Dynamic Mixture of Curriculum LoRA Experts for Continual Multimodal Instruction Tuning

• URL: http://arxiv.org/abs/2506.11672v1
• Date: Fri, 13 Jun 2025 11:03:46 GMT
• Title: Dynamic Mixture of Curriculum LoRA Experts for Continual Multimodal Instruction Tuning
• Authors: Chendi Ge, Xin Wang, Zeyang Zhang, Hong Chen, Jiapei Fan, Longtao Huang, Hui Xue, Wenwu Zhu,
• Abstract summary: Continual multimodal instruction tuning is crucial for adapting Multimodal Large Language Models (MLLMs) to evolving tasks.<n>We propose a novel Dynamic Mixture of Curriculum LoRA Experts (D-MoLE) method, which automatically evolves MLLM's architecture with controlled parameter budgets to continually adapt to new tasks.<n>Specifically, we propose a dynamic layer-wise expert allocator, which automatically allocates LoRA experts across layers to resolve architecture conflicts.<n>Then, we propose a gradient-based inter-modal continual curriculum, which adjusts the update ratio of each module in MLLM based on the difficulty of each
• Score: 45.019751165506946
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Continual multimodal instruction tuning is crucial for adapting Multimodal Large Language Models (MLLMs) to evolving tasks. However, most existing methods adopt a fixed architecture, struggling with adapting to new tasks due to static model capacity. We propose to evolve the architecture under parameter budgets for dynamic task adaptation, which remains unexplored and imposes two challenges: 1) task architecture conflict, where different tasks require varying layer-wise adaptations, and 2) modality imbalance, where different tasks rely unevenly on modalities, leading to unbalanced updates. To address these challenges, we propose a novel Dynamic Mixture of Curriculum LoRA Experts (D-MoLE) method, which automatically evolves MLLM's architecture with controlled parameter budgets to continually adapt to new tasks while retaining previously learned knowledge. Specifically, we propose a dynamic layer-wise expert allocator, which automatically allocates LoRA experts across layers to resolve architecture conflicts, and routes instructions layer-wisely to facilitate knowledge sharing among experts. Then, we propose a gradient-based inter-modal continual curriculum, which adjusts the update ratio of each module in MLLM based on the difficulty of each modality within the task to alleviate the modality imbalance problem. Extensive experiments show that D-MoLE significantly outperforms state-of-the-art baselines, achieving a 15% average improvement over the best baseline. To the best of our knowledge, this is the first study of continual learning for MLLMs from an architectural perspective.

Related papers

• VL-Cogito: Progressive Curriculum Reinforcement Learning for Advanced Multimodal Reasoning [69.44871115752055]
  We propose an advanced multimodal reasoning model trained via a novel Progressive Curriculum Reinforcement Learning (PCuRL) framework.<n>PCuRL systematically guides the model through tasks of gradually increasing difficulty, substantially improving its reasoning abilities across diverse multimodal contexts.<n>The framework introduces two key innovations: (1) an online difficulty soft weighting mechanism, dynamically adjusting training difficulty across successive RL training stages; and (2) a dynamic length reward mechanism, which encourages the model to adaptively regulate its reasoning path length according to task complexity.
  arXiv  Detail & Related papers  (2025-07-30T12:23:21Z)
• Mixture-of-Experts Meets In-Context Reinforcement Learning [29.866936147753368]
  In this paper, we introduce textbfT2MIR (textbfToken- and textbfTask-wise textbfMoE for textbfIn-context textbfRL), an innovative framework that introduces architectural advances of mixture-of-experts (MoE) into transformer-based decision models.<n> Comprehensive experiments show that T2MIR significantly facilitates in-context learning capacity and outperforms various types of baselines.
  arXiv  Detail & Related papers  (2025-06-05T06:29:14Z)
• Separable Mixture of Low-Rank Adaptation for Continual Visual Instruction Tuning [16.873306091966693]
  Visual instruction tuning enables large language models (MLLMs) to handle a wide range of vision tasks by framing them as language-based instructions. We identify a dual form of catastrophic forgetting in CVIT, where MLLMs forget previously learned visual understanding but also experience a decline in instruction following abilities. We introduce the Separable Mixture of Low-Rank Adaptation (SMoLoRA) framework, which employs separable routing through two distinct modules. This dual-routing design enables specialized adaptation in both domains, preventing forgetting while improving performance.
  arXiv  Detail & Related papers  (2024-11-21T09:00:15Z)
• AT-MoE: Adaptive Task-planning Mixture of Experts via LoRA Approach [0.6906005491572401]
  This paper introduces the Adaptive Task-planing Mixture of Experts(AT-MoE) architecture. We first train task-specific experts via LoRA approach to enhance problem-solving capabilities and interpretability in specialized areas. We then introduce a layer-wise adaptive grouped routing module that optimize module fusion based on complex task instructions.
  arXiv  Detail & Related papers  (2024-10-12T13:03:15Z)
• MoExtend: Tuning New Experts for Modality and Task Extension [61.29100693866109]
  MoExtend is an effective framework designed to streamline the modality adaptation and extension of Mixture-of-Experts (MoE) models. MoExtend seamlessly integrates new experts into pre-trained MoE models, endowing them with novel knowledge without the need to tune pretrained models.
  arXiv  Detail & Related papers  (2024-08-07T02:28:37Z)
• Multimodal Instruction Tuning with Conditional Mixture of LoRA [51.58020580970644]
  This paper introduces a novel approach that integrates multimodal instruction tuning with Low-Rank Adaption (LoRA)<n>It innovates upon LoRA by dynamically constructing low-rank adaptation matrices tailored to the unique demands of each input instance.<n> Experimental results on various multimodal evaluation datasets indicate that MixLoRA not only outperforms the conventional LoRA with the same or even higher ranks.
  arXiv  Detail & Related papers  (2024-02-24T20:15:31Z)
• Continual Instruction Tuning for Large Multimodal Models [30.438442723421556]
  Multi-task joint instruction tuning can facilitate the model's continual learning ability and forgetting. We propose task-similarity-informed regularization and model expansion methods for continual instruction tuning of LMMs.
  arXiv  Detail & Related papers  (2023-11-27T15:04:48Z)
• JiuZhang 2.0: A Unified Chinese Pre-trained Language Model for Multi-task Mathematical Problem Solving [77.51817534090789]
  We propose textbfJiuZhang2.0, a unified Chinese PLM specially for multi-task mathematical problem solving. Our idea is to maintain a moderate-sized model and employ the emphcross-task knowledge sharing to improve the model capacity in a multi-task setting.
  arXiv  Detail & Related papers  (2023-06-19T15:45:36Z)
• Fully Online Meta-Learning Without Task Boundaries [80.09124768759564]
  We study how meta-learning can be applied to tackle online problems of this nature. We propose a Fully Online Meta-Learning (FOML) algorithm, which does not require any ground truth knowledge about the task boundaries. Our experiments show that FOML was able to learn new tasks faster than the state-of-the-art online learning methods.
  arXiv  Detail & Related papers  (2022-02-01T07:51:24Z)
• A Simple General Approach to Balance Task Difficulty in Multi-Task Learning [4.531240717484252]
  In multi-task learning, difficulty levels of different tasks are varying. We propose a Balanced Multi-Task Learning (BMTL) framework. The proposed BMTL framework is very simple and it can be combined with most multi-task learning models.
  arXiv  Detail & Related papers  (2020-02-12T04:31:34Z)

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.