AVCD: Mitigating Hallucinations in Audio-Visual Large Language Models through Contrastive Decoding

• URL: http://arxiv.org/abs/2505.20862v2
• Date: Tue, 30 Sep 2025 12:14:50 GMT
• Title: AVCD: Mitigating Hallucinations in Audio-Visual Large Language Models through Contrastive Decoding
• Authors: Chaeyoung Jung, Youngjoon Jang, Joon Son Chung,
• Abstract summary: We propose Audio-Visual Contrastive Decoding (AVCD) to model trimodal interactions and suppress hallucinations in large language models (MLLMs)<n>To improve efficiency, we introduce entropy-guided adaptive decoding, which skips unnecessary decoding steps based on the model's confidence in its predictions.
• Score: 38.71842806548495
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Hallucination remains a major challenge in multimodal large language models (MLLMs). To address this, various contrastive decoding (CD) methods have been proposed that contrasts original logits with hallucinated logits generated from perturbed inputs. While CD has shown promise in vision-language models (VLMs), it is not well-suited for AV-LLMs, where hallucinations often emerge from both unimodal and cross-modal combinations involving audio, video, and language. These intricate interactions call for a more adaptive and modality-aware decoding strategy. In this paper, we propose Audio-Visual Contrastive Decoding (AVCD)-a novel, training-free decoding framework designed to model trimodal interactions and suppress modality-induced hallucinations in AV-LLMs. Unlike previous CD methods in VLMs that corrupt a fixed modality, AVCD leverages attention distributions to dynamically identify less dominant modalities and applies attentive masking to generate perturbed output logits. To support CD in a trimodal setting, we also reformulate the original CD framework to jointly handle audio, visual, and textual inputs. Finally, to improve efficiency, we introduce entropy-guided adaptive decoding, which selectively skips unnecessary decoding steps based on the model's confidence in its predictions. Extensive experiments demonstrate that AVCD consistently outperforms existing decoding methods. Especially, on the AVHBench dataset, it improves accuracy by 2% for VideoLLaMA2 and 7% for video-SALMONN, demonstrating strong robustness and generalizability. Our code is available at https://github.com/kaistmm/AVCD.

