计算机科学

计算机科学 ›› 2026, Vol. 53 ›› Issue (6A): 250400056-7.doi: 10.11896/jsjkx.250400056

• 人工智能 • 上一篇    下一篇

基于混合推理的人物科普教育认知智能体

郑家祺1, 彭世豪1, 赵俊杰2, 洪道诚1, 朱丹丹1, 桑晋秋1, 张桂戌1   

  1. 1 华东师范大学计算机科学与技术学院 上海 200062
    2 浙江工商大学信息与电子工程学院 杭州 310000
  • 出版日期:2026-06-16 发布日期:2026-06-12
  • 通讯作者: 洪道诚(hongdc@dase.ecnu.edu.cn)
  • 作者简介:(51275901093@stu.ecnu.edu.cn)
  • 基金资助:
    国家自然科学基金(62577023,61977025,L2224002)

Cognitive LLM Agent for Mater Education Based on Hybrid Reasoning

ZHENG Jiaqi1, PENG Shihao1, ZHAO Junjie2, HONG Daocheng1, ZHU Dandan1, SANG Jinqiu1, ZHANG Guixu1   

  1. 1 School of Computer Science and Technology,East China Normal University,Shanghai 200062,China
    2 School of Information and Electronic Engineering,Zhejiang Gongshang University,Hangzhou 310000,China
  • Published:2026-06-16 Online:2026-06-12
  • About author:ZHENG Jiaqi,born in 2002,postgra-duate.His main research interests include large language model and saliency prediction.
    HONG Daocheng,born in 1981,Ph.D,associate professor,is a member of CCF(No.E5370M).His main research interest is intelligent information proces-sing and application.
  • Supported by:
    National Natural Science Foundation of China(62577023,61977025,L2224002).

PDF (PC)

42

摘要： 大语言模型(Large Language Model)驱动的智能体技术正引领教育领域的认知变革,推动传统静态问答系统向具备动态知识整合与智能交互能力的数字导师演进。然而,现有通用大模型在人物科普教育场景中仍面临知识幻觉和教学策略针对性不足两大挑战。对此,提出基于意图增强型混合推理机制的人物科普教育认知智能体MECA(Master Education Cognitive LLM Agent,师大先生),基于感知层、推理层、行动层三层认知架构,构建“意图感知→知识推理→教育执行”的闭环决策范式。MECA引入动态认知增强机制,在感知层采用轻量级语言模型解析用户意图,在推理层设计意图增强型混合推理机制,融合领域知识、用户需求与教学策略进行多维推理,以提升知识生成的精准性与个性化适配能力。同时,为夯实数据基底,结合人工审查与大语言模型的语言理解能力,构建了国内首个面向人物科普教育领域的高质量问答数据集,涵盖上海地区具有重大贡献的知名学者、教育家、两院院士等著名人物的信息,包括生平事迹、学术理念、教育贡献等多个维度,填补了国内人物科普教育领域高质量语料的空白。实验结果表明,师大先生在多维度测评指标上均实现显著提升,增强了人物科普教育的认知交互能力与知识精准输出水平,为教育智能体的构建提供了可推广的范式,推动教育智能化向专业化、动态化方向发展。

关键词: Agent系统, 混合推理, 人物科普, 提示词工程, 意图分析

Abstract: LLM-driven agents have led a cognitive revolution in education,transforming traditional static Q&A systems into digital tutors with dynamic knowledge integration and intelligent interaction capabilities.However,existing general-purpose LLMs face significant challenges in the context of master popularization education,including knowledge hallucination,static knowledge representation limitations,and insufficient adaptability of teaching strategies.To address these challenges,this paper proposes the Master Education Cognitive LLM Agent(MECA),chatMaster,based on a three-layer cognitive architecture comprising a perception layer,reasoning layer and action layer.MECA establishes a closed-loop decision-making paradigm of “Intention Perception → Knowledge Reasoning → Educational Execution.”It introduces a dynamic cognitive enhancement mechanism for improving the precision of knowledge generation and personalized adaptation capabilities.In the proposed framework,the perception layer employs a lightweight language model to analyze user intent,while the reasoning layer incorporates an intent-enhanced hybrid reasoning mechanism that integrates domain knowledge,user needs,and teaching strategies for multi-dimensional reasoning.To build a solid data foundation,it develops the first high-quality Q&A dataset specifically for master education information in China,leveraging both human review and LLM-based language understanding.This dataset covers renowned scholars,educators and academicians in Shanghai,focusing on key figures with significant contributions.It encompasses multiple dimensions,including biographical details,academic ideologies,and educational contributions,addressing the lack of high-quality domain-specific corpora in this field.Experimental results demonstrate that chatMaster achieves significant improvements across multiple evaluation metrics,enhancing the cognitive interaction capabilities and precision of knowledge dissemination in master popularization education information.This research provides a generalizable paradigm for constructing educational agents,promoting the evolution of intelligent education towards greater specialization and dynamism.

