Scholarlink

<p>This paper explores the convergence of <i>Agentive Cognitive Construction Grammar</i> (AgCCxG) and <i>neuro-symbolic AI</i> (NSAI) for modeling human cognition and language processing. AgCCxG conceptualizes language as an embodied, predictive, and semiotic system operating through a Markov Blanket, structuring cognition via differentiation, optimization, and predictive control. NSAI integrates neural networks’ pattern recognition with symbolic AI’s reasoning capabilities, mirroring dual-system models of human cognition. I argue that AgCCxG provides a neurobiologically plausible foundation for enhancing NSAI’s predictive modeling, enabling AI to progress from statistical correlation toward meaning-driven computation. By incorporating semiotic agency, embodied inference, and context-aware reasoning, this integration advances explainable AI, scientific discovery, and personalized education. The synergy addresses critical challenges including the hallucination problem, with symbolic reasoning serving as a corrective mechanism for neural outputs. The future of artificial intelligence requires principled integration of predictive processing, semiotic agency, and embodied cognition—principles that have shaped human language and thought for millennia. This represents a significant step toward bridging human and machine intelligence in more theoretically sound and ethically responsible ways.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Bridging embodied cognition and AI: Agentive Cognitive Construction Grammar as a backing theory for neuro-symbolic AI

  • 期刊论文

  • 发表日期：

    2025年6月26日

查看全文
  • Sergio Torres-Martínez

摘要

This paper explores the convergence of Agentive Cognitive Construction Grammar (AgCCxG) and neuro-symbolic AI (NSAI) for modeling human cognition and language processing. AgCCxG conceptualizes language as an embodied, predictive, and semiotic system operating through a Markov Blanket, structuring cognition via differentiation, optimization, and predictive control. NSAI integrates neural networks’ pattern recognition with symbolic AI’s reasoning capabilities, mirroring dual-system models of human cognition. I argue that AgCCxG provides a neurobiologically plausible foundation for enhancing NSAI’s predictive modeling, enabling AI to progress from statistical correlation toward meaning-driven computation. By incorporating semiotic agency, embodied inference, and context-aware reasoning, this integration advances explainable AI, scientific discovery, and personalized education. The synergy addresses critical challenges including the hallucination problem, with symbolic reasoning serving as a corrective mechanism for neural outputs. The future of artificial intelligence requires principled integration of predictive processing, semiotic agency, and embodied cognition—principles that have shaped human language and thought for millennia. This represents a significant step toward bridging human and machine intelligence in more theoretically sound and ethically responsible ways.