{"id":"W6948185420","doi":"10.48448/4gcb-xp21","title":"HELPD: Mitigating Hallucination of LVLMs by Hierarchical Feedback Learning with Vision-enhanced Penalty Decoding","year":2024,"lang":"en","type":"other","venue":"Underline Science Inc.","topic":"Data Management and Algorithms","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"","keywords":"Decoding methods; Object (grammar); Sentence; Image (mathematics); Association (psychology); Encoding (memory); Modality (human–computer interaction); Generative grammar; Channel (broadcasting)","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":[0.001369412,0.001322201,0.0008257783,0.0006034615,0.0003904257,0.0009199165,0.001977048,0.001348654,0.003668737],"category_scores_gemma":[0.006863969,0.0003559864,0.0006068098,0.0004568952,0.0007762656,0.001542154,0.002717026,0.002040775,0.001363311],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007252072,"about_ca_system_score_gemma":0.001196233,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004638283,"about_ca_topic_score_gemma":0.007959769,"domain_scores_codex":[0.9992174,0.0002552092,0.00004953834,0.0001743327,0.0002181467,0.00008540858],"domain_scores_gemma":[0.9981286,0.001011303,0.0001449006,0.000257095,0.0003204409,0.0001376073],"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":[0.0005167814,0.0003104152,0.001753719,0.0002938776,0.0001032544,0.0003594201,0.0002327107,0.271367,0.02568533,0.006097684,0.01540352,0.6778763],"study_design_scores_gemma":[0.00003050113,0.00008198625,0.0001290121,0.000009253329,0.000009849818,0.00004918572,0.00002181278,0.9898197,0.005724098,0.003179861,0.0009332408,0.00001151473],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0363133,0.0008734607,0.9515455,0.0006126598,0.0001749212,0.0001063824,0.0002352821,0.007876118,0.002262415],"genre_scores_gemma":[0.6700521,0.0003315684,0.3174008,0.001150855,0.0001645096,0.0002236636,0.001379244,0.0008981224,0.008399191],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004638283,"threshold_uncertainty_score":0.01227313,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.008228171800360307,"score_gpt":0.2692292689642038,"score_spread":0.2610010971638436,"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."}}