IEEE CogMI 2026 Call for Papers

Cognitive machine intelligence (CogMI) takes up one of the central questions in modern AI research: what does it mean for an artificial system to perceive, reason, remember, decide, and communicate, and what can the cognitive, neural, and behavioral sciences tell us about how to build, study, and deploy such systems well? As AI becomes infrastructure for scientific discovery, healthcare, and the coordination of complex social and economic systems, the answers to these open questions have practical consequences. Making progress on them calls for work that is both technically rigorous and interdisciplinary.

The goal of the IEEE Conference on Cognitive Machine Intelligence (IEEE CogMI) is to create a forum for research that places cognition at the center of how we build, study, and deploy AI and machine learning. We welcome contributions that draw on theories of human cognition to inform the design of AI, that analyze cognitive-like phenomena in artificial systems, and that build AI capable of richer cognitive behavior. The conference is deliberately interdisciplinary, bringing together researchers and practitioners from computer science and engineering, psychology and neuroscience, the social and behavioral sciences, philosophy, and ethics, legal studies, and public policy. We are especially interested in work that treats AI as a partner to human judgment rather than a substitute for it, and that examines the conditions under which human–machine collaboration produces reliable and accountable systems.

---

List of Topics

Topics of interest include, but are not limited to:

Cognitive Foundations and Architectures

• Cognitive architectures, computational models of mind, and dual-process (System 1/System 2) theories in AI, including agent architectures, models of executive function, adaptive behavior, bounded rationality, and developmental learning trajectories
• Neuroscience-inspired AI (NeuroAI) and neuromorphic computing, including brain-inspired architectures, biologically plausible learning rules, neural coding principles, spiking neural networks, predictive coding, and free-energy frameworks
• Memory, attention, and information processing in artificial systems, including memory-augmented networks, episodic replay, models of forgetting and consolidation, selective attention, cognitive load, information bottlenecks, chunking, and content-addressable memory

Reasoning and Knowledge

• Commonsense, causal, and abstract reasoning, including analogical and relational reasoning, compositional generalization, spatial and temporal reasoning, counterfactual thinking, and reasoning under uncertainty and incomplete information
• Neuro-symbolic AI and hybrid architectures, including integrating neural learning with symbolic knowledge representation, differentiable logic programming, neural theorem proving, and bridging perception, language, and formal reasoning
• Knowledge representation, ontology, and conceptual structure, including concept formation and categorization, belief revision, epistemic reasoning, knowledge graphs, commonsense knowledge bases, and schema-based understanding of events
• Probabilistic and Bayesian models of cognition, including probabilistic computing, Bayesian learning, rational analysis, reasoning under uncertainty, and probabilistic programming as cognitive modeling frameworks

Language, Communication, and Meaning

• Large language and vision-language models as cognitive and linguistic objects of study, including what LLMs and VLLMs reveal and fail to reveal about human cognition, language, vision, and meaning; multimodal cognition and visual-linguistic binding; scaling laws and intelligence; in-context learning as cognitive flexibility; chain-of-thought as verbal reasoning
• Natural language processing, generation, and speech through a cognitive lens, including computational psycholinguistics, models of language acquisition, semantic grounding and compositionality, pragmatics, discourse, text analytics, and speech recognition/generation
• Dialogue systems, virtual agents, chatbots, and conversational AI, including communication and common ground theory, cooperative dialogue, audience design, perspective-taking, and interactive storytelling

Perception, Embodiment, and Action

• Perception, grounding, and embodied cognition, including multimodal integration, language grounding to vision and robotics, affordance learning, intuitive physics, object permanence, active perception, and simulation-based understanding
• Computer vision and image processing as cognitive perception, including visual reasoning, scene understanding, visual question answering, saliency, gaze modeling, and cognitive models of visual search and recognition
• Cognitive robotics, autonomous systems, and human-robot interaction, including sensorimotor learning, cognitive motor planning, spatial navigation, embodied interaction, situated cognition in physical agents, and collaborative robot behavior

