研究生计划 | Active Inference Institute

在本页

部分

研究人员开发与研究所和生态系统相一致的研究、应用或公共资源，定期提交进度报告，并产生推动该领域进步且公开可访问的成果。

同学会成员做什么

研究员在定义的项目或关注领域上进行工作，如研究、软件开发、教育材料、本体论、应用或生态系统支持，并在定期间隔报告进展。报告是公开的：研究员通过社区通讯、项目会议或公开帖子分享更新。这些输出积累成一整套作品，供整个积极推断社区及其外界使用。

Fellows pathwayParticipate Weekly updatesCommunity

一个团队需要什么？

Fellowship 需要有意义的时间投入，并且有一个明确的范围。 Fellows 被期望在他们定义的工作上取得进展，定期沟通这些进展，并交付对他人有用的结果。具体要求——报告频率、交付物 —— 在 fellowship 开始时由 Fellow 和 Institute 确定。 Research Fellow 角色是一个无薪、非雇员职位，默认任期为两年，可能可以续期。

当前研究 fellows

The Institute's current Research Fellows, their appointment start, project focus, ORCID iD, and their own description of the research are listed in the table below. The roster is generated from the Institute's fellows registry, so it reflects the current appointments rather than a snapshot fixed in prose.

谁应该申请

同心协力计划”适合那些已经通过志愿服务、实习或独立研究与研究所有过接触，并准备承担更大、更持续责任的人。预计申请人需要具备积极推断的相关经验——这些同心成员可以在定义的问题上高度自主地工作。如果你是研究所的新手，通过志愿服务或实习开始是推荐的路径。

Volunteer first Internship

申请要求

完整的研究 fellow 申请包包括申请人已阅读并同意 fellows 计划条款的确认；一份一到六页的研究提案（不包括引用）包含摘要、研究问题和目标、意义、方法论、与研究所使命的一致性、预期成果和时间表；当前简历；一封至三封推荐信；以及最多五份代表性的出版物或产品。已完成的包和问题发送到 blanket@activeinference.institute，并在主题行中包含 [RESEARCH FELLOWS]。

输出和公开记录

研究奖学金的输出旨在对公众有用。论文、软件仓库、教育课程材料、本体贡献、数据集和项目文档均被视为奖学金输出。该研究所支持研究员使其工作在更广泛社区中可查找、可引用的和可重用的。已发表的输出是研究所研究和教育工作的永久公开记录的一部分。

GitHub organizationProjects

如何成为同仁

同进开始于关于你拟议工作及其与研究所和生态系统目标对齐的对话。同进路径页面描述了表达兴趣并提出同进范围的当前流程。大多数同进项目源自现有参与——积极贡献者希望正式化和加深他们的参与。

Fellows pathwayParticipate

Fellowship 条款

研究员职位无薪且非员工，受一份签署的聘用信和一系列项目条款所规范。 fellows定期（例如每季度）提交进展报告，描述他们的研究活动、成就和挑战，并参与年度评估。研究院不主张对fellows私人工作的任何访问权，且fellow在研究院项目的参与不影响fellow或研究院的知识产权权利。fellow被鼓励但非强制要求将‘研究员，Active Inference Institute’列为相关输出中的专业联系人。

最佳下一步行动

Where to go next

Fellows pathwayParticipate 过滤资源全球目录程序实习

当前研究 fellows

9 current Research Fellows

Each fellow works on a defined scope of research and reports on it publicly. Rows come from the Institute's Research Fellows roster.

Current Research Fellows at the Active Inference Institute.
Fellow	Position	Fellow since	Focus	ORCID	研究
Alexander Hemming	Research Fellow	3/2026	Foundations of Ideology	—	Cross-disciplinary framework rooted in political neuroscience offering a normative account of how ideologies form, update, and interact, grounded in the Free Energy Principle.
Jake Hooper	Research Fellow	3/2026	—	—	—
Mahault Albarracin	Research Fellow	3/2026	AI2C: Active Inference Alignment Coalition	0000-0003-0916-4645	Proposal for the Active Inference Alignment Coalition, a multi-institution consortium producing large-scale proposals for AI alignment grounded in the Free Energy Principle and Active Inference, with a standardized governance framework, reusable proposal modules, and shared infrastructure for evidence synthesis, benchmarking, and stakeholder engagement.
Hongju Pae	Research Fellow	11/2025	Perspective Architectures for Coherent and Attuned Artificial Agency	0000-0002-5174-8858	Develops computational and theoretical frameworks for modeling how artificial agents form perspective, maintain coherent affect, and achieve developmental alignment, integrating Active Inference with phenomenological insights. Builds simulation prototypes toward testable architectures of sentience for safe, ecologically grounded machine minds.
Sheila Macrine	Research Fellow	11/2025	Multi-dimensional taxonomy of agency	0000-0002-8600-0938	Extends the framework from the forthcoming book 'Embodied Intelligence: Multidisciplinary Perspectives on Natural, Artificial, and Hybrid Systems' (MIT Press, 2026). Proposes a multi-dimensional taxonomy of agency formalizing nine qualitative dimensions — intentionality, autonomy, reflexivity, and others — as tunable Active Inference parameters, distinguishing agency (the driver) from intelligence (the engine).
Shagor Rahman	Research Fellow	8/2025	Myth of Objectivity	0009-0004-0460-0078	Investigates how morality and symbolic thought co-evolved (the Myth of Objectivity Hypothesis): the capacity to model shared cultural expectations enabled humans to transcend individual perspectives. Uses multi-agent active inference simulations and transcendental model selection, and offers a perspective on AI and AGI alignment.
Robert Worden	Research Fellow	10/2024	Wave Hypothesis	0000-0001-7304-2752	Two main research interests: 3-D spatial cognition, proposing wave-based rather than pure neural-firing models of consciousness in the thalamus and insect central body; and language, covering its evolution, learning, and brain mechanisms.
Jean-François Cloutier	Research Fellow	5/2024	Symbolic Cognitive Robotics	0009-0001-1841-2279	I seek to find out what it takes, at a minimum, for a robot to learn, on its own, how to survive in a world it knows initially almost nothing about. Continuation of a project of many years programming simple autonomous robots to develop and ground understanding of cognition. Drawn into Active Inference, Kantian epistemology, map vs territory, biosemiotics, mortal computing, collective intelligence, autopoiesis and constraint closure.
John Boik	Research Fellow	5/2024	Cognitive Narrative Ecosystem (CogNarr)	0000-0003-1289-7997	Continuation of work described in a book and two series of concept papers exploring science-driven de novo development of new cognitive architectures fit for purpose. First series: new societal (economic and governance) systems as components of a society's cognitive architecture. Second series: an online ecosystem facilitating cognition in the large-group setting.