FineFilter: A Fine-grained Noise Filtering Mechanism for Retrieval-Augmented Large Language Models

2025-02-17

AI Summary

In retrieval-augmented generation (RAG), retrieved documents often contain noise that obscures critical answer clues. To address this, we propose a sentence-level Min-Max optimization framework for fine-grained noise filtering. First, a context-aware clue extractor identifies the answer-bearing sentence. Second, a relevance re-ranker is trained using feedback from a generative module to improve discriminative capability. Third, a differentiable truncation optimizer dynamically prunes redundant content by minimizing the number of essential clues required for correct answer generation. This work introduces the first sentence-level Min-Max noise filtering paradigm, enabling modular, multi-stage fine-tuning. Evaluated on three QA benchmarks, our method achieves up to 11.3% absolute accuracy gain on complex reasoning tasks and reduces inference cost by up to 42%, significantly outperforming state-of-the-art baselines.

Technology Category

• Natural Language Processing: Question Answering
• Search and Optimization: Learning to Search
• Data Mining & Knowledge Management: Intelligent Query Processing

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Web Mining and Content Analysis: Community question answering

Abstract

Retrieved documents containing noise will hinder Retrieval-Augmented Generation (RAG) from detecting answer clues, necessitating noise filtering mechanisms to enhance accuracy.Existing methods use re-ranking or summarization to identify the most relevant sentences, but directly and accurately locating answer clues from these large-scale and complex documents remains challenging. Unlike these document-level operations, we treat noise filtering as a sentence-level MinMax optimization problem: first identifying the potential clues from multiple documents using contextual information, then ranking them by relevance, and finally retaining the least clues through truncation. In this paper, we propose FineFilter, a novel fine-grained noise filtering mechanism for RAG consisting of a clue extractor, a re-ranker, and a truncator. We optimize each module to tackle complex reasoning challenges: (1) Clue extractor firstly uses sentences containing the answer and similar ones as fine-tuned targets, aiming at extracting sufficient potential clues; (2) Re-ranker is trained to prioritize effective clues based on the real feedback from generation module, with clues capable of generating correct answer as positive samples and others as negative; (3) Truncator takes the minimum clues needed to answer the question (truncation point) as fine-tuned targets, and performs truncation on the re-ranked clues to achieve fine-grained noise filtering. Experiments on three QA datasets demonstrate that FineFilter significantly outperforms baselines in terms of performance and inference cost. Further analysis on each module shows the effectiveness of our optimizations for complex reasoning.

Problem

• Enhance accuracy in Retrieval-Augmented Generation
• Identify answer clues in complex documents
• Optimize sentence-level noise filtering mechanisms

Innovation

• Sentence-level MinMax optimization for noise filtering
• Fine-grained filtering via clue extractor and re-ranker
• Truncation based on minimal required answer clues