Learning and Adaptation

• Machine learning, neural networks, and deep learning as models of cognition, including few-shot and meta-learning as rapid generalization, curriculum and continual learning, transfer and domain adaptation, inductive biases as cognitive constraints, and self-supervised learning as perceptual development
• Reinforcement learning, planning, and goal-directed behavior, including model-based vs. model-free learning, hierarchical goal decomposition, world models, imagination and look-ahead, curiosity-driven exploration, and test-time adaptive reasoning
• Learning with relational and graph-structured data, including graph neural networks for relational and structural cognition, knowledge organization, relational reasoning, and learning over interconnected conceptual structures

Social Cognition and Behavior

• Theory of mind, social reasoning, and multi-agent cognition, including modeling beliefs, desires, and intentions of others; joint attention; cooperation and coordination; cultural transmission; negotiation; mechanism design; and algorithmic game theory
• Social computing, computational social science, and social psychology of AI, including modeling collective behavior and social dynamics, group decision-making, social norms, social influence, opinion formation, and AI-mediated social interaction
• Affective computing, emotion, and behavioral science, including emotion recognition and modeling, sentiment analysis, empathy in AI, behavioral models of decision-making, heuristics and biases, and the cognitive-behavioral foundations of human-AI systems

Creativity and Imagination

• Computational creativity, imagination, and generative cognition, including AI for art, music, design, and scientific ideation; divergent thinking; conceptual blending; generative models as accounts of imagination; evaluating novelty and meaning; and ethical issues of creative AI (authorship, appropriation)

Human-AI Interaction and Trust

• Human-AI collaboration and mixed-initiative systems, including augmented cognition, human-in-the-loop learning, cognitive aspects of shared decision-making, adaptive interfaces, designing for appropriate mental models of AI, and cognitive workload in human-AI teams
• Trust, reliance, and mental models of intelligent machines, including cognitive foundations of trust calibration, over-reliance and under-reliance, transparency as a trust mechanism, and building/maintaining trust in AI across contexts

Explainability, Safety, and Alignment

• Explainability, interpretability, and transparency, including cognitive models of explanation, XAI methods (feature attribution, concept-based, natural language), probing neural representations, attention visualization, argumentative explanations, and human evaluation of explanations
• AI safety, value alignment, and responsible AI, including bias, fairness, and equity through cognitive and social science lenses; reward modeling and preference specification; adversarial robustness; moral reasoning; accountability; and the cognitive science of AI risk perception
• Privacy, security, and adversarial machine learning, including privacy-preserving AI, adversarial resilience as cognitive reliability, deceptive alignment, and the societal-cognitive impacts of surveillance and data practices

Evaluation, Philosophy, and Societal Impact

• Cognitive benchmarking and psychometric evaluation of AI, including testing AI on human cognitive tasks, comparing human and machine reasoning/perception/language, behavioral experiments with AI, compositionality and systematicity metrics, and evaluation methodology for cognitive AI
• Consciousness, metacognition, and self-awareness in AI, including introspection, confidence calibration, self-monitoring and self-correction, computational theories of consciousness, self-models, and philosophical questions about machine understanding
• Philosophy of AI, ethics, and the nature of intelligence, including the grounding problem, narrow vs. general intelligence, moral status of cognitive AI, governance and regulation, and the distinction between understanding and pattern matching
• Societal and cognitive impacts of AI, including effects on human attention, memory, skill, and autonomy; AI in education, health, science, business, and public welfare; cognitive ergonomics of AI-generated content; and responsible deployment across application domains

Cross-Cutting and Emerging

• Emerging frontiers in cognitive AI, including agentic AI and autonomous tool use, foundation models as cognitive models, retrieval-augmented cognition, multi-agent cognitive ecosystems, AI for scientific discovery, developmental AI, world models as internal simulators, synthetic data as cognitive rehearsal, and AI-assisted cognitive science

Important dates:

Submission Guidelines

• Research Track Paper Submission

  We invite original research papers that have not been previously published and are not currently under review for publication elsewhere. Papers submitted to Research track should be up to 10 pages, anonymously in the standard two column IEEE proceedings format, excluding the bibliography, well-marked appendices, and supplementary material. which can be found at IEEE Manuscript Templates for Conference Proceedings . Authors should not change the font or the margins of the IEEE format. Papers should avoid revealing authors’ identity in the text. When referring to their previous work, authors are required to cite their papers in the third person, without identifying themselves. The papers should be submitted at the research track of the conference in EasyChair.

