SelfAug: Mitigating Catastrophic Forgetting in Retrieval-Augmented Generation via Distribution Self-Alignment

• URL: http://arxiv.org/abs/2509.03934v1
• Date: Thu, 04 Sep 2025 06:50:47 GMT
• Title: SelfAug: Mitigating Catastrophic Forgetting in Retrieval-Augmented Generation via Distribution Self-Alignment
• Authors: Yuqing Huang, Rongyang Zhang, Qimeng Wang, Chengqiang Lu, Yan Gao, Yi Wu, Yao Hu, Xuyang Zhi, Guiquan Liu, Xin Li, Hao Wang, Enhong Chen,
• Abstract summary: Large language models (LLMs) have revolutionized natural language processing through their capabilities in understanding and executing diverse tasks.<n> supervised fine-tuning, particularly in Retrieval-Augmented Generation (RAG) scenarios, often leads to catastrophic forgetting.<n>We propose SelfAug, a self-distribution alignment method that aligns input sequence logits to preserve the model's semantic distribution.
• Score: 49.86376148975563
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Recent advancements in large language models (LLMs) have revolutionized natural language processing through their remarkable capabilities in understanding and executing diverse tasks. While supervised fine-tuning, particularly in Retrieval-Augmented Generation (RAG) scenarios, effectively enhances task-specific performance, it often leads to catastrophic forgetting, where models lose their previously acquired knowledge and general capabilities. Existing solutions either require access to general instruction data or face limitations in preserving the model's original distribution. To overcome these limitations, we propose SelfAug, a self-distribution alignment method that aligns input sequence logits to preserve the model's semantic distribution, thereby mitigating catastrophic forgetting and improving downstream performance. Extensive experiments demonstrate that SelfAug achieves a superior balance between downstream learning and general capability retention. Our comprehensive empirical analysis reveals a direct correlation between distribution shifts and the severity of catastrophic forgetting in RAG scenarios, highlighting how the absence of RAG capabilities in general instruction tuning leads to significant distribution shifts during fine-tuning. Our findings not only advance the understanding of catastrophic forgetting in RAG contexts but also provide a practical solution applicable across diverse fine-tuning scenarios. Our code is publicly available at https://github.com/USTC-StarTeam/SelfAug.

Related papers

• Enhancing Generalization of Depth Estimation Foundation Model via Weakly-Supervised Adaptation with Regularization [21.788680301776207]
  We propose WeSTAR, a parameter-efficient framework that performs Weakly supervised Self-Training Adaptation with Regularization.<n>We first adopt a dense self-training objective as the primary source of structural self-supervision.<n>To further improve robustness, we introduce semantically-aware hierarchical normalization.
  arXiv  Detail & Related papers  (2025-11-18T08:16:16Z)
• Did Models Sufficient Learn? Attribution-Guided Training via Subset-Selected Counterfactual Augmentation [61.248535801314375]
  Subset-Selected Counterfactual Augmentation (SS-CA)<n>We develop Counterfactual LIMA to identify minimal spatial region sets whose removal can selectively alter model predictions.<n>Experiments show that SS-CA improves generalization on in-distribution (ID) test data and achieves superior performance on out-of-distribution (OOD) benchmarks.
  arXiv  Detail & Related papers  (2025-11-15T08:39:22Z)
• Attributing Response to Context: A Jensen-Shannon Divergence Driven Mechanistic Study of Context Attribution in Retrieval-Augmented Generation [32.30660197797758]
  We introduce a novel Jensen-Shannon Divergence driven method to Attribute Response to Context (ARC-JSD)<n>We demonstrate superior accuracy and significant computational efficiency improvements compared to the previous surrogate-based method.<n>Our mechanistic analysis reveals specific attention heads and multilayer perceptron (MLP) layers responsible for context attribution.
  arXiv  Detail & Related papers  (2025-05-22T09:04:03Z)
• Continuous Visual Autoregressive Generation via Score Maximization [69.67438563485887]
  We introduce a Continuous VAR framework that enables direct visual autoregressive generation without vector quantization.<n>Within this framework, all we need is to select a strictly proper score and set it as the training objective to optimize.
  arXiv  Detail & Related papers  (2025-05-12T17:58:14Z)
• Generate, Discriminate, Evolve: Enhancing Context Faithfulness via Fine-Grained Sentence-Level Self-Evolution [61.80716438091887]
  GenDiE (Generate, Discriminate, Evolve) is a novel self-evolving framework that enhances context faithfulness through fine-grained sentence-level optimization.<n>By treating each sentence in a response as an independent optimization unit, GenDiE effectively addresses the limitations of previous approaches.<n>Experiments on ASQA (in-domain LFQA) and ConFiQA datasets demonstrate that GenDiE surpasses various baselines in both faithfulness and correctness.
  arXiv  Detail & Related papers  (2025-03-03T16:08:33Z)
• Distributional Refinement Network: Distributional Forecasting via Deep Learning [0.8142555609235358]
  A key task in actuarial modelling involves modelling the distributional properties of losses. We propose a Distributional Refinement Network (DRN), which combines an inherently interpretable baseline model with a flexible neural network. DRN captures varying effects of features across all quantiles, improving predictive performance while maintaining adequate interpretability.
  arXiv  Detail & Related papers  (2024-06-03T05:14:32Z)
• Self-Distillation Bridges Distribution Gap in Language Model Fine-Tuning [29.72531930169291]
  Self-Distillation Fine-Tuning (SDFT) is a novel approach that bridges the distribution gap by guiding fine-tuning with a distilled dataset generated by the model itself to match its original distribution. Experimental results on the Llama-2-chat model across various benchmarks demonstrate that SDFT effectively mitigates catastrophic forgetting.
  arXiv  Detail & Related papers  (2024-02-21T10:06:08Z)
• General Greedy De-bias Learning [163.65789778416172]
  We propose a General Greedy De-bias learning framework (GGD), which greedily trains the biased models and the base model like gradient descent in functional space. GGD can learn a more robust base model under the settings of both task-specific biased models with prior knowledge and self-ensemble biased model without prior knowledge.
  arXiv  Detail & Related papers  (2021-12-20T14:47:32Z)
• Generalized Zero-Shot Learning via VAE-Conditioned Generative Flow [83.27681781274406]
  Generalized zero-shot learning aims to recognize both seen and unseen classes by transferring knowledge from semantic descriptions to visual representations. Recent generative methods formulate GZSL as a missing data problem, which mainly adopts GANs or VAEs to generate visual features for unseen classes. We propose a conditional version of generative flows for GZSL, i.e., VAE-Conditioned Generative Flow (VAE-cFlow)
  arXiv  Detail & Related papers  (2020-09-01T09:12:31Z)
• Generalizing Variational Autoencoders with Hierarchical Empirical Bayes [6.273154057349038]
  We present Hierarchical Empirical Bayes Autoencoder (HEBAE), a computationally stable framework for probabilistic generative models. Our key contributions are two-fold. First, we make gains by placing a hierarchical prior over the encoding distribution, enabling us to adaptively balance the trade-off between minimizing the reconstruction loss function and avoiding over-regularization.
  arXiv  Detail & Related papers  (2020-07-20T18:18:39Z)

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.