Linguistics; Pure mathematics","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.03232665,0.001443487,0.003146399,0.005434036,0.002359569,0.006237093,0.004103728,0.003178323,0.00203251],"category_scores_gemma":[0.1875594,0.001257714,0.001331561,0.01140983,0.004738429,0.01028347,0.004297268,0.005045828,0.001755088],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002697545,"about_ca_system_score_gemma":0.001396684,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005865411,"about_ca_topic_score_gemma":0.003048641,"domain_scores_codex":[0.9463965,0.02757302,0.006068022,0.009639301,0.00951334,0.0008098026],"domain_scores_gemma":[0.838033,0.1232817,0.008644593,0.02039782,0.008679267,0.0009637548],"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":[0.0007847045,0.0003240744,0.04245194,0.002760476,0.001324825,0.0005365004,0.003939576,0.06905963,0.009293888,0.2957718,0.03474575,0.5390069],"study_design_scores_gemma":[0.00007775119,0.0003012068,0.02044461,0.0004874762,0.0001438789,0.001711378,0.002064078,0.3168943,0.007764201,0.6267745,0.02301651,0.000320089],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.05839867,0.006480102,0.92264,0.004206839,0.0007548587,0.000275369,0.00119391,0.0007269094,0.005323303],"genre_scores_gemma":[0.6896952,0.002828836,0.2976128,0.0012496,0.001396493,0.0009567851,0.00259661,0.0004971903,0.003166529],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.03232665,"threshold_uncertainty_score":0.1709617,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02659254907497569,"score_gpt":0.2515398599478576,"score_spread":0.2249473108728819,"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."}}