Towards Understanding Token Evolution in Continual Learning of MLLMs

Xu-Sheng Cao Yi-Fan Meng Jia-Ze Li Jian-Zhong Ju Zhen-Bo Luo Jian Luan Fei Yang Ming-Ming Cheng Xia-Lei Liu

Sep 2026 · International Journal of Computer Vision · Vol 134 · 0 citations · 54 references

Domain Adaptation and Few-Shot Learning

TL;DR

This work revisits catastrophic forgetting from an interpretability perspective and introduces correctness-guided layer alignment (CLA), which allows the new model to inherit multi-layer semantic consistency from its predecessor while learning from ground-truth supervision.

DOI Semantic Scholar

View source

Similar papers

Preprint Aug 2026

A Model Merging Approach for Continual MLLM Unlearning

This work introduces Merging for Continual Unlearning (MCU), an approach that dynamically merges multiple one-shot unlearning adapters into a unified adapter upon receiving each new unlearning request and achieves superior unlearning effectiveness while preserving both retained knowledge and general multimodal utility.

Yuhang Wang, Lin-Lin Zhang, Haoxuan Ji et al. · 0 citations

#artificial intelligence Preprint Sep 2026

ConvMem: Convolutional Memory for Long-Context Reasoning

While Large Language Models (LLMs) have demonstrated impressive capabilities, they often struggle with extremely long contexts due to fixed context limits. To address this, sequential approaches like MemAgent extend the effective context by reading text in segments and iteratively updating a fixed-size memory. However,...

Hongming Zhang, Zhao-Zhen Gu, Fengshuo Bai et al. · 0 citations

#artificial intelligence Preprint Sep 2026

Dual-Latent Memory Routing for Vision-Language Reasoning

DLMR is a parameter-efficient mechanism that equips MLLMs with Dual Latent Memories: a visual memory that compresses image evidence and a reasoning memory that tracks intermediate conclusions and constraints with only a small number of additional trainable parameters.

Hao-Xuan Ma, Jin-Fei Qi, Yi-Cheng Xiao et al. · 0 citations

Preprint Aug 2026

Beyond Forgetting: Diagnosing and Harnessing Shared Reasoning in Continual RLVR

This work introduces Continual Prompt Replay (CPR), which harnesses shared reasoning to improve learning on the arriving and future tasks by replaying previous-task prompts and regenerating their responses with the current policy.

Li-Rui Luo, Guo-Xi Zhang, Hong-Ming Xu et al. · 0 citations

Sep 2026

Making Large Vision–Language Models Better Few-Shot Learners

Few-shot classification (FSC) aims to emulate the human ability to rapidly learn new concepts from a handful of examples. Large Vision-Language Models (LVLMs), with their rich prior knowledge and powerful visio-linguistic understanding capabilities, are emerging as a highly promising paradigm for FSC. This paper invest...

Chuan-Yi Zhang, Fan Liu, Yi Xu et al. · IEEE Transactions on Image P... · 0 citations

Preprint Aug 2026

AWARe: Mitigating Catastrophic Forgetting via Activation-Weighted Adaptive REtention

Activation-Weighted Adaptive REtention (AWARe) is a fine-tuning method that mitigates catastrophic forgetting by dynamically controlling parameter updates based on activation patterns, and operates without modifying model architectures, ensuring compatibility with existing inference engines.

Jun-Yong Liao, Jinfan Lv, Guoming Wang et al. · 1 citation