Heterogeneity-Aware Dataset Scheduling for Efficient Audio Large Language Model Training

• URL: http://arxiv.org/abs/2605.19101v1
• Date: Mon, 18 May 2026 20:41:08 GMT
• Title: Heterogeneity-Aware Dataset Scheduling for Efficient Audio Large Language Model Training
• Abstract summary: Grouped Sequential Training ( GST) is an efficient way to train Audio Large Language Models (ALLMs) across diverse datasets.<n> GST achieves 30--40% faster convergence than standard parallel training while maintaining or even surpassing the performance of mix-all training.<n>Our results provide both theoretical insights and a practical, model-agnostic framework for efficient large-scale ALLM optimization.
• Score: 8.070081784771899
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Training general-purpose Audio Large Language Models (ALLMs) across diverse datasets is essential for holistic audio understanding, yet it faces significant challenges due to dataset heterogeneity, which often leads to conflicting gradients and slow convergence. Despite its impact, how to explicitly manage this heterogeneity during training remains underexplored, with current practices relying primarily on uniform mixture. In this work, we analyze multi-dataset AudioQA training from a convergence perspective and propose Grouped Sequential Training (GST). GST strategically organizes datasets into affinity-aware groups and introduces them via a progressive scheduling protocol, effectively balancing the stability of parallel training with the efficiency of sequential optimization. To ensure scalability, we develop gradient-based affinity metrics that capture inter-dataset relationships without the prohibitive cost of empirical transferability estimation. Extensive evaluations on 14 AudioQA datasets spanning speech, music, and environmental sounds demonstrate that GST achieves 30--40\% faster convergence than standard parallel training while maintaining or even surpassing the performance of mix-all training. Our results provide both theoretical insights and a practical, model-agnostic framework for efficient large-scale ALLM optimization.

Related papers

• SAW: Stage-Aware Dynamic Weighting for Multi-Objective Reinforcement Learning in Large Language Models [66.92079232778707]
  We propose Stage-Aware Dynamic Weighting (SAW) as a lightweight, algorithm-agnostic dynamic weighting mechanism.<n>SAW reweights each dimension's reward or advantage contribution by its relative informativeness within the batch.<n> Experiments on tool-calling and text summarization tasks demonstrate that SAW consistently improves both training efficiency and final performance.
  arXiv  Detail & Related papers  (2026-06-05T10:00:19Z)
• Data Mixing for Large Language Models Pretraining: A Survey and Outlook [52.66761353708619]
  Large language models (LLMs) rely on pretraining on massive and heterogeneous corpora.<n>Unlike sample-level data selection, data mixing optimize domain-level sampling weights to allocate limited budgets more effectively.<n>This paper provides a comprehensive review of data mixing for LLM pretraining.
  arXiv  Detail & Related papers  (2026-03-25T13:30:40Z)
• DRL: Discriminative Representation Learning with Parallel Adapters for Class Incremental Learning [63.65467569295623]
  We propose the Discriminative Representation Learning (DRL) framework to specifically address these challenges.<n>To conduct incremental learning effectively and yet efficiently, the DRL's network is built upon a PTM.<n>Our DRL consistently outperforms other state-of-the-art methods throughout the entire CIL period.
  arXiv  Detail & Related papers  (2025-10-14T03:19:15Z)
• A Scalable Pretraining Framework for Link Prediction with Efficient Adaptation [16.82426251068573]
  Link Prediction (LP) is a critical task in graph machine learning.<n>Existing methods face key challenges including limited supervision from sparse connectivity.<n>We explore pretraining as a solution to address these challenges.
  arXiv  Detail & Related papers  (2025-08-06T17:10:31Z)
• Dynamic Loss-Based Sample Reweighting for Improved Large Language Model Pretraining [55.262510814326035]
  Existing reweighting strategies primarily focus on group-level data importance.<n>We introduce novel algorithms for dynamic, instance-level data reweighting.<n>Our framework allows us to devise reweighting strategies deprioritizing redundant or uninformative data.
  arXiv  Detail & Related papers  (2025-02-10T17:57:15Z)
• Enhancing Generalization via Sharpness-Aware Trajectory Matching for Dataset Condensation [37.77454972709646]
  We introduce Sharpness-Aware Trajectory Matching (SATM), which enhances the generalization capability of learned synthetic datasets.<n>Our approach is mathematically well-supported and straightforward to implement along with controllable computational overhead.
  arXiv  Detail & Related papers  (2025-02-03T22:30:06Z)
• Aligning Instruction Tuning with Pre-training [61.50161961371844]
  We propose Aligning Instruction Tuning with Pre-training (AITP) to align instruction tuning with pre-training distributions.<n>We show consistent performance improvements with AITP on three fully open large language models (LLMs) across eight benchmarks.
  arXiv  Detail & Related papers  (2025-01-16T08:27:40Z)
• Improving Generalization of Speech Separation in Real-World Scenarios: Strategies in Simulation, Optimization, and Evaluation [45.76004686788507]
  We present a novel data simulation pipeline that produces diverse training data from a range of acoustic environments and content. We propose new training paradigms to improve quality of a general speech separation model.
  arXiv  Detail & Related papers  (2024-08-28T20:26:34Z)
• Towards Robust Few-shot Class Incremental Learning in Audio Classification using Contrastive Representation [1.3586572110652484]
  Few-shot class-incremental learning addresses challenges arising from limited incoming data. We propose supervised contrastive learning to refine the representation space, enhancing discriminative power and leading to better generalization.
  arXiv  Detail & Related papers  (2024-07-27T14:16:25Z)
• Take the Bull by the Horns: Hard Sample-Reweighted Continual Training Improves LLM Generalization [165.98557106089777]
  A key challenge is to enhance the capabilities of large language models (LLMs) amid a looming shortage of high-quality training data. Our study starts from an empirical strategy for the light continual training of LLMs using their original pre-training data sets. We then formalize this strategy into a principled framework of Instance-Reweighted Distributionally Robust Optimization.
  arXiv  Detail & Related papers  (2024-02-22T04:10:57Z)
• DYNAMITE: Dynamic Interplay of Mini-Batch Size and Aggregation Frequency for Federated Learning with Static and Streaming Dataset [23.11152686493894]
  Federated Learning (FL) is a distributed learning paradigm that can coordinate heterogeneous edge devices to perform model training without sharing private data. This paper introduces novel analytical models and optimization algorithms that leverage the interplay between batch size and aggregation frequency to navigate the trade-offs among convergence, cost, and completion time for dynamic FL training.
  arXiv  Detail & Related papers  (2023-10-20T08:36:12Z)

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.