CRAFT: Forgetting-Aware Intervention-Based Adaptation for Continual Learning

• URL: http://arxiv.org/abs/2605.05732v2
• Date: Fri, 08 May 2026 04:10:34 GMT
• Title: CRAFT: Forgetting-Aware Intervention-Based Adaptation for Continual Learning
• Abstract summary: CRAFT is a continual learning framework that avoids updating model weights.<n>It first routes each task to a group of similar tasks based on output-distribution divergence.<n>It then fine-tunes the model using a Kullback-Leibler (KL) divergence against the group's prior state.<n>Finally, it merges interventions for the updated task into the shared representation using the same KL signal.
• Score: 7.180013165859104
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large language models (LLMs) can acquire new capabilities through fine-tuning, but continual adaptation often leads to catastrophic forgetting. We propose CRAFT, a continual learning framework that avoids updating model weights by instead learning low-rank interventions on hidden representations. CRAFT proceeds in three stages: it first routes each task to a group of similar tasks based on output-distribution divergence; it then fine-tunes the model using a Kullback-Leibler (KL) divergence against the group's prior state, which directly controls forgetting and determines convergence; finally, it merges interventions for the updated task into the shared representation using the same KL signal. This design unifies routing, regularization, and merging through a single KL-based objective. CRAFT improves overall performance and reduces forgetting compared to strong LoRA-based approaches across multiple benchmarks and model scales, while remaining robust to task ordering. These results suggest that controlling adaptation in representation space, guided by output-space divergence, provides a scalable and principled approach to continual learning in LLMs.

Related papers

• RW-LoRA: Communication-Efficient Decentralized LoRA Fine-Tuning via Random Walks [20.21667584370137]
  Fine-tuning computation methods such as LoRA have become a standard approach for adapting large foundation models.<n>Most existing distributed LoRA methods rely on centralized aggregation, and gossip-based decentralized LoRA requires repeated synchronization among multiple model copies.<n>In this paper, we propose a random-based LoRA finetuning scheme. Instead of maintaining multiple model replicas, a single model token traverses the network is updated sequentially.
  arXiv  Detail & Related papers  (2026-08-31T08:24:24Z)
• dVLA-RL: Reinforcement Learning over Denoising Trajectories for Discrete Diffusion Vision-Language-Action Models [49.497309561043004]
  We propose textbfdVLA-RL, shifting the learning objective from the marginal action probability to the joint probability of the sampled generation path.<n>Our approach achieves a success rate of textbf99.7% on LIBERO.<n>It also establishes strong VLA-based results on RoboTwin 2.0 by delivering a textbf30.6% improvement over the SFT baseline.
  arXiv  Detail & Related papers  (2026-06-22T17:19:03Z)
• Reinforcing Structured Chain-of-Thought for Video Understanding [49.3086326382764]
  We introduce Summary-Driven Reinforcement Learning (SDRL)<n>SDRL is a single-stage RL framework that obviates the need forSupervised Fine-Tuning (SFT)<n>Our method achieves state-of-the-art performance on seven public VideoQA datasets.
  arXiv  Detail & Related papers  (2026-03-26T22:11:14Z)
• SLE-FNO: Single-Layer Extensions for Task-Agnostic Continual Learning in Fourier Neural Operators [0.4369550829556577]
  New experimental conditions or simulation regimes may differ significantly, requiring extrapolation and model updates without re-access to prior data.<n>This creates a need for continual learning (CL) frameworks that can adapt to distribution shifts while preventing catastrophic forgetting.<n>We introduce a new architecture-based approach (SLE-FNO) combining a Single-Layer Extension (SLE) with the Fourier Neural Operator (FNO) to support efficient CL.
  arXiv  Detail & Related papers  (2026-03-20T18:30:38Z)
• Decomposing and Composing: Towards Efficient Vision-Language Continual Learning via Rank-1 Expert Pool in a Single LoRA [50.97792275353563]
  We introduce a novel framework that restructures a single Low-Rank Adaptation (LoRA) module as a decomposable Rank-1 Expert Pool.<n>Our method learns to dynamically compose a sparse, task-specific update by selecting from this expert pool, guided by the semantics of the [Guided] token.
  arXiv  Detail & Related papers  (2026-01-30T10:54:51Z)
• Model Merging via Multi-Teacher Knowledge Distillation [11.543771846135021]
  We introduce a novel flatness-aware PAC-Bayes generalization bound specifically for the model merging setting.<n>We frame model merging as multi-teacher knowledge distillation on scarce, unlabeled data.<n>We formally demonstrate that minimizing the student-teacher Kullback-Leibler divergence directly tightens the upper bound on the merged model's excess risk.
  arXiv  Detail & Related papers  (2025-12-24T17:10:44Z)
• Steering Vision-Language-Action Models as Anti-Exploration: A Test-Time Scaling Approach [78.4812458793128]
  We propose textbfTACO, a test-time-scaling framework that applies a lightweight pseudo-count estimator as a high-fidelity verifier of action chunks.<n>Our method resembles the classical anti-exploration principle in offline reinforcement learning (RL), and being gradient-free, it incurs significant computational benefits.
  arXiv  Detail & Related papers  (2025-12-02T14:42:54Z)
• Dual-Stream Diffusion for World-Model Augmented Vision-Language-Action Model [62.889356203346985]
  We propose DUal-STream diffusion (DUST), a world-model augmented VLA framework that handles the modality conflict.<n>DUST achieves up to 6% gains over a standard VLA baseline and implicit world-modeling methods.<n>On real-world tasks with the Franka Research 3, DUST outperforms baselines in success rate by 13%.
  arXiv  Detail & Related papers  (2025-10-31T16:32:12Z)
• STABLE: Gated Continual Learning for Large Language Models [0.0]
  STABLE is a gated continual self editing framework that constrains forgetting during sequential updates.<n>Each candidate edit is evaluated against a stability budget using one of three metrics.<n>Experiments on the Qwen-2.5-7B model show that gating effectively mitigates forgetting while preserving adaptability.
  arXiv  Detail & Related papers  (2025-10-17T16:14:05Z)
• Train with Perturbation, Infer after Merging: A Two-Stage Framework for Continual Learning [57.514786046966265]
  We propose textbfPerturb-and-Merge (P&M), a novel continual learning framework that integrates model merging into the CL paradigm to mitigate forgetting.<n>Our proposed approach achieves state-of-the-art performance on several continual learning benchmark datasets.
  arXiv  Detail & Related papers  (2025-05-28T14:14:19Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.