DeSTA2.5-Audio: Toward General-Purpose Large Audio Language Model with Self-Generated Cross-Modal Alignment

• URL: http://arxiv.org/abs/2507.02768v1
• Date: Thu, 03 Jul 2025 16:28:25 GMT
• Title: DeSTA2.5-Audio: Toward General-Purpose Large Audio Language Model with Self-Generated Cross-Modal Alignment
• Authors: Ke-Han Lu, Zhehuai Chen, Szu-Wei Fu, Chao-Han Huck Yang, Sung-Feng Huang, Chih-Kai Yang, Chee-En Yu, Chun-Wei Chen, Wei-Chih Chen, Chien-yu Huang, Yi-Cheng Lin, Yu-Xiang Lin, Chi-An Fu, Chun-Yi Kuan, Wenze Ren, Xuanjun Chen, Wei-Ping Huang, En-Pei Hu, Tzu-Quan Lin, Yuan-Kuei Wu, Kuan-Po Huang, Hsiao-Ying Huang, Huang-Cheng Chou, Kai-Wei Chang, Cheng-Han Chiang, Boris Ginsburg, Yu-Chiang Frank Wang, Hung-yi Lee,
• Abstract summary: We introduce DeSTA2.5-Audio, a general-purpose Large Audio Language Model (LALM)<n>It is designed for robust auditory perception and instruction-following, without requiring task-specific audio instruction-tuning.<n>DeSTA2.5-Audio achieves state-of-the-art or competitive performance across a wide range of audio-language benchmarks.
• Score: 94.0709779805955
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: We introduce DeSTA2.5-Audio, a general-purpose Large Audio Language Model (LALM) designed for robust auditory perception and instruction-following, without requiring task-specific audio instruction-tuning. Recent LALMs typically augment Large Language Models (LLMs) with auditory capabilities by training on large-scale, manually curated or LLM-synthesized audio-instruction datasets. However, these approaches have often suffered from the catastrophic forgetting of the LLM's original language abilities. To address this, we revisit the data construction pipeline and propose DeSTA, a self-generated cross-modal alignment strategy in which the backbone LLM generates its own training targets. This approach preserves the LLM's native language proficiency while establishing effective audio-text alignment, thereby enabling zero-shot generalization without task-specific tuning. Using DeSTA, we construct DeSTA-AQA5M, a large-scale, task-agnostic dataset containing 5 million training samples derived from 7,000 hours of audio spanning 50 diverse datasets, including speech, environmental sounds, and music. DeSTA2.5-Audio achieves state-of-the-art or competitive performance across a wide range of audio-language benchmarks, including Dynamic-SUPERB, MMAU, SAKURA, Speech-IFEval, and VoiceBench. Comprehensive comparative studies demonstrate that our self-generated strategy outperforms widely adopted data construction and training strategies in both auditory perception and instruction-following capabilities. Our findings underscore the importance of carefully designed data construction in LALM development and offer practical insights for building robust, general-purpose LALMs.

Related papers

• PAL: Probing Audio Encoders via LLMs -- A Study of Information Transfer from Audio Encoders to LLMs [16.820927353576774]
  The integration of audio perception capabilities into Large Language Models (LLMs) has enabled significant advances in Audio-LLMs.<n>We conceptualize effective audio-LLM interaction as the LLM's ability to proficiently probe the audio encoder representations to satisfy textual queries.<n>This paper presents a systematic investigation on how architectural design choices can affect that.
  arXiv  Detail & Related papers  (2025-06-12T07:23:07Z)
• 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)
• IFEval-Audio: Benchmarking Instruction-Following Capability in Audio-based Large Language Models [18.11667976818302]
  IFEval-Audio contains 280 audio-instruction-answer triples across six diverse dimensions.<n>Each example pairs an audio input with a text instruction, requiring the model to generate an output that follows a specified structure.<n>We benchmark state-of-the-art audio LLMs on their ability to follow audio-involved instructions.
  arXiv  Detail & Related papers  (2025-05-22T15:15:29Z)
• DeSTA2: Developing Instruction-Following Speech Language Model Without Speech Instruction-Tuning Data [84.01401439030265]
  Recent end-to-end speech language models (SLMs) have expanded upon the capabilities of large language models (LLMs)<n>We present a simple yet effective automatic process for creating speech-text pair data.<n>Our model demonstrates general capabilities for speech-related tasks without the need for speech instruction-tuning data.
  arXiv  Detail & Related papers  (2024-09-30T07:01:21Z)
• Enhancing Audio-Language Models through Self-Supervised Post-Training with Text-Audio Pairs [3.8300818830608345]
  Multi-modal contrastive learning strategies for audio and text have rapidly gained interest.<n>The ability of these models to understand natural language and temporal relations is still a largely unexplored and open field for research.<n>We propose to equip the multi-modal ALMs with temporal understanding without loosing their inherent prior capabilities of audio-language tasks with a temporal instillation method TeminAL.
  arXiv  Detail & Related papers  (2024-08-17T18:53:17Z)
• Meerkat: Audio-Visual Large Language Model for Grounding in Space and Time [73.7845280328535]
  We present Meerkat, an audio-visual LLM equipped with a fine-grained understanding of image and audio. Meerkat can tackle challenging tasks such as audio referred image grounding, image guided audio temporal localization, and audio-visual fact-checking. We achieve state-of-the-art performance on all these downstream tasks with a relative improvement of up to 37.12%.
  arXiv  Detail & Related papers  (2024-07-01T23:32:25Z)
• WavLLM: Towards Robust and Adaptive Speech Large Language Model [93.0773293897888]
  We introduce WavLLM, a robust and adaptive speech large language model with dual encoders, and a prompt-aware LoRA weight adapter. We validate the proposed model on universal speech benchmarks including tasks such as ASR, ST, SV, ER, and also apply it to specialized datasets like Gaokao English listening comprehension set for SQA, and speech Chain-of-Thought (CoT) evaluation set.
  arXiv  Detail & Related papers  (2024-03-31T12:01:32Z)
• Text-to-Audio Generation using Instruction-Tuned LLM and Latent Diffusion Model [23.058939018350603]
  Large language models (LLM) allow many interesting properties, such as, instruction- and chain-of-thought-based fine-tuning. We adopt such an instruction-tuned LLM Flan-T5 as the text encoder for text-to-audio (TTA) generation. Our approach TANGO outperforms the state-of-the-art AudioLDM on most metrics and stays comparable on the rest on AudioCaps test set.
  arXiv  Detail & Related papers  (2023-04-24T07:45:28Z)
• WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing [102.45426364965887]
  We propose a new pre-trained model, WavLM, to solve full-stack downstream speech tasks. WavLM is built based on the HuBERT framework, with an emphasis on both spoken content modeling and speaker identity preservation. We scale up the training dataset from 60k hours to 94k hours of public audio data, and optimize its training procedure for better representation extraction.
  arXiv  Detail & Related papers  (2021-10-26T17:55:19Z)

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.