中文体育类核心期刊

中国人文社会科学期刊AMI综合评价(A刊)核心期刊

《中文社会科学引文索引》(CSSCI)来源期刊

美国《剑桥科学文摘》(CSA)收录期刊

中国高校百佳科技期刊

逆向工程:具身智能与具身的人运动预测的神经哲学阐释

Reverse Engineering: A Neurophilosophical Interpretation of Embodied AI and Embodied Athletes' Action Prediction

  • 摘要: 整合神经哲学、认知科学与工程实践,以逆向工程为方法论路径,即以具身智能的算法实践为可解剖的工程对象,逆向推导人类体育运动预测这一“黑箱”的内在机制,聚焦脑-身-环境耦合,旨在揭示体育运动的具身本质——主动预测。具身智能的兴起标志着从预编程人形机器人向类身范式的转捩,强调运动预测作为处理不确定性的核心问题,通过主动推理与自由能最小化实现感知-行动闭环,将智能嵌入身体与环境的动态互动中,超越了传统符号主义与联结主义的局限。借助具身智能的算法机制,逆向探究贝叶斯脑理论,阐释具身的人类运动者预测策略的主动性、精度加权及自我证明特征。进一步剖析具身智能与人类运动者在预测中的马尔可夫性同构,逆推人类延展心智的数学特征,突出身体先验性、时间主观调节与形态计算的生态优势。尽管当前具身智能泛化能力不足,无法实现人类运动者的具身预测,但能够为体育的身智科学研究奠定基本逻辑基础,进而为体育学从运动本身出发探问运动机制提供可量化解释的逆向工程路径,推动体育科学的研究范式从脑神经中心论向脑-身-环境整体论转变。

     

    Abstract: Integrating neurophilosophy, cognitive science and engineering practice, this article deploys reverse engineering as its methodological pathway — taking the algorithmic praxis of embodied AI as a dissectable engineering object from which to reversely infer the internal mechanisms of action prediction in human sports, a domain that otherwise remains a black box — and focuses on brain-body-environment coupling, with the aim of disclosing the essence of embodied sports: active prediction. The study argues that the emergence of embodied AI marks an inflection from pre-programmed humanoid robots toward body-analogical paradigm, foregrounding action prediction as the central problem of uncertainty management: through active inference and free energy minimization, intelligence is instantiated within the dynamic interplay of body and environment as a perception-action loop, thereby surpassing the constraints of classical symbolism and connectionism. By leveraging the algorithmic mechanisms of embodied intelligence, this research reversely explores Bayesian brain theory to explicate the active, precision-weighting, and self-evidencing characteristics of predictive strategies of humans in sports. It further analyzes the Markovian isomorphism between embodied intelligence and humans in sports in predictive processes, reversely inferring the mathematical signatures of the human extended mind while accentuating the ecological advantages of bodily priors, subjective temporal modulation, and morphological computation. Although embodied intelligence presently exhibits insufficient generalization capacity to replicate the embodied prediction of humans in sports, it nonetheless furnishes sports science — conceived as a discipline of body-mind intelligence — with foundational logical scaffolding, and in so doing provides a quantitatively interpretable reverse-engineering pathway for sports science to interrogate the mechanism in sport from the sport itself, pushing a paradigmatic shift in sports science from neurocentric frameworks toward brain-body-environment holism.

     

/

返回文章
返回