ALEX:A Light Editing-knowledge Extractor

2025-11-17

AI Summary

To address the static knowledge limitation of large language models (LLMs) and their difficulty in adapting to dynamic information, this paper proposes ALEX, a lightweight knowledge editing framework. Its core innovation is a novel hierarchical memory architecture that organizes knowledge via semantic clustering and integrates an inference-aware query synthesis (IQS) module with a dynamic evidence adjudication (DEA) engine, enabling efficient two-stage retrieval and multi-step reasoning alignment. This design reduces knowledge retrieval complexity from O(N) to O(K + N/C), significantly enhancing scalability and efficiency. On the MQUAKE benchmark, ALEX achieves substantial improvements in both multi-hop answer accuracy (MultiHop-ACC) and hop-wise reasoning path reliability (HopWise-ACC), while compressing the search space by over 80%. The framework thus provides a scalable, high-precision, and lightweight solution for dynamic knowledge updating.

Technology Category

• Knowledge Representation and Reasoning: Knowledge Acquisition
• Data Mining & Knowledge Management: Intelligent Query Processing
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

The static nature of knowledge within Large Language Models (LLMs) makes it difficult for them to adapt to evolving information, rendering knowledge editing a critical task. However, existing methods struggle with challenges of scalability and retrieval efficiency, particularly when handling complex, multi-hop questions that require multi-step reasoning. To address these challenges, this paper introduces ALEX (A Light Editing-knowledge Extractor), a lightweight knowledge editing framework. The core innovation of ALEX is its hierarchical memory architecture, which organizes knowledge updates (edits) into semantic clusters. This design fundamentally reduces retrieval complexity from a linear O(N) to a highly scalable O(K+N/C). Furthermore, the framework integrates an Inferential Query Synthesis (IQS) module to bridge the semantic gap between queries and facts , and a Dynamic Evidence Adjudication (DEA) engine that executes an efficient two-stage retrieval process. Experiments on the MQUAKE benchmark demonstrate that ALEX significantly improves both the accuracy of multi-hop answers (MultiHop-ACC) and the reliability of reasoning paths (HopWise-ACC). It also reduces the required search space by over 80% , presenting a promising path toward building scalable, efficient, and accurate knowledge editing systems.

Problem

• Addresses scalability and efficiency challenges in editing LLM knowledge
• Improves multi-hop question answering accuracy through hierarchical memory
• Reduces retrieval complexity from linear O(N) to scalable O(K+N/C)

Innovation

• Hierarchical memory architecture organizes knowledge edits
• Inferential Query Synthesis bridges query-fact semantic gap
• Dynamic Evidence Adjudication enables two-stage retrieval process