{"id":"W4385565351","doi":"10.18653/v1/2023.acl-long.99","title":"Precise Zero-Shot Dense Retrieval without Relevance Labels","year":2023,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":233,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Relevance (law); Computer science; Similarity (geometry); Relevance feedback; Embedding; Vector space model; Artificial intelligence; Information retrieval; Encoder; Zero (linguistics); Natural language processing; Document retrieval; Vector space; Encoding (memory); Language model; Image retrieval; Pattern recognition (psychology); Mathematics; Image (mathematics)","routes":{"ca_aff":true,"ca_fund":true,"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":[0.001621794,0.0008307018,0.001286729,0.0008250775,0.000632689,0.001264828,0.002247638,0.001122285,0.004331301],"category_scores_gemma":[0.006049732,0.0005486913,0.0005497888,0.0008188074,0.001278667,0.005105434,0.003087473,0.001366953,0.002513336],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000774376,"about_ca_system_score_gemma":0.001243583,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003943651,"about_ca_topic_score_gemma":0.007530736,"domain_scores_codex":[0.9988398,0.0002458384,0.00007671615,0.0003577325,0.0003377027,0.0001422949],"domain_scores_gemma":[0.997454,0.0009534964,0.0001629694,0.0009844882,0.0003374449,0.0001075986],"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":[0.0009339598,0.0005870601,0.001682117,0.0006989691,0.0001285203,0.0002655615,0.0006928482,0.0389984,0.07328229,0.02010466,0.01372411,0.8489015],"study_design_scores_gemma":[0.000203441,0.0009475786,0.002181564,0.00007830985,0.000130765,0.001205066,0.0005656751,0.8439937,0.08150802,0.05312939,0.01591879,0.0001376615],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05352547,0.0010467,0.9326018,0.0002709103,0.0000831902,0.0002569932,0.0003688929,0.007766546,0.004079486],"genre_scores_gemma":[0.587177,0.0005957969,0.3942788,0.000590756,0.0001326958,0.0002461056,0.002084445,0.000523384,0.01437084],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004331301,"threshold_uncertainty_score":0.01448959,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05173755000251323,"score_gpt":0.2928313881534951,"score_spread":0.2410938381509819,"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."}}