{"id":"W4392864334","doi":"10.1145/3649896","title":"Realizing Efficient On-Device Language-based Image Retrieval","year":2024,"lang":"en","type":"article","venue":"ACM Transactions on Multimedia Computing Communications and Applications","topic":"Multimodal Machine Learning Applications","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto; Centre for Social Innovation","funders":"","keywords":"Computer science; Information retrieval; Image retrieval; Latency (audio); Ranking (information retrieval); Modal; Language model; Deep learning; Context (archaeology); Artificial intelligence; Image (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":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","sts"],"consensus_categories":[],"category_scores_codex":[0.0005630614,0.0003161331,0.0002349779,0.000490926,0.001407275,0.0005229971,0.002588872,0.0001221921,0.00001872055],"category_scores_gemma":[0.00008834914,0.0003197179,0.0001453463,0.001596399,0.0002697748,0.0001489468,0.0001622593,0.0009525082,0.0003776212],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001309918,"about_ca_system_score_gemma":0.0001442395,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002002178,"about_ca_topic_score_gemma":0.0000126558,"domain_scores_codex":[0.9977533,0.0002173537,0.0004865892,0.0008286006,0.0003449214,0.0003691922],"domain_scores_gemma":[0.9916998,0.00337153,0.0001183504,0.004444713,0.0001467469,0.0002188556],"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":[0.00001667639,0.001211524,0.00003986481,0.0001338044,0.0001140512,0.00000463066,0.002602668,0.0235949,0.01324713,0.06105962,0.00009563346,0.8978795],"study_design_scores_gemma":[0.0003352388,0.00006746731,0.0006148394,0.0001593842,0.00004229713,0.00001256082,0.0001526375,0.9846281,0.002149586,0.0004922606,0.01099981,0.0003457864],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.004185896,0.0005338694,0.977363,0.01406582,0.00009517479,0.0008969984,0.00005705989,0.0015457,0.001256427],"genre_scores_gemma":[0.6191716,0.00007215189,0.3799933,0.0003581358,0.00005638073,0.000226385,0.00004790532,0.00003386847,0.00004033411],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9610332,"threshold_uncertainty_score":0.9999255,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02008231110073879,"score_gpt":0.3266956285550984,"score_spread":0.3066133174543596,"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."}}