Turning Back Without Forgetting: Selective Backward Refinement for Parameter-Efficient Continual Learning

• URL: http://arxiv.org/abs/2606.01379v2
• Date: Sun, 07 Jun 2026 14:03:46 GMT
• Title: Turning Back Without Forgetting: Selective Backward Refinement for Parameter-Efficient Continual Learning
• Abstract summary: We propose a replay-free framework that enables controlled backward transfer in prompt-based continual learning.<n>SABER determines when backward refinement is beneficial using complementary task-correlation criteria.<n>Experiments across multiple continual learning benchmarks and diverse pretrained backbones, including T5-Large, LLaMA, and Qwen, demonstrate that SABER consistently achieves positive backward transfer.
• Score: 22.058497718429518
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: While prompt-based parameter-efficient continual learning mitigates catastrophic forgetting by isolating task-specific prompts, this isolation also limits later tasks from improving earlier ones, leaving backward knowledge transfer underexplored. We address this limitation by proposing Selective bAckward refinement for positive Backward knowledge transfER (SABER), a replay-free framework that enables controlled backward transfer in prompt-based continual learning. SABER determines when backward refinement is beneficial using complementary task-correlation criteria based on prompt-gradient geometry and loss-distribution similarity, and how to perform refinement safely by restricting updates to non-interfering directions in the prompt parameter space. Extensive experiments across multiple continual learning benchmarks and diverse pretrained backbones, including T5-Large, LLaMA, and Qwen, demonstrate that SABER consistently achieves positive backward transfer while maintaining strong overall average performance. Code is available at https://github.com/OptMN-Lab/SABER-ICML-2026/.

Related papers

• Rethinking Transfer in Continual Learning: A Replay-Based Realisation [14.73400672149583]
  Continual learning studies how language models can continually acquire new tasks without expensive retraining from scratch.<n>Existing methods, whether rehearsal-based (replaying stored past data) or rehearsal-free (regularising or isolating parameters), overwhelmingly target one objective: preventing catastrophic forgetting.<n>We propose Transfer-Selective Replay (TSR), which selects replay data predicted to benefit the incoming task rather than replaying past examples indiscriminately.
  arXiv  Detail & Related papers  (2026-07-17T03:18:17Z)
• Hidden Forgetting in Continual Multimodal Learning: When Accuracy Survives but Grounding Fails [6.708653593551749]
  We study whether a continually adapted MLLM can preserve not only what it answers, but how it uses visual, textual, OCR, chart, and document evidence.<n>We identify emphhidden evidence-use forgetting, where answer accuracy is retained while the model silently shifts toward different or less grounded evidence channels.<n>We propose textscRCL, a replay-free reliance-constrained continual learning framework.
  arXiv  Detail & Related papers  (2026-07-02T10:50:37Z)
• MIITA: Memory-Induced Inference-Time Adaptation for Continual Learning with Small Language Models [64.2820121997882]
  MIITA is a Memory-Induced Inference-Time Adaptation framework for supervised CL under constrained storage.<n>It stores supervised experiences as compact correction-direction prototypes with semantic anchors, and retrieves them at inference time using semantic and uncertainty-based cues.<n>Experiments show that MIITA consistently improves final performance and mitigates forgetting under fixed memory budgets.
  arXiv  Detail & Related papers  (2026-05-20T03:03:04Z)
• Task Switching Without Forgetting via Proximal Decoupling [43.88767637201092]
  In continual learning, the primary challenge is to learn new information without forgetting old knowledge.<n>A common solution addresses this trade-off through regularization, penalizing changes to parameters critical for previous tasks.<n>We propose a different approach that separates task learning from stability enforcement via operator splitting.
  arXiv  Detail & Related papers  (2026-04-20T21:28:13Z)
• Spectral Imbalance Causes Forgetting in Low-Rank Continual Adaptation [58.3773038915023]
  Continual learning aims to adapt pre-trained models to sequential tasks without forgetting previously acquired knowledge.<n>Most existing approaches treat continual learning as avoiding interference with past updates, rather than considering what properties make the current task-specific update naturally preserve previously acquired knowledge.<n>We address this problem using a projected first-order method compatible with standard deep-dots used in vision-language models.
  arXiv  Detail & Related papers  (2026-01-31T13:27:02Z)
• In-Context Reinforcement Learning through Bayesian Fusion of Context and Value Prior [53.21550098214227]
  In-context reinforcement learning promises fast adaptation to unseen environments without parameter updates.<n>We introduce SPICE, a Bayesian ICRL method that learns a prior over Q-values via deep ensemble and updates this prior at test-time.<n>We prove that SPICE achieves regret-optimal behaviour in both bandits and finite-horizon MDPs, even when pretrained only on suboptimal trajectories.
  arXiv  Detail & Related papers  (2026-01-06T13:41:31Z)
• Dynamic Mixture of Progressive Parameter-Efficient Expert Library for Lifelong Robot Learning [64.88132446226474]
  A generalist agent must continuously learn and adapt throughout its lifetime, achieving efficient forward transfer while minimizing forgetting.<n>Previous work has explored parameter-efficient fine-tuning for single-task adaptation, effectively steering a frozen pretrained model with a small number of parameters.<n>We propose Dynamic Mixture of Progressive Efficient Expert Library (DMPEL) for lifelong robot learning.
  arXiv  Detail & Related papers  (2025-06-06T11:13:04Z)
• SLCA++: Unleash the Power of Sequential Fine-tuning for Continual Learning with Pre-training [68.7896349660824]
  We present an in-depth analysis of the progressive overfitting problem from the lens of Seq FT. Considering that the overly fast representation learning and the biased classification layer constitute this particular problem, we introduce the advanced Slow Learner with Alignment (S++) framework. Our approach involves a Slow Learner to selectively reduce the learning rate of backbone parameters, and a Alignment to align the disjoint classification layers in a post-hoc fashion.
  arXiv  Detail & Related papers  (2024-08-15T17:50:07Z)
• SHERL: Synthesizing High Accuracy and Efficient Memory for Resource-Limited Transfer Learning [63.93193829913252]
  We propose an innovative METL strategy called SHERL for resource-limited scenarios. In the early route, intermediate outputs are consolidated via an anti-redundancy operation. In the late route, utilizing minimal late pre-trained layers could alleviate the peak demand on memory overhead.
  arXiv  Detail & Related papers  (2024-07-10T10:22:35Z)
• Beyond Anti-Forgetting: Multimodal Continual Instruction Tuning with Positive Forward Transfer [21.57847333976567]
  Multimodal Continual Instruction Tuning (MCIT) enables Multimodal Large Language Models (MLLMs) to meet continuously emerging requirements without expensive retraining. MCIT faces two major obstacles: catastrophic forgetting (where old knowledge is forgotten) and negative forward transfer. We propose Prompt Tuning with Positive Forward Transfer (Fwd-Prompt) to address these issues.
  arXiv  Detail & Related papers  (2024-01-17T12:44:17Z)
• AdaS: Adaptive Scheduling of Stochastic Gradients [50.80697760166045]
  We introduce the notions of textit"knowledge gain" and textit"mapping condition" and propose a new algorithm called Adaptive Scheduling (AdaS) Experimentation reveals that, using the derived metrics, AdaS exhibits: (a) faster convergence and superior generalization over existing adaptive learning methods; and (b) lack of dependence on a validation set to determine when to stop training.
  arXiv  Detail & Related papers  (2020-06-11T16:36:31Z)

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.