Continual Learning for VLA Models

VLA: Vision-Language-Action

Continual learning for vision-language-action models asks how a robot can adapt to new tasks, environments, and feedback while retaining useful earlier skills. Common approaches include replaying demonstrations, adapting with LoRA or task-specific adapters, constraining changes to important parameters, and organizing skills in modular experts or reusable memory. Methods are tested on sequential manipulation tasks in simulation benchmarks such as LIBERO and in real-world task streams, where replay budgets, task diversity, and deployment constraints matter. The central tension is that pretrained models can sometimes resist forgetting, yet targeted fine-tuning can still erase skills, leaving the challenge of reliably balancing adaptation and retention.

Latest papers

23 papers5 papers in the last four weeks, up 25% on the four weeks before. 0.0% of all new papers.

SARAM: Sparse Adapter Allocation with Reuse and Merging for Lifelong Robot Learning
Oct 17, 2026SeongJun Jeong, Chungwoo Lee, Gi-Cheon Kang +1Adapter-Based Continual Learning Continual Robot Learning
Seoul National University · Ajou University
NestedVLA: Learning to Consolidate and Generate Skills for Vision-Language-Action Models
Oct 16, 2026Guodong Du, Wanyu LinVision-Language-Action Models Continual Learning for VLA Models
Department of Computing, The Hong Kong Polytechnic University
Learning from Runtime Feedback through Failure-Bank Self-Evolution for Vision-Language-Action Models
Sep 30, 2026Mingyue Cui, Zheyuan Liu, Yihan Zhu +2Continual Learning for VLA Models Robotic Manipulation
University of Notre Dame
Inline Memory Meets Reusable Skills: Memory-centric Framework for Vision-Language-Action Model
Sep 30, 2026Zaijing Li, Rui Shao, Bing Hu +3Memory-Augmented VLMs VLA Model Adaptation
Harbin Institute of Technology (Shenzhen) · Pengcheng Laboratory
F4R: Failure-Driven Recognition, Reconstruction, Refinement, and Redeployment for Continual Robot Self-Improvement
Sep 28, 2026Zhuoyuan Yu, Jiacheng Wang, Tianle Liu +13Continual Robot Learning Sim-to-Real Transfer
Nanyang Technological University · Xi’an Jiaotong University · Dexmal
SEES: A Self-Evolving Embodied System via Failure-Guided VLA Policy Adaptation
Sep 26, 2026Ziwen Li, Hanlue Zhang, Zhenyang Ren +9Long-Horizon Robotic Manipulation Continual Learning for VLA Models
Mohamed bin Zayed University of Artificial Intelligence · University of Sydney · University of Oxford +3
SAMBAR: Selective Anchoring via Method of Multipliers for Balanced Knowledge Acquisition and Retention in Vision-Language-Action Models
Sep 26, 2026Aayushi Shrivastava, Xunlan Zhou, Hongrui Zhao +2Continual Learning for VLA Models Vision-Language-Action Models
Department of Mechanical Engineering, University of California Berkeley, Berkeley, CA 94709, USA
Brain-Inspired Hierarchical Modularity for General Continual Learning
Sep 21, 2026Hongwei Yan, Kanglei Zhou, Qi Cheng +7Continual Learning for LLMs Continual Learning for VLA Models
School of Life Sciences, Tsinghua University, Beijing, China. · IDG/McGovern Institute for Brain Research, Tsinghua University, Beijing, China. · Tsinghua-Peking Center for Life Sciences, Beijing, China. +2
FAN: Foresight Action Normalization for Continual Adaptation of Vision-Language-Action Models
Sep 18, 2026Yijun Hong, Jiarun Zhu, Xiaoquan Sun +6Continual Learning for VLA Models Vision-Language-Action Models
HKU · SUSTech · EIT, Ningbo +2
OrthoSkillVLA: Continual Skill Learning via Gradient-Informed Skill Subspace Adaptation
Aug 20, 2026Jiaqi Wang, Zhou Fang, Qiongfeng Shi +1Continual Learning for VLA Models Continual Robot Learning
School of Computer Science and Engineering, Southeast University, China · Key Laboratory of New Generation Artificial Intelligence Technology and Its Interdisciplinary Applications, Ministry of Education, China · School of Electronic Science and Engineering, Southeast University, China
StellaVLA: In-Context Structured Demonstration for Generalizable Vision-Language-Action Models
Aug 12, 2026Siyu Xu, Yunke Wang, Zijian Wang +6Reasoning for VLA Models Continual Learning for VLA Models
StellarEdge AI Technical Team
In-Context VLA: Endowing Vision-Language-Action Models with Language via In-Context Post-Training and Agentic Tool Use
Aug 6, 2026Jiarui Yang, Wen Huang, Jiale Zhang +2Reasoning for VLA Models ICL for Robotics
Nankai University · Tsinghua University · Swiss Federal Institute of Technology in Lausanne (EPFL)
Towards Human-like Physical Intelligence: LifelongVision-Language-Action Learning for Robotic Manipulation
Jul 16, 2026Yao He, Gan Sun, Wenqi Liang +2Continual Robot Learning Robot Skill Learning
South China University of Technology · University of Trento
RECALL: Recovery Experience Collection for Active Lifelong Learning in Vision-Language-Action Models
Jun 22, 2026Ulas Berk Karli, Tesca FitzgeraldContinual Robot Learning Robot Skill Learning
Department of Computer Science Yale University United States
PHASER: Phase-Aware and Semantic Experience Replay for Vision-Language-Action Models
Jun 2, 2026Ziyang Chen, Shaoguang Wang, Weiyu Guo +5Continual Learning for VLA Models Robotic Manipulation
Thrust of AI, HKUST(Guangzhou), Guangzhou, China · AI2 Robotics, Shenzhen, China
Can VLA Models Learn from Real-World Data Continually without Forgetting?
May 26, 2026Jiarun Zhu, Yijun Hong, Xiaoquan Sun +5Continual Robot Learning Continual Learning for VLA Models
1HKU · 5SUSTech · 2INFIFORCE +3
Preserving Foundational Capabilities in Flow-Matching VLAs through Conservative SFT
May 9, 2026Tianyi Zhang, Shaopeng Zhai, Haoran Zhang +2Supervised Fine-Tuning Continual Learning for VLA Models
Shanghai Artificial Intelligence Laboratory
Sentinel-VLA: A Metacognitive VLA Model with Active Status Monitoring for Dynamic Reasoning and Error Recovery
May 2, 2026Wenhao Li, Xiu Su, Dan Niu +6Continual Learning for LLMs Robotics
University of Sydney · Central South University · Southeast University +2
Simple Recipe Works: Vision-Language-Action Models are Natural Continual Learners with Reinforcement Learning
Mar 12, 2026Jiaheng Hu, Jay Shim, Chen Tang +4Continual Learning for VLA Models VLA Model Adaptation
UT Austin · UCLA · NTU +1
AtomicVLA: Unlocking the Potential of Atomic Skill Learning in Robots
Mar 8, 2026Likui Zhang, Tao Tang, Zhihao Zhan +9Robot Skill Learning VLMs for Robotics
Sun Yat-sen University · Yinwang Intelligent Technology Co. Ltd. · Peng Cheng Laboratory
Information-Theoretic Constraints for Continual Vision-Language-Action Alignment
Mar 6, 2026Libang Zhao, Qixin Zeng, Hongyin Zhang +1Continual Learning for VLA Models Vision-Language Alignment
Westlake University, China · University of Southampton, UK
Pretrained Vision-Language-Action Models are Surprisingly Resistant to Forgetting in Continual Learning
Mar 4, 2026Huihan Liu, Changyeon Kim, Bo Liu +2Continual Robot Learning Continual Learning for VLA Models
The University of Texas at Austin · KAIST · Microsoft Superintelligence.
Continually Evolving Skill Knowledge in Vision Language Action Model
Nov 22, 2025Yuxuan Wu, Guangming Wang, Zhiheng Yang +4Continual Learning for VLA Models Continual Robot Learning
Shanghai Jiao Tong University · Shanghai Innovation Institute · University of Cambridge +4

