[1]李春文 陆思聪 吴热冰 丁青青 刘华平 李东海等.从人工智能学科发展到人机会话关键问题的探析与展望[J].清华大学教育研究,2022,(03):25-32.
　LI Chun-wen LU Si-cong WU Re-bing DING Qing-qing LIU Hua-ping LI Dong-hai.Analysis and Prospect from Development of Artificial Intelligence to Key Issues of Human-machine Conversation[J].TSINGHUA JOURNAL OF EDUCATION,2022,(03):25-32.

点击复制

从人工智能学科发展到人机会话关键问题的探析与展望()

分享到：

清华大学教育研究[ISSN:1001-4519/CN:11-1610/G4]

卷:

期数:

2022年03期

页码:

25-32

栏目:

人工智能专题

出版日期:

2022-06-20

文章信息/Info

Title:

Analysis and Prospect from Development of Artificial Intelligence to Key Issues of Human-machine Conversation

作者:

李春文 ¹ 陆思聪¹ 吴热冰¹ 丁青青² 刘华平³ 李东海⁴等

1.清华大学 自动化系；2.清华大学 电机工程与应用电子技术系；3.清华大学 计算机科学与技术系；4.清华大学 能源与动力工程系

Author(s):

LI Chun-wen¹ LU Si-cong¹ WU Re-bing¹ DING Qing-qing² LIU Hua-ping³ LI Dong-hai⁴

1.Department of Automation, Tsinghua University; 2.Department of Electrical Engineering, Tsinghua University; 3.Department of Computer Science and Technology, Tsinghua University; 4.Department of Energy and Power Engineering, Tsinghua University

关键词:

人工智能; 自然语言处理; 控制论; 人机会话; 深度学习; 情感计算; 人工智能与教育

Keywords:

artificial intelligence;  natural language processing;  cybernetics;  human-machine conversation;  deep learning;  affective computing;  AI education

分类号:

G434

文献标志码:

A

摘要:

沿着人工智能从起源到应用的演化脉络，本文分析研究了其学科发展路径上的一些关键问题。首先，对人工智能领域的产生到专家系统的这一发展阶段进行了历史回顾和定位分析，概括了当前人工智能在理论方法和应用展开方面的基本领域分支及其分布状态。进一步，从拟人与超人这一话题切入，探讨了人机会话中智能的判别标准及对未来发展的预期。然后，分析了当前深度学习方法的自主与内蕴特性，并从不同角度分析了自然语言处理领域中智能会话面临的困难，探索其以人格与情感为代表的产生根源和可能的解决方向，并特别论述了普适人工智能与教育发展的特殊联系。

Abstract:

From the origin to the application of artificial intelligence, some key problems along with the development of this discipline are analyzed.Firstly, a historical review and evaluation are carried out to the development stage from the birth of the artificial intelligence field to the expert system, following which we summarize the fundamental branches and their distribution in the theory and application of artificial intelligence.Further, induced from the topic of anthropomorphic and superman, the discriminative criteria of intelligence in human-machine conversation and the expectation of future development are discussed.Then the autonomous and intrinsic characteristics of the current deep learning methods are analyzed.Furthermore, from different angles, the difficulties faced by intelligent conversations in the field of natural language processing are analyzed, and the causes(typically by the personality and emotion)and possible solutions are discussed.Then the special relationship between general artificial intelligence and the development of education is mentioned.

常用功能