Parameter Isolation

Follow this topic Ask about this topic

Latest papers

8 papers2 papers in the last four weeks, down 33% on the four weeks before. 0.0% of all new papers.

Continual Adaptation Across Modalities: Learning and Forgetting in Unified Multimodal Models
Oct 28, 2026Xinrui Wang, Zhihao Chen, Jia Tang +4Parameter Isolation Unified Multimodal Models
Nanjing University of Aeronautics and Astronautics · Computer Vision Center, Barcelona · EPFL +3
SPARS: A Stability-Plasticity-Aware Region Selection Framework for Continual Learning
Oct 14, 2026Teif Aldaajani, Huda Hakami, Maha A. ThafarRegularization-Based Continual Learning Parameter Isolation
Department of Computer Science, College of Computers and Information Technology, Taif University, Taif, 21944, Saudi Arabia .
Continual Learning of Dynamical Systems in Recurrent Neural Networks through Recyclable Unit Gating
Sep 29, 2026Sima Hashemi, Daniel Durstewitz, Georgia KoppeDynamical Systems Continual Learning
Faculty of Mathematics and Computer Science, Interdisciplinary Center for Scientific Computing, Heidelberg University, Heidelberg, Germany · Department of Theoretical Neuroscience, Central Institute of Mental Health (CIMH), Medical Faculty Mannheim, Heidelberg University, Heidelberg, Germany · Hector Institute for AI in Psychiatry & Department of Psychiatry and Psychotherapy, CIMH, Medical Faculty Mannheim, Heidelberg University, Heidelberg, Germany +1
Learning When to Recur: Token-Adaptive Recursion for Imbalanced Ophthalmic Domain Incremental Learning
Sep 26, 2026Nanxi Yu, Kang Li, Ye Du +3Parameter Isolation Ophthalmic Image Analysis
School of Mechanical and Electrical Engineering, University of Electronic Science and Technology of China, Sichuan, China · Department of Biomedical Engineering and Sports Technology, The Hong Kong Polytechnic University, Hong Kong SAR, China · School of Future Technology, South China University of Technology, Guangdong, China +1
Parameter isolation with domain-specific experts for incremental audio classification
Sep 13, 2026Jongyeon Park, Do-Hyeon Lim, Sang-won Park +4Parameter Isolation Domain-Incremental Learning
Dept. of AI Convergence, Gwangju Institute of Science and Technology (GIST) · Dept. of EECS, Gwangju Institute of Science and Technology (GIST) · AunionAI Co., Ltd., Gwangju, Korea +1
JumpLoRA: Sparse Adapters for Continual Learning in Large Language Models
Apr 17, 2026Alexandra Dragomir, Ioana Pintilie, Antonio Barbalau +6Parameter-Efficient Fine-Tuning Low-Rank Adaptation
Bitdefender · ETH Zurich · University of Bucharest
Parameter Importance is Not Static: Evolving Parameter Isolation for Supervised Fine-Tuning
Apr 15, 2026Zekai Lin, Chao Xue, Di Liang +8Parameter Isolation Supervised Fine-Tuning
Tencent Hunyuan · University of New South Wales · Tencent Yuanbao +2
MedCL-Bench: Benchmarking stability-efficiency trade-offs and scaling in biomedical continual learning
Mar 17, 2026Min Zeng, Shuang Zhou, Zaifu Zhan +1Parameter Isolation Continual Learning
Division of Computational Health Sciences, Department of Surgery, University of Minnesota, Minneapolis, 55455, MN, USA.

Oct 28, 2026J/K move · Enter open · S save

Continual Adaptation Across Modalities: Learning and Forgetting in Unified Multimodal Models

Xinrui Wang, Zhihao Chen, Jia Tang, Alexandra Gomez-Villa+3

Nanjing University of Aeronautics and Astronautics · Computer Vision Center, Barcelona · EPFL+3

Parameter Isolation Unified Multimodal Models Continual Learning Catastrophic Forgetting

Open paper AskPDF arXiv

Unified multimodal models (UMMs), which can process and generate diverse modalities within a single framework, have recently emerged as a powerful paradigm. However, when these models undergo continual learning on multiple tasks across different modalities, their learning and forgetting dynamics remain poorly understood. In this paper, we systematically investigate how architectural design and adaptation method shape intra-modal learning and cross-modal interference between the generation and understanding sides. Specifically, we report three findings. First, the severity of cross-modal interference decreases with the degree of modality-specific parameter isolation. Second, a within-model asymmetry: the generation side of a UMM is markedly more fragile than its understanding side. Third, what post-training adds is the first to be forgotten: purely post-training capabilities such as personalized generation collapse by 84% to 90%, while pretraining capabilities remain stable. Together, these results clarify how parameter isolation and post-training fragility shape UMMs' learning and forgetting dynamics, offering a starting point for developing continually evolving UMMs.