{"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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004107013,0.00008738061,0.0001464905,0.00005940889,0.0001506053,0.00003279773,0.0005907486,0.00002100463,0.00007890349],"category_scores_gemma":[0.00002760836,0.00007240992,0.00003360114,0.0002784432,0.00001593241,0.0001881495,0.0002243875,0.0001161449,7.362378e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00005612249,"about_ca_system_score_gemma":0.0001387536,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0003434014,"about_ca_topic_score_gemma":0.00005526759,"domain_scores_codex":[0.998912,0.00003592648,0.0001875932,0.0002978705,0.0003750795,0.0001915492],"domain_scores_gemma":[0.9992998,0.00005172049,0.00008558703,0.0004145941,0.0001063819,0.0000419505],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00003483608,0.0002282169,0.002677127,0.0001428183,0.00006738691,0.00001123402,0.002741365,0.063067,0.004521709,0.9119744,0.0004976733,0.01403627],"study_design_scores_gemma":[0.002111597,0.001161092,0.0003514422,0.00009885695,0.00002822778,0.00006388487,0.0003154956,0.8540879,0.005906288,0.1340312,0.001288133,0.0005557818],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03552918,0.00004898024,0.9600589,0.001396953,0.0001081035,0.0003791323,0.000004695055,0.0001088877,0.002365171],"genre_scores_gemma":[0.738776,3.876882e-7,0.2606446,0.0001888342,0.00001717071,0.00008494559,0.000001617017,0.000005919635,0.0002804715],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.791021,"threshold_uncertainty_score":0.2952791,"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."}}