Oct 17, 2026J/K move · Enter open · S save

SARAM: Sparse Adapter Allocation with Reuse and Merging for Lifelong Robot Learning

SeongJun Jeong, Chungwoo Lee, Gi-Cheon Kang, Byoung-Tak Zhang

Seoul National University · Ajou University

Adapter-Based Continual Learning Continual Robot Learning Continual Learning for VLA Models Storage-Efficient Continual Learning

Open paper AskPDF arXiv

Lifelong robot learning requires acquiring new tasks sequentially without forgetting earlier ones. With large pre-trained vision-language-action (VLA) backbones, adapter-based methods effectively mitigate this challenge. However, their task-specific adapters and selection modules accumulate with the task stream, creating a scalability challenge in on-disk storage and deployment cost under fixed onboard resource budgets. We propose SARAM, a lifelong learning framework that limits storage growth while keeping GPU memory and per-step inference cost independent of the task stream length. For each new task, SARAM reuses frozen adapters from previous tasks and introduces new adapters only at sparsely selected layers, with a duplication penalty encouraging them to capture information not already represented by the adapter library. At deployment, a compact signature derived from backbone activation statistics identifies the task once per episode, and the corresponding task update is merged into the backbone, leaving only the backbone in the control loop. Across four LIBERO suites, SARAM matches or exceeds adapter-based baselines while adding up to 4.5x less on-disk state per task. GPU memory usage and control frequency remain comparable to those of the backbone as the task stream grows. Experiments on a real UR5 arm with a different backbone further demonstrate retention throughout a sequential task stream.