Key words: Agent system, Hybrid reasoning, Master popularization education, Prompt engineering, Intent analysis

中图分类号: 

  • TP182
[1] LIU M,YANG M,WU Z M,et al.Development,Application Status and Future Prospect of Large Language Model Agents for Education[J].Modern Educational Technology,2024,34(11):5.
[2] PETRONI F,ROCKT ASCHEL T,RIEDEL S,et al.Lan-guage models as knowledge bases? [C]//Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing(EMNLP-IJCNLP).2019:2463-2473.
[3] KASNECI E,SEβLER K,KÜCHEMANN S,et al.ChatGPT for good? On opportunities and challenges of large language models for education[J].Learning and Individual Differences,2023,103:102274.
[4] FLORIDI L,CHIRIATTI M.GPT-3:Its nature,scope,limits,and consequences[J].Minds and Machines,2020,30(4):681-694.
[5] YANG A,YANG B,ZHANG B,et al.Qwen2.5 technical report[J].arXiv:2412.15115,2024.
[6] KRYĆCÍNSKI W,MCCANN B,XIONG C,et al.Evaluating the factual consistency of abstractive text summarization[C]//Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing(EMNLP).2020:9332-9346.
[7] WANG Y,WANG X,HUANG X,et al.Intent-aware recommendation via disentangled graph contrastive learning[C]//Proceedings of the 32th International Joint Conference on Artificial Intelligence.2023:2343-2351.
[8] RADFORD A,NARASIMHAN K,SALIMANS T,et al.Improving Language Understanding by Generative Pre-training[J/OL].https://s3-us-west-2.amazonaws.com/openai-assets/researchcovers/languageunderstandingpaper.pdf.
[9] LI Z Z,ZHANG D,ZHANG M L,et al.From system 1 to system 2:A survey of reasoning large language models[J].arXiv:2502.17419,2025.
[10] CHU Z,CHEN J,CHEN Q,et al.Navigate through enigmatic labyrinth a survey of chain of thought reasoning:Advances,frontiers and future[C]//Proceedings of the 2024 Conference of the Association for Computational Linguistics.2024.
[11] HU E J,SHEN Y,WALLIS P,et al.Lora:Low-rank adaptation of large language models[C]//Proceedings of the ICLR 2022 Conference.2022.
[12] HAYOU S,GHOSH N,YU B.Lora+:Efficient low rank adaptation of large models[J].arXiv:2402.12354,2024.
[13] BONAN M,HAYLEY R,ELIOR S,et al.Recent advances in natural language processing via large pre-trained language models:A survey[J].arXiv:2111.01243,2021.
[14] HUANG Z,XU W,YU K.Bidirectional LSTM-CRF models for sequence tagging[J].arXiv:1508.01991,2015.
[15] LIU Y,NI X,SUN J T,et al.Unsupervised transactional query classification based on webpage form understanding[C]//Proceedings of the 20th ACM International Conference on Information and Knowledge Management.2011:57-66.
[16] CAO W,WU Y,SUN Y,et al.A review on multimodal zero-shot learning[J].Wiley Interdisciplinary Reviews:Data Mining and Knowledge Discovery,2023,13(2):e1488.
[17] AMATRIAIN X.Prompt design and engineering:Introduction and advanced methods[J].arXiv:2401.14423,2024.
[18] LEVY O,SEO M,CHOI E,et al.Zero-shot relation extraction via reading comprehension[C]//Proceedings of the 2017 Conference on Computational Natural Language Learning.2017.
[19] WEI X,CUI X Y,CHENG N,et al.Zero-shot information extraction via chatting with chatgpt[J].arXiv:2302.10205,2023.
[20] WU X,TSIOUTSIOULIKLIS K.Thinking with KnowledgeGraphs:Enhancing LLM Reasoning Through Structured Data[J].arXiv:2412.10654,2024.
