AudioDER: A Deduplication-Enhanced Reasoning Dataset for Post-Training Large Audio-Language Models

📄 arXiv: 2606.14591v1 📥 PDF

作者: Hui Geng, Yi Su, Han Yin, Tianjiao Wan, Qisheng Xu, Jiaxin Chen, Zijian Gao, Hengzhu Liu, Xie Chen, Kele Xu

分类: cs.SD, cs.AI

发布日期: 2026-06-12

💡 一句话要点

提出AudioDER以解决音频推理数据集冗余问题

🎯 匹配领域: 支柱九：具身大模型 (Embodied Foundation Models)

关键词: 音频推理 数据去重 多模态学习 后训练 音频语言模型 监督学习 链式思维

📋 核心要点

  1. 现有音频语言模型在复杂音频推理任务中表现不佳，主要由于训练数据的冗余性和多样性不足。
  2. 本文提出了一种冗余感知的数据构建流程，通过声学相似性去重和整合多种注释格式，提升数据集的多样性。
  3. 实验结果显示，在AudioDER上进行后训练，Qwen2-Audio-7B-Instruct在多个音频推理基准上性能显著提升。

📝 摘要（中文）

大型音频语言模型（LALMs）在音频理解任务上表现出色，但在复杂音频推理方面仍存在挑战。现有音频语言数据集往往存在显著冗余，导致样本高度相似，增加了标注成本并限制了数据多样性。为此，本文提出了一种冗余感知的数据构建流程，旨在为LALMs构建推理导向的监督数据。通过对原始音频数据集进行声学相似性基础的去重，整合现有音频标题和问答对，最终构建了包含约191k样本的AudioDER数据集。实验表明，在AudioDER上进行后训练显著提升了Qwen2-Audio-7B-Instruct在多个音频推理基准上的表现。

🔬 方法详解

问题定义：本文旨在解决现有音频语言数据集中的冗余问题，导致样本相似性高，影响模型的学习效果和推理能力。

核心思路：通过声学相似性去重，提升数据集的多样性，并将不同类型的注释整合为统一的多选格式，以增强推理导向的监督。

技术框架：整体流程包括声学相似性去重、注释整合和基于Qwen3-30B生成链式思维（CoT）推理依据，最终构建AudioDER数据集。

关键创新：提出了一种冗余感知的数据构建方法，显著提升了数据集的多样性和有效性，区别于传统的音频语言数据集构建方式。

关键设计：在去重过程中，采用声学相似性度量，确保样本之间的多样性；在注释整合中，设计了统一的多选格式，以便于后续的推理训练。生成的CoT推理依据为模型提供了更清晰的推理路径。

🖼️ 关键图片

fig_0
fig_1
fig_2

📊 实验亮点

在多个音频推理基准上进行的实验表明，Qwen2-Audio-7B-Instruct在AudioDER数据集上进行后训练后，性能显著提升，具体提升幅度在不同基准上均表现出一致的改善，验证了数据集的有效性和实用性。

🎯 应用场景

该研究的潜在应用领域包括音频理解、语音识别和音乐分析等。通过提升音频语言模型的推理能力，AudioDER可以为智能音频助手、自动化内容生成和多模态交互等应用提供更强的支持，推动相关技术的发展与应用。

📄 摘要（原文）

Large Audio-Language Models (LALMs) have shown strong performance on a wide range of audio understanding tasks, yet they still struggle with complex audio reasoning. A practical way to improve such capabilities is post-training, whose effectiveness critically depends on the quality and diversity of training data. However, existing audio-language datasets often contain substantial redundancy, where many samples are highly similar in acoustic content and thus provide overlapping supervisory signals. Such redundancy not only increases annotation cost, but also limits corpus diversity and reduces the effectiveness of post-training. To address this issue, we propose a redundancy-aware data construction pipeline for building reasoning-oriented supervision for LALMs. Specifically, we first perform acoustic similarity-based deduplication across raw audio datasets to improve corpus diversity. We then integrate existing audio captions and question-answer pairs into a unified multiple-choice format. Based on these unified annotations, we leverage Qwen3-30B to generate chain-of-thought (CoT) rationales for reasoning-oriented supervision. Based on this pipeline, we construct AudioDER, a reasoning-oriented post-training dataset containing approximately 191k samples spanning sound, speech, and music. Each sample consists of an audio clip, a multiple-choice question, four answer candidates, an audio caption, and a CoT rationale. Extensive experiments show that post-training on AudioDER consistently improves the performance of Qwen2-Audio-7B-Instruct on multiple audio reasoning benchmarks, including MMAU-mini, MMSU, and MMAR. We hope AudioDER can serve as a valuable resource for advancing audio reasoning research and the development of more capable LALMs.