LoRA-Loop: Closing the Synthetic Replay Cycle for Continual VLM Learning

• URL: http://arxiv.org/abs/2507.13568v2
• Date: Tue, 29 Jul 2025 03:29:57 GMT
• Title: LoRA-Loop: Closing the Synthetic Replay Cycle for Continual VLM Learning
• Authors: Kaihong Wang, Donghyun Kim, Margrit Betke,
• Abstract summary: Continual learning for vision-language models has achieved remarkable performance through synthetic replay.<n>However, real-world downstream applications often exhibit domain-specific nuances and fine-grained semantics not captured by generators.<n>We propose a LoRA-enhanced synthetic-replay framework that injects task-specific low-rank adapters into a frozen Stable Diffusion model.
• Score: 15.385906975878628
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Continual learning for vision-language models has achieved remarkable performance through synthetic replay, where samples are generated using Stable Diffusion to regularize during finetuning and retain knowledge. However, real-world downstream applications often exhibit domain-specific nuances and fine-grained semantics not captured by generators, causing synthetic-replay methods to produce misaligned samples that misguide finetuning and undermine retention of prior knowledge. In this work, we propose a LoRA-enhanced synthetic-replay framework that injects task-specific low-rank adapters into a frozen Stable Diffusion model, efficiently capturing each new task's unique visual and semantic patterns. Specifically, we introduce a two-stage, confidence-based sample selection: we first rank real task data by post-finetuning VLM confidence to focus LoRA finetuning on the most representative examples, then generate synthetic samples and again select them by confidence for distillation. Our approach integrates seamlessly with existing replay pipelines-simply swap in the adapted generator to boost replay fidelity. Extensive experiments on the Multi-domain Task Incremental Learning (MTIL) benchmark show that our method outperforms previous synthetic-replay techniques, achieving an optimal balance among plasticity, stability, and zero-shot capability. These results demonstrate the effectiveness of generator adaptation via LoRA for robust continual learning in VLMs.

Related papers

• Hybrid Autoregressive-Diffusion Model for Real-Time Streaming Sign Language Production [0.0]
  We introduce a hybrid approach combining autoregressive and diffusion models to generate Sign Language Production (SLP) models.<n>To capture fine-grained body movements, we design a Multi-Scale Pose Representation module that separately extracts detailed features from distinct arttors.<n>We also introduce a Confidence-Aware Causal Attention mechanism that utilizes joint-level confidence scores to dynamically guide the pose generation process.
  arXiv  Detail & Related papers  (2025-07-12T01:34:50Z)
• Taming Flow Matching with Unbalanced Optimal Transport into Fast Pansharpening [10.23957420290553]
  We propose the Optimal Transport Flow Matching framework to achieve one-step, high-quality pansharpening.<n>The OTFM framework enables simulation-free training and single-step inference while maintaining strict adherence to pansharpening constraints.
  arXiv  Detail & Related papers  (2025-03-19T08:10:49Z)
• Synthetic Data is an Elegant GIFT for Continual Vision-Language Models [52.343627275005026]
  GIFT is a novel continual fine-tuning approach to overcome catastrophic forgetting in Vision-Language Models.<n>We employ a pre-trained diffusion model to recreate both pre-training and learned downstream task data.<n>Our method consistently outperforms previous state-of-the-art approaches across various settings.
  arXiv  Detail & Related papers  (2025-03-06T09:09:18Z)
• Exploring Representation-Aligned Latent Space for Better Generation [86.45670422239317]
  We introduce ReaLS, which integrates semantic priors to improve generation performance.<n>We show that fundamental DiT and SiT trained on ReaLS can achieve a 15% improvement in FID metric.<n>The enhanced semantic latent space enables more perceptual downstream tasks, such as segmentation and depth estimation.
  arXiv  Detail & Related papers  (2025-02-01T07:42:12Z)
• Unlocking Tuning-Free Few-Shot Adaptability in Visual Foundation Models by Recycling Pre-Tuned LoRAs [76.40876036912537]
  Large Language Models (LLMs) demonstrate strong few-shot adaptability without requiring fine-tuning.<n>Current Visual Foundation Models (VFMs) require explicit fine-tuning with sufficient tuning data.<n>We propose a framework, LoRA Recycle, that distills a meta-LoRA from diverse pre-tuned LoRAs with a meta-learning objective.
  arXiv  Detail & Related papers  (2024-12-03T07:25:30Z)
• Unleashing the Power of Large Language Models in Zero-shot Relation Extraction via Self-Prompting [21.04933334040135]
  We introduce the Self-Prompting framework, a novel method designed to fully harness the embedded RE knowledge within Large Language Models.<n>Our framework employs a three-stage diversity approach to prompt LLMs, generating multiple synthetic samples that encapsulate specific relations from scratch.<n> Experimental evaluations on benchmark datasets show our approach outperforms existing LLM-based zero-shot RE methods.
  arXiv  Detail & Related papers  (2024-10-02T01:12:54Z)
• Mitigating Catastrophic Forgetting in Large Language Models with Self-Synthesized Rehearsal [49.24054920683246]
  Large language models (LLMs) suffer from catastrophic forgetting during continual learning. We propose a framework called Self-Synthesized Rehearsal (SSR) that uses the LLM to generate synthetic instances for rehearsal.
  arXiv  Detail & Related papers  (2024-03-02T16:11:23Z)
• Take the Bull by the Horns: Hard Sample-Reweighted Continual Training Improves LLM Generalization [165.98557106089777]
  A key challenge is to enhance the capabilities of large language models (LLMs) amid a looming shortage of high-quality training data. Our study starts from an empirical strategy for the light continual training of LLMs using their original pre-training data sets. We then formalize this strategy into a principled framework of Instance-Reweighted Distributionally Robust Optimization.
  arXiv  Detail & Related papers  (2024-02-22T04:10:57Z)
• ReEval: Automatic Hallucination Evaluation for Retrieval-Augmented Large Language Models via Transferable Adversarial Attacks [91.55895047448249]
  This paper presents ReEval, an LLM-based framework using prompt chaining to perturb the original evidence for generating new test cases. We implement ReEval using ChatGPT and evaluate the resulting variants of two popular open-domain QA datasets. Our generated data is human-readable and useful to trigger hallucination in large language models.
  arXiv  Detail & Related papers  (2023-10-19T06:37:32Z)
• MRHER: Model-based Relay Hindsight Experience Replay for Sequential Object Manipulation Tasks with Sparse Rewards [11.79027801942033]
  We propose a novel model-based RL framework called Model-based Relay Hindsight Experience Replay (MRHER) MRHER breaks down a continuous task into subtasks with increasing complexity and utilizes the previous subtask to guide the learning of the subsequent one. We show that MRHER exhibits state-of-the-art sample efficiency in benchmark tasks, outperforming RHER by 13.79% and 14.29%.
  arXiv  Detail & Related papers  (2023-06-28T09:51:25Z)

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.