ACL Findings2026
Temporal Contrastive Decoding: A Training-Free Method for Large Audio-Language Models
Yanda Li, Yuhan Liu, Zirui Song, Yunchao Wei, Martin Takác, Salem Lahlou
Abstract
Large audio-language models (LALMs) generalize across speech, sound, and music, but unified decoders can exhibit a temporal smoothing bias: transient acoustic cues may be underutilized in favor of temporally smooth context that is better supported by language priors, leading to less specific audio-grounded outputs. We propose Temporal Contrastive Decoding (TCD), a training-free decoding method for unified LALMs that mitigates this effect at inference time. TCD constructs a temporally blurred slow-path view by smoothing the input waveform and re-encoding it, then contrasts next-token logits from the original and slow-path views. The contrastive signal is applied as a token-level logit update restricted to a small candidate set. A self-normalized stability score sets the blur window and update scale, and a step-wise gate based on uncertainty and audio reliance activates the update only when needed. Experiments on MMAU and AIR-Bench show consistent improvements on strong unified LALMs. We further conduct ablations and an architectural applicability study to analyze the contributions of key components and how TCD behaves across large audio-language model designs.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
Lineage
Builds on26
- Plug and Play Language Models: A Simple Approach to Controlled Text GenerationSumanth Dathathri, Andrea Madotto, Janice Lan, Jane Hung et al.ICLR 2020 · 1,166 citations
- Self-Consistency Improves Chain of Thought Reasoning in Language ModelsXuezhi Wang, Jason Wei, Dale Schuurmans, Quoc V. Le et al.ICLR 2023 · 681 citations
- SALMONN: Towards Generic Hearing Abilities for Large Language ModelsChangli Tang, Wenyi Yu, Guangzhi Sun, Xianzhao Chen et al.ICLR 2024 · 557 citations
- DoLa: Decoding by Contrasting Layers Improves Factuality in Large Language ModelsYung-Sung Chuang, Yujia Xie, Hongyin Luo, Yoon Kim et al.ICLR 2024 · 354 citations
- Audio Flamingo 3: Advancing Audio Intelligence with Fully Open Large Audio Language ModelsSreyan Ghosh, Arushi Goel, Jaehyeon Kim, Sonal Kumar et al.NeurIPS 2025 · 299 citations
Related papers
- AVCD: Mitigating Hallucinations in Audio-Visual Large Language Models through Contrastive DecodingChaeyoung Jung, Youngjoon Jang, Joon Son ChungNeurIPS 2025 · 33 citations
- Mitigating Object Hallucinations in Large Vision-Language Models through Visual Contrastive DecodingSicong Leng, Hang Zhang, Guanzheng Chen, Xin Li et al.CVPR 2024
- Multi-Frequency Contrastive Decoding: Alleviating Hallucinations for Large Vision-Language ModelsBingqian Liu, Fu Zhang, Guoqing Chen, Jingwei ChengEMNLP 2025
- Alleviating Hallucinations in Large Language Models through Multi-Model Contrastive Decoding and Dynamic Hallucination DetectionChenyu Zhu, Yefeng Liu, Hao Zhang, Aowen Wang et al.NeurIPS 2025 · 7 citations
- MAD: Modality-Adaptive Decoding for Mitigating Cross-Modal Hallucinations in Multimodal Large Language ModelsSangyun Chung, Se Yeon Kim, Youngchae Chee, Yong Man RoCVPR 2026 · 5 citations
