{"id":"W2112908352","doi":"10.1016/j.tcs.2014.05.005","title":"New space/time tradeoffs for top- k document retrieval on sequences","year":2014,"lang":"en","type":"article","venue":"Theoretical Computer Science","topic":"Algorithms and Data Compression","field":"Computer Science","cited_by":16,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Waterloo","funders":"Fondo Nacional de Desarrollo Científico y Tecnológico","keywords":"Document retrieval; Computer science; Space (punctuation); Information retrieval; Combinatorics; Mathematics","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":[0.008816387,0.002071754,0.00268066,0.004639383,0.001762235,0.00613607,0.003854572,0.003580687,0.01460745],"category_scores_gemma":[0.06420129,0.00121435,0.001005941,0.007002582,0.002494219,0.0185024,0.004136935,0.003440441,0.003891533],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.004198353,"about_ca_system_score_gemma":0.002948422,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002612919,"about_ca_topic_score_gemma":0.004451335,"domain_scores_codex":[0.9895548,0.003368799,0.001011952,0.0009299397,0.004138984,0.0009955752],"domain_scores_gemma":[0.9523143,0.03506189,0.001709874,0.005751139,0.003600839,0.001561913],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0042815,0.0007983999,0.002455313,0.00124605,0.0001745735,0.0002255038,0.0006821005,0.1717075,0.0380253,0.1590838,0.02398112,0.5973388],"study_design_scores_gemma":[0.0001520235,0.0003324334,0.0006609624,0.00008066284,0.00009556149,0.0004598896,0.0002208187,0.8687865,0.01011988,0.1150236,0.003976028,0.00009163847],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.08620887,0.01406729,0.8785967,0.003170483,0.0006522165,0.0002775748,0.001111208,0.003048237,0.01286726],"genre_scores_gemma":[0.4084955,0.004508425,0.5719308,0.000649229,0.001545547,0.0003531489,0.001348781,0.001256038,0.009912517],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01460745,"threshold_uncertainty_score":0.04886681,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.008556120600705158,"score_gpt":0.2540201630994239,"score_spread":0.2454640424987188,"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."}}