mlsanityData to 14 Oct 2026

Reasoning Portability: Guiding Continual Learning for MLLMs in the RLVR Era

arXiv 2605.18903 ↗First posted 17 May 2026cs.LG cs.CVcited by 0

On the map: LLM reasoning → Multimodal reasoning

also in 1 more
Abstract (arXiv)

Vision-Language Models in Continual Learning (VLM-CL) aim to continuously adapt to new multimodal tasks while retaining prior knowledge. The emerging paradigm that couples Multimodal Large Language Models (MLLMs) with Reinforcement Learning with Verifiable Rewards (RLVR) calls for a new pattern to guide continual adaptation. Advances in reasoning capability now make it feasible to impose constraints at the reasoning level. We formalize portability, a sample-level measure of how reusable the previous policy's behavior is on a new task, and empirically show that reasoning-level signals remain reliable on out-of-distribution samples while answer-level signals do not. We instantiate this as Reasoning Portability (RP) and propose Reasoning-based Dynamic Balance Continual Learning (RDB-CL), which modulates the per-sample Kullback-Leibler regularization in RLVR according to RP: a tight anchor preserves reusable reasoning on high-RP samples, while a relaxed anchor on low-RP samples permits exploration of new reasoning pathways. Experiments show that RDB-CL consistently outperforms baselines, improving Last accuracy by +12.0% over the vanilla RLVR baseline.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Multimodal reasoning, one of 17 directions in LLM reasoning. Builds on 47 of its 70 references that are in the map. First posted on arXiv 17 May 2026.

Builds on

47 of 70 references.

Closest ancestors · 8

Same direction, newest first

  1. 2025Look-Back: Implicit Visual Re-focusing in MLLM Reasoning2 Jul
  2. 2025DeepEyes: Incentivizing "Thinking with Images" via Reinforcement Learning20 May
  3. 2025VL-Rethinker: Incentivizing Self-Reflection of Vision-Language Models with Reinforcement Learning10 Apr
  4. 2025Perception-R1: Pioneering Perception Policy with Reinforcement Learning10 Apr
  5. 2025Reason-RFT: Reinforcement Fine-Tuning for Visual Reasoning of Vision Language Models26 Mar
  6. 2025MM-Eureka: Exploring the Frontiers of Multimodal Reasoning with Rule-based Reinforcement Learning10 Mar
  7. 2025Imagine while Reasoning in Space: Multimodal Visualization-of-Thought13 Jan
  8. 2024Fine-Tuning Large Vision-Language Models as Decision-Making Agents via Reinforcement Learning16 May

Other references · 39

  1. 2020Language Models are Few-Shot Learnerscited by 32,419
  2. 2017Proximal Policy Optimization Algorithmscited by 15,215
  3. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 9,379
  4. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 8,065
  5. 2022Self-Consistency Improves Chain of Thought Reasoning in Language Modelscited by 5,708
  6. 2023InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuningcited by 2,876
  7. 2025DAPO: An Open-Source LLM Reinforcement Learning System at Scalecited by 2,810
  8. 2016Learning without Forgettingcited by 2,496
  9. 2022Language Models (Mostly) Know What They Knowcited by 1,856
  10. 2024Tulu 3: Pushing Frontiers in Open Language Model Post-Trainingcited by 1,053
  11. 2025Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?cited by 982
  12. 2018VizWiz Grand Challenge: Answering Visual Questions from Blind Peoplecited by 858
27 more
  1. 2024DeepSeek-V3 Technical Reportcited by 811
  2. 2022Large Language Models Can Self-Improvecited by 664
  3. 2025The Entropy Mechanism of Reinforcement Learning for Reasoning Language Modelscited by 485
  4. 2025R1-VL: Learning to Reason with Multimodal Large Language Models via Step-wise Group Relative Policy Optimizationcited by 301
  5. 2024Boosting Continual Learning of Vision-Language Models via Mixture-of-Experts Adapterscited by 199
  6. 2025RL's Razor: Why Online Reinforcement Learning Forgets Lesscited by 176
  7. 2025The Surprising Effectiveness of Negative Reinforcement in LLM Reasoningcited by 162
  8. 2024RL-VLM-F: Reinforcement Learning from Vision Language Foundation Model Feedbackcited by 156
  9. 2025Retaining by Doing: The Role of On-Policy Data in Mitigating Forgettingcited by 94
  10. 2024Internal Consistency and Self-Feedback in Large Language Models: A Surveycited by 50
  11. 2024Self-Consistency Preference Optimizationcited by 50
  12. 2025Weight Ensembling Improves Reasoning in Language Modelscited by 38
  13. 2024Mind the Interference: Retaining Pre-trained Knowledge in Parameter Efficient Continual Learning of Vision-Language Modelscited by 34
  14. 2025Unlocking Exploration in RLVR: Uncertainty-aware Advantage Shaping for Deeper Reasoningcited by 34
  15. 2025HiDe-LLaVA: Hierarchical Decoupling for Continual Instruction Tuning of Multimodal Large Language Modelcited by 31
  16. 2024Uncertainty-Aware Evaluation for Vision-Language Modelscited by 29
  17. 2025MLLM-CL: Continual Learning for Multimodal Large Language Modelscited by 23
  18. 2024TrustScore: Reference-Free Evaluation of LLM Response Trustworthinesscited by 19
  19. 2025Continual Learning for VLMs: A Survey and Taxonomy Beyond Forgettingcited by 18
  20. 2023Continual Multimodal Knowledge Graph Constructioncited by 15
  21. 2024Advancing Cross-domain Discriminability in Continual Learning of Vision-Language Modelscited by 15
  22. 2025Why Reinforcement Fine-Tuning Enables MLLMs Preserve Prior Knowledge Better: A Data Perspectivecited by 14
  23. 2024Uncertainty Aware Learning for Language Model Alignmentcited by 13
  24. 2024Adapt-∞: Scalable Continual Multimodal Instruction Tuning via Dynamic Data Selectioncited by 9
  25. 2025Self-Training Large Language Models with Confident Reasoningcited by 8
  26. 2024Multi-granularity Knowledge Transfer for Continual Reinforcement Learningcited by 6
  27. 2024Exploiting the Semantic Knowledge of Pre-trained Text-Encoders for Continual Learningcited by 6

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Beyond Forgetting: Diagnosing and Harnessing Shared Reasoning in Continual RLVR
  2. 2025Reinforced MLLM: A Survey on RL-Based Reasoning in Multimodal Large Language Models
  3. 2026CORA: Analyzing and bridging thinking-answer gap in Multimodal RLVR via Consistency-Oriented Reasoning Alignment
  4. 2025Observe-R1: Unlocking Reasoning Abilities of MLLMs with Dynamic Progressive Reinforcement Learning
  5. 2026RL Forgets! Towards Continual Policy Optimization
  6. 2025Perception-R1: Advancing Multimodal Reasoning Capabilities of MLLMs via Visual Perception Reward
  7. 2026Bridging Perception and Reasoning: Token Reweighting for RLVR in Multimodal LLMs
  8. 2025Enhancing the Outcome Reward-based RL Training of MLLMs with Self-Consistency Sampling