[21] LEWIS P,PEREZ E,PIKTUS A,et al.Retrieval-augmentedgeneration for knowledge-intensive NLP tasks[C]//Procee-dings of the 2020 Conference in Neural Information Processing Systems.2020:9459-9474.
[22] MODRAN H A,BOGDAN I C,URSUIU D,et al.LLM intelligent agent tutoring in higher education courses using a RAG approach[C]//International Conference on Interactive Collaborative Learning.Cham:Springer Nature Switzerland,2024:589-599.
[23] QURESHI N A,LIASKOS S,PERINI A.Reasoning aboutadaptive requirements for self-adaptive systems at runtime[C]//Proceedings of the 2011 2nd International Workshop on Requirements@ Run.Time.IEEE,2011:16-22.
[24] ASIF M,KHAN T A,SONG W C.Leveraging Cognitive Machine Reasoning and NLP for Automated Intent-Based Networking and e2e Service Orchestration[J].IEEE Access,2025.
[25] TANG L M,YU R N,DONG Q W,et al.A review of non-intrusive sensing based personalized resource recommendations for help-seekers in education[J].Journal of East China Normal University.Natural Sc,2018(5):17-29.
[26] ZHANG J Y,WANG T K,YAO C Y,et al.Construction and Evaluation of Intelligent Question Answering System for Electric Power Knowledge Base Based on Large Language Model [J].Computer Science,2024,51(12):286-292.
[27] LIU S Y,WANG L H,LIU W,et al.Research Paradigm and Platform Construction of Digital Humanities [J].Documentation,Information & Knowledge,2022,39(1):6-29.
[28] HEY T,TANSLEY S,TOLLE K.The Fourth Paradigm:Data-intensive Scientific Discovery [M].Beijing:Science Press,2012.
[29] YANG H,LIU X Y,DAN W C.FinGPT:Open-Source Financial Large Language Models[C]//Proceedings of the 2023 Confe-rence of IJCAI.2023.
[30] CUI J,LI Z,YAN Y,et al.Chatlaw:Open-source legal large language model with integrated external knowledge bases[J].ar-Xiv:2306.16092,2023.
[31] RADFORD A,WU J,CHILD R,et al.Language models are unsupervised multitask learners[J].OpenAI blog,2019,1(8):9.
[32] SIVARAJKUMAR S,WANG Y.HealthPrompt:a zero-shotlearning paradigm for clinical natural language processing[C]//Proceedings of the 2023 Conference on AMIA Annual Sympo-sium Proceedings.2023,2022:972.
[33] BROWN T,MANN B,RYDER N,et al.Language models are few-shot learners[J].Advances in Neural Information Proces-sing Systems,2020,33:1877-1901.
[34] SHAO M,BASIT A,KARRI R,et al.Survey of different large language model architectures:Trends,benchmarks,and challenges[J].IEEE Access,2024,12:188664-188706.
[35] LING H,HONG D C,ZHANG C H.Research on tacit know-ledge integration:a synthesis of social ties and TMS[J].Know-ledge Management Research & Practice,2011,9(3):256-262.
[36] CHINCHOR N.MUC-6 named entity task definition(version2.1) [C]//Proceedings of the 6th Conference on Message Understanding.Columbia,Maryland,1995.
[37] SERBAN I,SORDONI A,BENGIO Y,et al.Bleu:a Method for Automatic Evaluation of Machine Translation[C]//Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics.Philadelphia,Pennsylvania,USA:Association for Computational Linguistics,2002:311-318.
[38] LIN C Y.ROUGE:A Package for Automatic Evaluation ofSummaries[C]//Proceedings of the Association for Computational Linguistics.2004:74-81.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
地址：重庆市两江新区洪湖西路18号    邮编：401121  
电话：023-63500828（编辑部） 023-67039612（会议/专题） 023-67039623（财务/发票）
E-mail：jsjkx12@163.com
本系统由北京玛格泰克科技发展有限公司设计开发