HiDe-LLaVA: Hierarchical Decoupling for Continual Instruction Tuning of Multimodal Large Language Model (2025.acl-long)
Copied to clipboard
| Challenge: | Existing methods to improve instructionfollowing performance of MLLMs often trade off memory efficiency for performance gains, compromising overall efficiency. |
| Approach: | They propose a task-specific expansion and task-general fusion framework based on variations in Centered Kernel Alignment (CKA) similarity across different model layers when trained on diverse datasets. |
| Outcome: | The proposed framework improves performance compared to existing benchmarks. |
Similar Papers
Enhancing Multimodal Continual Instruction Tuning with BranchLoRA (2025.acl-long)
Copied to clipboard
| Challenge: | Existing approaches to fine tune Multimodal Large Language Models (MLLMs) are prone to Catastrophic Forgetting (CF) existing approaches rely on the Mixture-of-Experts (MoE) LoRA framework to preserve previous instruction alignments. |
| Approach: | They propose an asymmetric tuning-freezing mechanism to mitigate parameter inefficiency . branch-specific routers are introduced to ensure optimal branch distribution over time . |
| Outcome: | The proposed framework outperforms existing frameworks on the latest MCIT benchmarks. |
From Experts to Bases: Orthogonal Subspace Mixture for Continual Multimodal Instruction Tuning (2026.acl-long)
Copied to clipboard
| Challenge: | Existing parameter-efficient approaches to multimodal Continual Instruction Tuning suffer from knowledge interference and inefficient capacity expansion, limiting scalability. |
| Approach: | They propose a framework for multimodal Continual instruction tuning that decomposes adaptation weights into a globally shared pool of orthonormal bases to capture task-invariant knowledge. |
| Outcome: | Experiments show that MoBLoRA outperforms state-of-the-art methods while maintaining superior parameter efficiency. |
Align2LLaVA: Cascaded Human and Large Language Model Preference Alignment for Multi-modal Instruction Curation (2025.findings-acl)
Copied to clipboard
Hongzhe Huang, Jiang Liu, Zhewen Yu, Li Cai, Dian Jiao, Wenqiao Zhang, Siliang Tang, Juncheng Li, Hao Jiang, Haoyuan Li, Yueting Zhuang
| Challenge: | Recent advances in Multi-modal Large Language Models (MLLMs) introduce significant variability in data quality. |
| Approach: | They propose to use human and LLM preference alignment to compress large corpus of machine-generated multimodal instructions into a compact and high-quality form. |
| Outcome: | The proposed algorithm outperforms LLaVA-series models in MLLM benchmarks by 90% . it uses human and LLM preference alignment to compress a large dataset . |
Exploring Two-Phase Continual Instruction Fine-tuning for Multilingual Adaptation in Large Language Models (2026.findings-acl)
Copied to clipboard
| Challenge: | A key challenge for Large Language Models (LLMs) is improving their Multilingual instruction-following ability over time without deteriorating their ability in languages they already excel at, typically English. |
| Approach: | They propose a two-phase Continual Fine-tuning setup to improve a model's Multilingual adaptability by comparing an English-only LLM with a multilingual instruction dataset. |
| Outcome: | The proposed model improves on two-phase Continual Fine-tuning (CFT) setups on a multilingual instruction dataset. |
VFA: Empowering Multilingual MLLMs via Vision-Free Adaptation (2026.acl-long)
Copied to clipboard
Yixia Li, Yaqing Shi, Zhiwen Ruan, Dongdong Zhang, Lingjie Jiang, Shaohan Huang, Yun Chen, Guanhua Chen, Furu Wei
| Challenge: | Multimodal large language models have advanced rapidly, yet most remain English-centric . scaling multilingual multimodal instruction tuning is limited by the scarcity and high cost of non-English image–text supervision. |
| Approach: | They propose a framework that decouples multilingual language enhancement from visual alignment by composing complementary task vectors over a shared LLM backbone. |
| Outcome: | The proposed framework achieves competitive performance with a fully multimodally trained model using less than 2% of the text data. |
Progressive LoRA for Multimodal Continual Instruction Tuning (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing approaches to MCIT address Catastrophic Forgetting and Knowledge Transfer (KT) but using a fixed number of shared LoRA blocks across tasks can lead to knowledge interference. |
| Approach: | They propose a framework that uses a fixed number of shared LoRA blocks to reduce knowledge interference. |
| Outcome: | The proposed framework outperforms existing approaches on the latest MCIT benchmark. |
CoMMIT: Coordinated Multimodal Instruction Tuning (2025.emnlp-main)
Copied to clipboard
Xintong Li, Junda Wu, Tong Yu, Rui Wang, Yu Wang, Xiang Chen, Jiuxiang Gu, Lina Yao, Julian McAuley, Jingbo Shang
| Challenge: | et al., 2024) show that multimodal instruction tuning is more effective than baselines. |
| Approach: | They propose a multimodal balance coefficient that enables quantitative measurement of the balance of learning . they propose auxiliary regularization on the gradient to promote updating with larger step sizes . |
| Outcome: | The proposed method is more effective than baselines in MLLM instruction tuning. |
Demystifying Instruction Mixing for Fine-tuning Large Language Models (2024.acl-srw)
Copied to clipboard
| Challenge: | Instruction tuning is effective for aligning large language models with human instructions, but the procedure to optimizing the mixing of instruction datasets is still unclear. |
| Approach: | They categorize instructions into three primary types: NLP downstream tasks, coding, and general chat. |
| Outcome: | The proposed method improves performance of large language models (LLMs) but it is difficult to combine different instruction datasets to optimize overall performance. |
Multi-modal Preference Alignment Remedies Degradation of Visual Instruction Tuning on Language Models (2024.acl-long)
Copied to clipboard
| Challenge: | Multi-modal large language models (MLLMs) are expected to support multi-turn queries of interchanging image and text modalities in production. |
| Approach: | They propose to use visual-question-answering (VQA) datasets to annotate a 5k-sample VQA preference dataset and to investigate the degradation of VQA datasets. |
| Outcome: | The proposed model surpasses the instruction-following capabilities of the language model with DPO and SteerLM. |
Layer by Layer: Uncovering Where Multi-Task Learning Happens in Instruction-Tuned Large Language Models (2024.emnlp-main)
Copied to clipboard
| Challenge: | Pre-trained large language models retain task-specific knowledge, but where and to what extent they retain it remains unexplored. |
| Approach: | They investigate the task-specific information encoded in pre-trained LLMs and the effects of instruction tuning on their representations across over 60 NLP tasks. |
| Outcome: | The results show that pre-trained models retain task-specific knowledge . some tasks are already encoded in pre-train models, but others benefit from instruction tuning. |