A Large Language Model-Enhanced Knowledge Graph Multi-hop Reasoning Method for Assembly Process Question-Answering
Peilin Shao, Yifan Qie, Chao Chen, Nabil Anwer, Zhujia Li, Zhicheng Huang, Lihong Qiao, Xinzheng Xu, Yongqiang Wan
Journal of Computing and Information Science in Engineering
Problems Identified (1)
complex-query-understanding: Knowledge graph question-answering for assembly processes has difficulty understanding complex queries because assembly processes are highly complex and specialized.
Proposed Solutions (3)
llm-enhanced-kg-multihop-reasoning: The paper proposes an LLM-enhanced knowledge graph multi-hop reasoning method for assembly process question-answering.
qa-chain-decomposition: The method decomposes multi-hop knowledge graph question-answering into question-answering chain generation, multi-hop chain reasoning, and natural language answer generation subtasks.
graph-path-chain-reasoning: The method constructs a graph-path-based multi-hop reasoning model using core chain reasoning, attribute constraints, and task-oriented negative sampling.
Results (1)
comparative-effectiveness-validated:
Research Domain
Assembly process knowledge graph question-answering