Related papers

• MACD: Model-Aware Contrastive Decoding via Counterfactual Data [0.0]
  Video language models (Video-LLMs) are prone to hallucinations, often generating plausible but ungrounded content when visual evidence is weak, ambiguous, or biased.<n>We propose Model-aware Counterfactual Data based Contrastive Decoding (MACD), a new inference strategy that combines model-guided counterfactual construction with decoding.<n>Our approach uses the Video-LLM's own feedback to identify object regions most responsible for hallucination, generating targeted counterfactual inputs at the object level rather than arbitrary frame or temporal modifications.
  arXiv  Detail & Related papers  (2026-02-02T07:21:02Z)
• MAD: Modality-Adaptive Decoding for Mitigating Cross-Modal Hallucinations in Multimodal Large Language Models [45.58164536222542]
  Multimodal Large Language Models (MLLMs) suffer from cross-modal hallucinations, where one modality inappropriately influences generation about another, leading to fabricated output.<n>We propose Modality-Adaptive Decoding (MAD), a training-free method that adaptively weights modality-specific decoding branches based on task requirements.<n>Our approach demonstrates that explicit modality awareness through self-assessment is crucial for robust multimodal reasoning, offering a principled extension to existing contrastive decoding methods.
  arXiv  Detail & Related papers  (2026-01-29T02:30:32Z)
• MaskCD: Mitigating LVLM Hallucinations by Image Head Masked Contrastive Decoding [53.068815533016355]
  We propose image head Masked Contrastive Decoding (MaskCD) for large vision-language models (LVLMs)<n>Our approach utilizes the "image heads" in LVLMs, masking them to construct contrastive samples for contrastive decoding.<n>The results demonstrate that MaskCD effectively alleviates the phenomenon of hallucinations and retains the general capabilities of LVLMs.
  arXiv  Detail & Related papers  (2025-10-03T07:59:16Z)
• Implicit Counterfactual Learning for Audio-Visual Segmentation [50.69377287012591]
  We propose the implicit counterfactual framework (ICF) to achieve unbiased cross-modal understanding.<n>Due to the lack of semantics, heterogeneous representations may lead to erroneous matches.<n>We introduce the multi-granularity implicit text (MIT) involving video-, segment- and frame-level as the bridge to establish the modality-shared space.
  arXiv  Detail & Related papers  (2025-07-28T11:46:35Z)
• ASCD: Attention-Steerable Contrastive Decoding for Reducing Hallucination in MLLM [12.091189146069198]
  Multimodal Large Language Model (MLLM) often suffer from hallucinations.<n>They over-rely on partial cues and generate incorrect responses.<n>Recent methods like Visual Contrastive Decoding (VCD) and Instruction Contrastive Decoding (ICD) have been proposed to mitigate hallucinations.
  arXiv  Detail & Related papers  (2025-06-17T17:58:11Z)
• Fork-Merge Decoding: Enhancing Multimodal Understanding in Audio-Visual Large Language Models [13.887164304514101]
  The goal of this work is to enhance balanced multimodal understanding in audio-visual large language models (AV-LLMs)<n>In current AV-LLMs, audio and video features are typically processed jointly in the decoder.<n>We propose Fork-Merge Decoding (FMD), a simple yet effective inference-time strategy that requires no additional training or architectural modifications.
  arXiv  Detail & Related papers  (2025-05-27T08:22:56Z)
• From Alignment to Advancement: Bootstrapping Audio-Language Alignment with Synthetic Data [55.2480439325792]
  Audio-aware large language models (ALLMs) have recently made great strides in understanding and processing audio inputs.<n>These models are typically adapted from text-based large language models (LLMs) through additional training on audio-related tasks.<n>We propose a data generation framework that produces contrastive-like training data, designed to enhance ALLMs' ability to differentiate between present and absent sounds.
  arXiv  Detail & Related papers  (2025-05-26T16:08:41Z)
• AVadCLIP: Audio-Visual Collaboration for Robust Video Anomaly Detection [57.649223695021114]
  We present a novel weakly supervised framework that leverages audio-visual collaboration for robust video anomaly detection.<n>Our framework demonstrates superior performance across multiple benchmarks, with audio integration significantly boosting anomaly detection accuracy.
  arXiv  Detail & Related papers  (2025-04-06T13:59:16Z)
• Multi-Task Corrupted Prediction for Learning Robust Audio-Visual Speech Representation [23.406334722946163]
  We propose CAV2vec, a novel self-supervised speech representation learning framework to handle audio-visual joint corruption.<n>We suggest a unimodal multi-task learning, which distills cross-modal knowledge and aligns the corrupted modalities.<n>Our experiments on robust AVSR benchmarks demonstrate that the corrupted representation learning method significantly enhances recognition accuracy.
  arXiv  Detail & Related papers  (2025-01-23T05:11:19Z)
• Mitigating Hallucination for Large Vision Language Model by Inter-Modality Correlation Calibration Decoding [66.06337890279839]
  Large vision-language models (LVLMs) have shown remarkable capabilities in visual-language understanding for downstream multi-modal tasks.<n>LVLMs still suffer from generating hallucinations in complex generation tasks, leading to inconsistencies between visual inputs and generated content.<n>We propose an Inter-Modality Correlation Decoding (IMCCD) method to mitigate hallucinations in LVLMs in a training-free manner.
  arXiv  Detail & Related papers  (2025-01-03T17:56:28Z)
• Mitigating Object Hallucinations in Large Vision-Language Models through Visual Contrastive Decoding [125.05295513481035]
  We introduce Visual Contrastive Decoding (VCD), a simple and training-free method that contrasts output distributions derived from original and distorted visual inputs. The proposed VCD effectively reduces the over-reliance on statistical bias and unimodal priors, two essential causes of object hallucinations. Our experiments show that VCD, without either additional training or the usage of external tools, significantly mitigates the object hallucination issue across different LVLM families.
  arXiv  Detail & Related papers  (2023-11-28T16:26:35Z)
• STELLA: Continual Audio-Video Pre-training with Spatio-Temporal Localized Alignment [61.83340833859382]
  Continuously learning a variety of audio-video semantics over time is crucial for audio-related reasoning tasks. This is a nontemporal problem and poses two critical challenges: sparse-temporal correlation between audio-video pairs and multimodal correlation overwriting that forgets audio-video relations. We propose a continual audio-video pre-training method with two novel ideas.
  arXiv  Detail & Related papers  (2023-10-12T10:50:21Z)
• MM-Diffusion: Learning Multi-Modal Diffusion Models for Joint Audio and Video Generation [70.74377373885645]
  We propose the first joint audio-video generation framework that brings engaging watching and listening experiences simultaneously. MM-Diffusion consists of a sequential multi-modal U-Net for a joint denoising process by design. Experiments show superior results in unconditional audio-video generation, and zero-shot conditional tasks.
  arXiv  Detail & Related papers  (2022-12-19T14:11:52Z)
• Contrastive Audio-Visual Masked Autoencoder [85.53776628515561]
  Contrastive Audio-Visual Masked Auto-Encoder (CAV-MAE) Our fully self-supervised pretrained CAV-MAE achieves a new SOTA accuracy of 65.9% on VGGSound.
  arXiv  Detail & Related papers  (2022-10-02T07:29:57Z)
• Modality-Aware Contrastive Instance Learning with Self-Distillation for Weakly-Supervised Audio-Visual Violence Detection [14.779452690026144]
  We propose a modality-aware contrastive instance learning with self-distillation (MACIL-SD) strategy for weakly-supervised audio-visual learning. Our framework outperforms previous methods with lower complexity on the large-scale XD-Violence dataset.
  arXiv  Detail & Related papers  (2022-07-12T12:42:21Z)
• VLMixer: Unpaired Vision-Language Pre-training via Cross-Modal CutMix [59.25846149124199]
  This paper proposes a data augmentation method, namely cross-modal CutMix. CMC transforms natural sentences from the textual view into a multi-modal view. By attaching cross-modal noise on uni-modal data, it guides models to learn token-level interactions across modalities for better denoising.
  arXiv  Detail & Related papers  (2022-06-17T17:56:47Z)
• Switching Variational Auto-Encoders for Noise-Agnostic Audio-visual Speech Enhancement [26.596930749375474]
  We introduce the use of a latent sequential variable with Markovian dependencies to switch between different VAE architectures through time. We derive the corresponding variational expectation-maximization algorithm to estimate the parameters of the model and enhance the speech signal.
  arXiv  Detail & Related papers  (2021-02-08T11:45:02Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.