{"id":"W4392678259","doi":"10.1609/aaai.v38i16.29789","title":"Cross-Modal and Uni-Modal Soft-Label Alignment for Image-Text Retrieval","year":2024,"lang":"en","type":"preprint","venue":"Proceedings of the AAAI Conference on Artificial Intelligence","topic":"Image Retrieval and Classification Techniques","field":"Computer Science","cited_by":37,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Ottawa","funders":"","keywords":"Modal; Computer science; Artificial intelligence; Image (mathematics); Computer vision; Information retrieval; Materials science","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":[0.001888205,0.001210436,0.001198253,0.002198965,0.001054121,0.001712432,0.002275623,0.001710653,0.007139828],"category_scores_gemma":[0.006067685,0.0004670313,0.001279875,0.00211609,0.001142638,0.004046321,0.002663351,0.002454656,0.00601191],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001123586,"about_ca_system_score_gemma":0.001368343,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003443207,"about_ca_topic_score_gemma":0.005631846,"domain_scores_codex":[0.9979272,0.0004535885,0.0001277308,0.0006278784,0.0006485954,0.0002149756],"domain_scores_gemma":[0.997775,0.0005732173,0.0002485529,0.000684795,0.0006071816,0.0001112069],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0006821731,0.0004337996,0.001573979,0.0004642755,0.0001468292,0.0001787287,0.0003170237,0.04386969,0.07418788,0.01208501,0.0199318,0.8461287],"study_design_scores_gemma":[0.00004926115,0.0001496354,0.00122796,0.00002883843,0.00006960192,0.0002613164,0.0001571804,0.9135444,0.05245198,0.02298019,0.009009469,0.00007020457],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01784421,0.0006510364,0.9712815,0.0002582326,0.0001021179,0.0001582788,0.0002807075,0.006384598,0.003039411],"genre_scores_gemma":[0.3945019,0.0005963566,0.5848121,0.00095834,0.0002778548,0.000512959,0.002796582,0.00109489,0.01444906],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007139828,"threshold_uncertainty_score":0.02388513,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07936549470850399,"score_gpt":0.3473004146126378,"score_spread":0.2679349199041339,"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."}}