{"id":"W4389519126","doi":"10.18653/v1/2023.emnlp-main.448","title":"OssCSE: Overcoming Surface Structure Bias in Contrastive Learning for Unsupervised Sentence Embedding","year":2023,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"Queen's University","funders":"","keywords":"Natural language processing; Embedding; Sentence; Computer science; Artificial intelligence; Linguistics; Philosophy","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":[0.001897662,0.001543153,0.001310354,0.001261122,0.0008210345,0.00128321,0.002350134,0.001300276,0.008199535],"category_scores_gemma":[0.006338375,0.00065668,0.001032884,0.001238141,0.0009486637,0.005000766,0.003583896,0.002560169,0.003422891],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006593302,"about_ca_system_score_gemma":0.001155224,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003725331,"about_ca_topic_score_gemma":0.01219979,"domain_scores_codex":[0.998929,0.0004269026,0.00006215672,0.0003070307,0.0001973171,0.00007765433],"domain_scores_gemma":[0.9974579,0.001273738,0.00009066924,0.0006889905,0.000377567,0.0001111521],"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":[0.000464976,0.0002839066,0.001781174,0.000328251,0.0002032121,0.0001572091,0.0004904954,0.03195637,0.01978182,0.02754268,0.04155827,0.8754516],"study_design_scores_gemma":[0.0001414422,0.0001750789,0.0006062681,0.00003001137,0.00005488483,0.0001031885,0.0001354883,0.9281309,0.01164786,0.0475901,0.0113486,0.0000360434],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02434839,0.0007161218,0.9586677,0.0003085305,0.0002498284,0.0001694937,0.0008987148,0.01186058,0.00278063],"genre_scores_gemma":[0.2755433,0.0004628702,0.7020111,0.0004890669,0.0002832209,0.0005519893,0.007438653,0.002713123,0.01050679],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.008199535,"threshold_uncertainty_score":0.02743018,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05164863126802949,"score_gpt":0.3009311128956684,"score_spread":0.2492824816276389,"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."}}