Research map / Vision-language and multimodal / Cross-modal learning
MAD: Modality-Adaptive Decoding for Mitigating Cross-Modal Hallucinations in Multimodal Large Language Models
Sangyun Chung, Se Yeon Kim, Youngchae Chee, Yong Man Ro
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add MAD adaptive modality-weighted contrastive decoding | VideoLLaMA2-AV-7B · CMM overall | accuracy (%): 73.5 -> 81.3 | Improved |
| add MAD | Qwen2.5-Omni-7B · CMM overall | accuracy (%): 72.7 -> 81.4 | Improved |
| add MAD | VideoLLaMA2-AV-7B · AVHBench overall | accuracy (%): 77.4 -> 79.4 | Improved |
| add MAD | Qwen2.5-Omni-7B · AVHBench overall | accuracy (%): 76.9 -> 81.6 | Improved |
| add VCD-Extended uniform contrastive decoding | VideoLLaMA2-AV-7B · AVHBench overall | accuracy (%): 77.4 -> 70.4 | Hurt |
| add VCD-Extended to Qwen2.5-Omni | Qwen2.5-Omni-7B · CMM overall | accuracy (%): 72.7 -> 72.8 | No clear effect |
| add AVCD vs MAD | VideoLLaMA2-AV-7B · CMM overall | accuracy (%): 73.5 -> 75.8 (AVCD); 81.3 (MAD) | Mixed |
| use adaptive weights instead of uniform weights (ablation) | MAD · CMM overall | accuracy (%): 79.4 -> 81.3 | Improved |
| use soft weighting instead of argmax branch selection (ablation) | MAD · CMM overall | accuracy (%): 78.7 -> 81.3 | Improved |
| add audio weight w_a (ablation) | MAD · CMM overall | accuracy (%): 78.0 -> 81.3 | Improved |
| add visual weight w_v (ablation) | MAD · CMM overall | accuracy (%): 78.3 -> 81.3 | Improved |
| add audio-visual weight w_av (ablation) | MAD · CMM overall | accuracy (%): 78.9 -> 81.3 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- AVCD — AVCD: Mitigating Hallucinations in Audio-Visual Large Language Models through Contrastive Decoding (NeurIPS 2025) · better on 4 of 4 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- The Curse of Multi-Modalities: Evaluating Hallucinations of Large Multimodal Models across Language, Visual, and Audio (NeurIPS 2025 · similarity 0.93)
- MoD-DPO: Towards Mitigating Cross-modal Hallucinations in Omni LLMs using Modality Decoupled Preference Optimization (CVPR 2026 · similarity 0.93)
- AVCD: Mitigating Hallucinations in Audio-Visual Large Language Models through Contrastive Decoding (NeurIPS 2025 · similarity 0.93)
- AVHBench: A Cross-Modal Hallucination Benchmark for Audio-Visual Large Language Models (ICLR 2025 · similarity 0.93)
- Robust Multimodal Large Language Models Against Modality Conflict (ICML 2025 · similarity 0.93)
- CODE: Contrasting Self-generated Description to Combat Hallucination in Large Multi-modal Models (NeurIPS 2024 · similarity 0.92)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
