Author name cluster

Minghu Wang

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

2 papers

2 author rows

AAAI Conference 2026 Conference Paper

ALEX:A Light Editing-knowledge Extractor

Minghu Wang
ShuLiang Zhao
Yuanyuan Zhao
Hongxia Xu

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.

PDF Details DOI

ECAI Conference 2025 Conference Paper

ReAligner: Knowledge Editing via Semantic Refinement and Representation Diversification

Minghu Wang
Shuliang Zhao
Yuqing Li
Mengjun Yang
Mei He

Large language models (LLMs) have made remarkable progress in natural language processing tasks. However, efficiently integrating new knowledge and accurately answering multi-hop questions remain significant challenges. Existing methods often struggle with question semantic understanding and knowledge integration during reasoning. This paper proposes the ReAligner to address these issues. The ReAligner comprises three key components. The Multi-Representation Generator generates multiple semantically rich question representations by applying synonym replacement, syntactic variation, and semantic expansion, enhancing the model’s understanding of multi-hop questions. The Knowledge Relevance Filter filters out irrelevant facts from large fact sets, improving retrieval precision and computational efficiency. The Semantic Alignment Predictor determines the most relevant fact for each sub-question by leveraging the Semantic Sharpening Unit and a trained classifier. We evaluate the ReAligner model using four datasets and compare it with state-of-the-art methods. Experimental results show that the ReAligner model outperforms existing methods.

Details