{"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":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.001085132,0.0003138994,0.0003383078,0.0007886963,0.0002117915,0.0007187727,0.001994792,0.0001046305,0.00013039],"category_scores_gemma":[0.0001139858,0.0002512896,0.00004927042,0.001803138,0.0006130313,0.0009709439,0.0009865016,0.0005104592,0.0001647425],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00008539069,"about_ca_system_score_gemma":0.0002130646,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00007369381,"about_ca_topic_score_gemma":0.00005222434,"domain_scores_codex":[0.9966556,0.00005795184,0.0003815517,0.001126301,0.001263352,0.0005152145],"domain_scores_gemma":[0.9986514,0.00007993244,0.0003862602,0.0006134887,0.0001214691,0.0001474926],"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":[0.00001181371,0.0002780585,0.00007929732,0.0008338686,0.0001423843,0.00005745957,0.001224573,0.0007656118,0.01130821,0.04566874,0.1342673,0.8053627],"study_design_scores_gemma":[0.000821673,0.0007122303,0.00003907384,0.002944226,0.00006401797,0.00001695845,0.0003875521,0.9049979,0.004420578,0.002920445,0.08148397,0.001191406],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0001632146,0.0003371764,0.805418,0.0005987749,0.0005317614,0.0003077581,0.00001907483,0.0004421839,0.1921821],"genre_scores_gemma":[0.0760217,0.0001596639,0.5961229,0.0002141779,0.0003778192,0.00002640631,0.0001626556,0.000295757,0.3266189],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.9042323,"threshold_uncertainty_score":0.9999939,"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."}}