{"id":"W6966925072","doi":"10.48448/dsy5-tw90","title":"Unifying Multimodal Retrieval via Document Screenshot Embedding","year":2024,"lang":"en","type":"other","venue":"Underline Science Inc.","topic":"","field":"","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Document retrieval; Parsing; Task (project management); Information extraction; Embedding; Vector space model; Document clustering; Image retrieval","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":[0.0007198036,0.001529998,0.001042312,0.002439734,0.0003082173,0.001351518,0.001040071,0.0009287873,0.006616835],"category_scores_gemma":[0.003847175,0.0002719334,0.0008787866,0.001502403,0.0006186819,0.004005841,0.002695417,0.0008627928,0.004331669],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004621556,"about_ca_system_score_gemma":0.0005588934,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002700959,"about_ca_topic_score_gemma":0.003785301,"domain_scores_codex":[0.9989796,0.0002299539,0.00007716438,0.0003016314,0.0003062628,0.0001052495],"domain_scores_gemma":[0.9985331,0.0004687125,0.00008684253,0.0004669784,0.0003704274,0.0000738578],"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":[0.0007188178,0.0003240269,0.0007160209,0.0008950901,0.0001214363,0.0002067549,0.0003203158,0.008376423,0.09020607,0.003882022,0.03497146,0.8592616],"study_design_scores_gemma":[0.0002859383,0.001534813,0.007672447,0.0002806408,0.0004103617,0.001924719,0.001322745,0.5825557,0.255878,0.03418385,0.1135655,0.0003851947],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.06678873,0.005559934,0.8791543,0.0008350024,0.0005097418,0.0008677121,0.006211727,0.02624439,0.01382858],"genre_scores_gemma":[0.3503392,0.003139737,0.5935875,0.0009508082,0.0004799133,0.0007997004,0.01926712,0.001631415,0.02980473],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.006616835,"threshold_uncertainty_score":0.0221355,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02728345031207594,"score_gpt":0.3439891022885956,"score_spread":0.3167056519765196,"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."}}