{"id":"W4386728930","doi":"10.1145/3604915.3608845","title":"Large Language Models are Competitive Near Cold-start Recommenders for Language- and Item-based Preferences","year":2023,"lang":"en","type":"article","venue":"","topic":"Recommender Systems and Techniques","field":"Computer Science","cited_by":112,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"","keywords":"Leverage (statistics); Computer science; Preference; Cold start (automotive); Natural language processing; Artificial intelligence; Dialog box; Task (project management); Recommender system; Variety (cybernetics); Preference learning; Language model; Machine learning; World Wide Web","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":[0.007870412,0.002241438,0.002866269,0.001587106,0.00141974,0.002346864,0.002133778,0.002977737,0.004082182],"category_scores_gemma":[0.02514066,0.0008654931,0.002072377,0.001418762,0.0007721253,0.004658337,0.001676981,0.0040404,0.005666769],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001188811,"about_ca_system_score_gemma":0.001235089,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.013581,"about_ca_topic_score_gemma":0.03304776,"domain_scores_codex":[0.9952753,0.002264074,0.0002709947,0.001192827,0.0007267647,0.0002701406],"domain_scores_gemma":[0.979529,0.01462672,0.00058273,0.003310126,0.001348924,0.0006025113],"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":[0.009149906,0.003465391,0.02512228,0.002366089,0.002045304,0.0006491166,0.001519159,0.1616813,0.0182621,0.009513449,0.06297245,0.7032536],"study_design_scores_gemma":[0.0003642071,0.001154421,0.005484933,0.0001278858,0.0003047853,0.0005275538,0.0003282209,0.9681373,0.005002339,0.01118983,0.007183214,0.0001952329],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4039614,0.01608513,0.5424731,0.002542084,0.0009203998,0.0006413202,0.004696473,0.01322021,0.01545975],"genre_scores_gemma":[0.7598918,0.001712021,0.2183218,0.001775322,0.0004740694,0.0002783826,0.007881863,0.0004670433,0.009197808],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.013581,"threshold_uncertainty_score":0.04162318,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04197899489953127,"score_gpt":0.2851031101057967,"score_spread":0.2431241152062654,"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."}}