The Human Cognitive Simulation Framework proposes a governed cognitive AI architecture designed to improve personalization, adaptability, and long-term coherence in human AI interaction. The framework integrates short-term memory (conversation context), long-term memory (interaction context), cognitive processing modules, and managed knowledge persistence into a unified architectural model that ensures contextual continuity across sessions and controlled accumulation of relevant information. A central contribution is a unified memory architecture supervised by explicit governance mechanisms, including algorithmic relevance validation, selective persistence, and auditability. The framework incorporates differentiated processing modules for logical, creative, and analogical reasoning, enabling both structured task execution and complex contextual inference. Through dynamic and selective knowledge updating, the system augments the capabilities of large language models without modifying their internal parameters, relying instead on retrieval augmented generation and governed external memory. The proposed architecture addresses key challenges related to scalability, bias mitigation, and ethical compliance by embedding operational safeguards directly into the cognitive loop. These mechanisms establish a foundation for future work on continuous learning, sustainability, and multimodal cognitive interaction. This manuscript is a substantially revised and extended version of the previously released preprint (DOI:10.48550/arXiv.2502.04259).
Crisp: Cognitive Restructuring of Negative Thoughts through Multi-turn Supportive Dialogues
The CoAI Group, DCST, Tsinghua University(清华大学自动化研究所);University of Pennsylvania(宾夕法尼亚大学);Lingxin AI(灵犀AI);Harvard Graduate School of Education, Harvard University(哈佛大学教育研究生院);Department of Psychology and Behavioral Sciences, Zhejiang University(浙江大学心理学与行为科学系);Fuxi AI Lab(伏羲AI实验室)
The Architecture of Cognitive Amplification: Enhanced Cognitive Scaffolding as a Resolution to the Comfort-Growth Paradox in Human-AI Cognitive Integration
Catholic University of Sacred Heart(圣心天主教大学);Istituto Auxologico Italiano(意大利奥索洛吉科研究所)
Between Sense and Sensibility: Declarative narrativisation of mental models as a basis and benchmark for visuo-spatial cognition and computation focussed collaborative cognitive systems
University of Bremen(不来梅大学)
“\emph{sensing}”和“\emph{sensibility}”之间有什么区别?换句话说,什么样的认知过程介导了感知能力和可感知印象的形成-例如,抽象,类比,假设和理论的形成,信念和他们的修正,论点的形成-在特定领域的问题解决中,还是在日常生活,工作和简单地在环境中走动的常规活动中?如何才能知识和推理的能力,表现出人类在特定的问题背景下,被用来作为一个模型和基准的发展与人类的援助,保证和授权的协作认知(互动)系统?我们提出这些问题的背景下,一系列的辅助技术与\emph{visuo-spatial perception and cognition}任务,包括方面,如常识,创造力,专业领域的知识和解决问题的思维过程的应用。正在考虑的辅助技术包括:(a)人类活动解释;(b)高级认知旋转;(c)建筑和数字媒体创作等领域以人为本的创造性设计;(d)地理信息系统的定性分析。计算叙事不仅提供了丰富的认知基础,但他们也作为一个基准的功能性能在我们的发展计算认知辅助系统。我们认为,有关空间,行动和变化的计算叙事提供了一个有用的模型\emph{visual}和\emph{spatio-temporal thinking}在广泛的解决问题的任务和应用领域,协作认知系统可以提供一个辅助和授权功能。
Building Machines that Learn and Think with People
University of Cambridge(剑桥大学);Princeton University(普林斯顿大学);NYU(纽约大学);The Alan Turing Institute(艾伦·图灵研究所);MIT(麻省理工学院);Microsoft Research(微软研究院);University of Chicago(芝加哥大学)