ScholarSum: Student-Teacher Abstractive Summarization via Knowledge Graph Reasoning and Reflective Refinement

2026-06-17

AI Summary

This work addresses the longstanding challenge in scientific document summarization of balancing linguistic fluency with factual faithfulness. The authors propose a hierarchical student–teacher framework that first constructs a hierarchical knowledge graph from the source document to capture its macro-level logical structure, which guides the student model in generating an initial draft. The teacher model then performs iterative fact-checking and reflective rewriting of the summary by leveraging retrieved evidence from the original text. Integrating knowledge graph reasoning, community structure analysis, and retrieval-augmented generation, the approach significantly enhances factual consistency while preserving fluent language. Experimental results demonstrate consistent improvements over existing baselines in both summary completeness and factual fidelity.

Abstract

Abstractive summarization plays a crucial role in enabling efficient understanding of scientific literature, yet it inherently demands both linguistic fluency and factual faithfulness. Existing approaches often fail to reconcile these two requirements. Extractive methods rely on rigid sentence splicing that disrupts macro-level logical coherence, while large language model (LLM)-based generative approaches, despite mastering linguistic fluency, exhibit limited factual consistency. In this work, we propose ScholarSum, a hierarchical reflective graph-based framework that emulates a student-teacher writing process for fluent and faithful scientific summarization. ScholarSum first organizes the document into a hierarchical knowledge graph by segmenting it into semantically coherent units, whose multi-layered community structure captures global logic and macro-level themes. Guided by this global structure, the student generates an initial draft, which is subsequently refined through fine-grained evidence retrieval. To ensure factual consistency, a teacher-like reviewer then iteratively examines the draft, identifies unsupported content, and prompts targeted re-retrieval and rewriting until the summary meets rigorous quality standards. Extensive experiments demonstrate that ScholarSum significantly outperforms previous baselines in terms of both completeness and faithfulness. Our code is available at https://github.com/Xiaoyu-Tao/ScholarSum.

Problem

• abstractive summarization
• factual faithfulness
• scientific literature
• linguistic fluency
• knowledge graph

Innovation

• knowledge graph reasoning
• reflective refinement
• student-teacher framework
• abstractive summarization
• factual faithfulness