{"id":"W4401328745","doi":"10.1016/j.patrec.2024.08.003","title":"Enhancing zero-shot object detection with external knowledge-guided robust contrast learning","year":2024,"lang":"en","type":"article","venue":"Pattern Recognition Letters","topic":"Domain Adaptation and Few-Shot Learning","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":false,"ca_institutions":"Carleton University","funders":"National Natural Science Foundation of China","keywords":"Computer science; Discriminative model; Artificial intelligence; Robustness (evolution); Contrast (vision); Adversarial system; Machine learning; Consistency (knowledge bases); Object detection; Pattern recognition (psychology)","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":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0005417609,0.0002768075,0.0002093307,0.000376936,0.0003099875,0.0008683399,0.0002866366,0.00007544626,0.0001598682],"category_scores_gemma":[0.00005407847,0.0002602553,0.0001140789,0.0005330835,0.00005520999,0.0009328711,0.00006870332,0.0006154194,0.0009459375],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001423459,"about_ca_system_score_gemma":0.00005662205,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00004972308,"about_ca_topic_score_gemma":0.00006974064,"domain_scores_codex":[0.9978547,0.000274166,0.0003773562,0.0006778971,0.0003611714,0.0004547316],"domain_scores_gemma":[0.9991503,0.0002545853,0.0001292334,0.0002098865,0.0001143086,0.0001416841],"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":[0.00002067737,0.00003878416,0.0007614983,0.0001404545,0.00008448354,0.0001975031,0.003149832,0.003761521,0.1961517,0.00002493332,0.0002454866,0.7954231],"study_design_scores_gemma":[0.00353641,0.0007338711,0.009617581,0.003721791,0.0001804056,0.001787987,0.0007750064,0.7703112,0.1941567,0.0004169934,0.01217379,0.002588256],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1659989,0.0001341305,0.8304902,0.0005100851,0.000841066,0.0001920473,0.000001932061,0.0007214125,0.001110307],"genre_scores_gemma":[0.9889089,0.00001790999,0.008728473,0.001660056,0.0003292112,0.00006636403,0.00002010729,0.00004694651,0.0002220706],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.82291,"threshold_uncertainty_score":0.999985,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03665023461828474,"score_gpt":0.2559155425210343,"score_spread":0.2192653079027496,"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."}}