{"id":"W2217916065","doi":"10.1016/j.dam.2015.09.018","title":"Similarity based deduplication with small data chunks","year":2015,"lang":"en","type":"article","venue":"Discrete Applied Mathematics","topic":"Cloud Data Security Solutions","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":false,"ca_institutions":"IBM (Canada)","funders":"","keywords":"Computer science; Data deduplication; Backup; Hash function; Scalability; Petabyte; Extension (predicate logic); Data mining; Theoretical computer science; Database; Big data; Programming language","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":[0.001949505,0.0006004462,0.001211926,0.001195876,0.001201797,0.001601351,0.001269956,0.0009372677,0.002900745],"category_scores_gemma":[0.009563007,0.0004257405,0.0005620107,0.001959784,0.0006842489,0.003616208,0.002328293,0.0008597259,0.0009022799],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004318766,"about_ca_system_score_gemma":0.001118075,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0004598051,"about_ca_topic_score_gemma":0.0008748436,"domain_scores_codex":[0.9983056,0.000327699,0.0002169903,0.0002885994,0.0007250062,0.0001361368],"domain_scores_gemma":[0.9908592,0.00244019,0.0004280362,0.004679808,0.001369188,0.0002235496],"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":[0.00404319,0.0006069197,0.006687845,0.0008312968,0.0002958852,0.001256036,0.0007856581,0.09818389,0.07402565,0.0388253,0.02379463,0.7506637],"study_design_scores_gemma":[0.0002456071,0.0008556886,0.002709731,0.0001028731,0.0001756574,0.002127808,0.0006760257,0.7980712,0.1206218,0.05702304,0.01728363,0.0001069469],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1711268,0.002632729,0.8152438,0.0007390453,0.001018374,0.0005642794,0.0007338249,0.002627612,0.005313565],"genre_scores_gemma":[0.6536655,0.0005141456,0.3373725,0.0002769157,0.0003650591,0.000177139,0.001080532,0.0002042464,0.006343991],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.002900745,"threshold_uncertainty_score":0.01031011,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1239699788852311,"score_gpt":0.2851796968870112,"score_spread":0.1612097180017801,"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."}}