自然杂志 ›› 2024, Vol. 46 ›› Issue (2): 150-156.doi: 10.3969/j.issn.0253-9608.2024.01.013

• 科技进展 • 上一篇    下一篇

面向机器智能的情感分析

彭俊杰   

  1. 上海大学 计算机学院,上海 200444
  • 收稿日期:2023-12-07 出版日期:2024-04-25 发布日期:2024-04-19
  • 基金资助:
    2022年上海市服务业发展引导资金项目

Emotion analysis for machine intelligence

PENG Junjie   

  1. School of Computer Engineering and Science, Shanghai University, Shanghai 200444, China
  • Received:2023-12-07 Online:2024-04-25 Published:2024-04-19

摘要: 大模型的兴起,尤其是以大模型为基础的人工智能技术的发展与应用引起了全社会对人工智能技术的广泛关注。人工智能技术已经在很多方面达到甚至超越人类,可以替代人工完成一些技艺要求较高的工作。不过,在人工智能迈向人类智慧的道路上,仍然有许多难题亟待克服与解决,其中一个重要的问题就是情感分析与理解问题,如让机器准确理解人类的情感。针对该问题,文章从情感的定义与分类入手,对情感分析研究需要解决的问题,研究方向、研究现状、面临的挑战、应用前景与展望等进行了讨论。

关键词: 机器智能, 情感分析, 多模态, 信息融合, 特征提取

Abstract: The rise of large language models, especially the development and application of artificial intelligence technology based on large language models, has attracted widespread attentions from the whole society to artificial intelligence technology, as it can reach or even surpass humans in many aspects, and accomplish tasks with high skill requirements that can only be done by humans before. However, there are still many challenges that need to be overcome and solved on the path of artificial intelligence towards human intelligence. One of the important issues is emotional analysis and understanding, that is, how to accurately understand human emotions with artificial intelligence technology. In response to this issue, the paper presents the definition and classification of emotions, and discusses the problems to be solved, research directions, current research status, challenges, applications, and prospects of the study on emotion analysis.