首页|情感计算的意向困境

情感计算的意向困境

扫码查看
作为言语行为的人机情感对话中包含语义、语用及心灵三重意向,但现行情感计算领域所有基于向量语义学的神经网络模型,包括chatGPT的处理方式,仍只是形式层面,而不是真正的语义计算,其既未能实现人类基于自然语言理解基础之上的意义意向,也未能实现心灵的赋义意向.要想实现真正的意义理解,情感计算应从根本上转变哲学预设,尝试一种自顶向下的解决思路,从身体和世界出发追溯意义的来源,设计一种能实现意向弧的具身智能,以首先解决心灵的意向问题,进而解决由其派生的语言意向问题.
The Intentionality Dilemma of Affective Computing
This paper analyzes the semantic and pragmatic dilemmas of a tripartite intentionality in human-computer dialogues for affective communication with a connectionist approach based on vector semantics.The neural networks currently adopted for affective computing,including the chat GPT model,are still formal rather than semantic processing for their failure in understanding either the intentionality of human language or of the human mind.To achieve a true semantic understanding,a transformation of philosophical assumption with a top-down design which traces the source of meaning from the body and the world for an embodied intelligence should be made.

affective computingspeech actintentionalitysemanticstop-down

彭静

展开 >

安徽大学哲学学院,安徽 合肥 230039

情感计算 言语行为 意向性 语义 自顶向下

2024

汕头大学学报(人文社会科学版)
汕头大学

汕头大学学报(人文社会科学版)

CHSSCD
影响因子:0.291
ISSN:1001-4225
年,卷(期):2024.40(3)