华东师范大学学报(哲学社会科学版) ›› 2020, Vol. 52 ›› Issue (5): 90-96.doi: 10.16382/j.cnki.1000-5579.2020.05.009

• 人工智能与社会 • 上一篇    下一篇

人工群体智能是否可能?

何静   

  1. 华东师范大学哲学系(上海, 200241)
  • 出版日期:2020-09-15 发布日期:2020-09-19
  • 作者简介:何静,华东师范大学哲学系副教授(上海,200241)

How Is Swarm Intelligence Possible?

HE Jing   

  • Online:2020-09-15 Published:2020-09-19

摘要: 人工群体智能技术以自治体的方法论与神经网络的模型为基础,通过大量独立人工智能体间的简单交互形成超越个体维度的群体智能类型。一方面,这种人工智能的新发展向我们展现了在某些领域超越人类个体智能水平的新型智能形式;但另一方面也凸显了人工群体智能与人类群体智能在交互方式上的深层次差异。对于人类群体而言,交互从本质上来说是一个个体间共同—规定的耦合过程。这种共同参与、相互规定的社会交互过程在维系并塑造个体自治性的同时,形成了"复数"的"我"对世界的理解——群体智能。

关键词: 群体人工智能, 信息, 交互, 相互规定

Abstract: Based on the latest autonomous methodology and neural network model, swarm intelligence is a type of collective artificial intelligence beyond the individual artificial agent. On the one hand, this new development of artificial intelligence displays the new types of intelligence that may overpower individual human intelligence in some specific domains; on the other hand, it discloses some fundamental differences between swarm intelligence and human collective intelligence. For human beings, interaction is essentially a mutual regulative process among individuals. This mutual regulative social interaction sustains and shapes the individual autonomy, and at the same time constitutes "our"(plural "my") understanding of the world, that is, collective intelligence.

Key words: swarm intelligence, information, interaction, mutual regulation