• Vision Track Paper Submission (By Invitation Only)

  We invite original research vision papers that have not been previously published and are not currently under review for publication elsewhere. Vision contributions should focus on blue-sky ideas and research vision in the area that at least one of the lead senior authors are known for. Vision paper can be as short as 2 pages but should be no longer than 10 pages in the standard two column IEEE proceedings format, which can be found at IEEE Manuscript Templates for Conference Proceedings . The papers should be submitted at the vision track of the conference in EasyChair.

• Industry/Government Paper Submission

  We invite original industry papers that have not been previously published and are not currently under review for publication elsewhere. At least one co-author must be affiliated with industry or Government organizations, such as labs in DoE, DoD, NIH, NSA Labs. Papers submitted to Regular or Industry/Gov track should be no longer than 10 pages in the standard two column IEEE proceedings format, IEEE Manuscript Templates for Conference Proceedings . The papers should be submitted at the industry/gov track of the conference in EasyChair.

Submission Rounds: There are two submission rounds to IEEE CogMI 2026. Authors who submit in the first round may revise and resubmit their papers in the second round if their papers are not accepted initially. It is also possible to submit directly to Round 2; however, in that case, there won't be an opportunity for resubmission.

---

Use of Large Language Models (LLMs) and Generative AI Tools

Authors should note the following key points regarding use of large language models (LLMs) and other generative AI tools in the preparation of their submissions. These points are meant to complement IEEE's policy on AI generated text, available here: IEEE Submission Policies.

• Authorship is restricted to humans: Only human individuals may be listed as authors. Generative AI tools may not be credited as authors or co-authors. All listed authors must take full responsibility for the content of the submission.
• Disclosure of substantive use: The use of LLMs or other generative AI tools to generate or materially affect substantive content should be disclosed at the time of submission in an “AI Disclosure” statement placed at the end of the paper. The disclosure should identify the tool used and indicate which parts of the submission were generated or materially influenced. If such tools play a material role in the research methodology, analyses, experiments, or implementation, this role should also be appropriately described in the body of the paper. Authors may use the following format (if applicable):

  AI Disclosure: We used [Tool Name] to assist with [Brief Description of Use]. The tool materially affected [Sections X and Y]. More details can be found in [Section Z]. The authors verified the correctness and originality of all content including references.

  Use of AI tools solely for minor copy-editing or grammar/clarity improvements applied to the authors' own text does not require disclosure.
• Authors are fully responsible for their submissions: Authors are accountable for the accuracy, originality, and integrity of all material in their paper, including any content produced with AI assistance. This includes responsibility for errors, plagiarism, misrepresentation, or fabricated content (e.g., “hallucinatory” references) generated by such tools.

Potential concerns regarding compliance with these principles will be handled in accordance with applicable IEEE publication policies and conference procedures. Failure to comply with the disclosure policy or submissions that contain falsified, hallucinated or plagiarized contents may result in the submission being desk rejected at any stage during the submission.

---

Workshops Proposals

Proposals for half-day or full day workshops that focus on IEEE CogMI 2026 related themes are solicited. Workshop proposals should be at most 6 pages, including a biographical sketch of each instructor, and submitted to EasyChair.

---

Panels Proposals

Proposals for panel discussions that focus on future visions relevant to Cognitive Machine Intelligence are preferred. Potential panel organizers should submit a panel proposal of at most five pages, including biographical sketches of the proposed panelists to the Panel Chairs.

---

Tutorials Proposals

Proposals for full and half-day tutorials are solicited. Tutorials are intended to enhance the technical program, and as such they should be relevant to collaborative computing, networking, internet technologies, worksharing, and applications. Potential tutorial presenters should submit a tutorial proposal of at most three pages, including: description of potential audience and background knowledge expected from the audience, if any; tutorial description; biographical sketch(s) of presenter(s).

---

Review Policy

IEEE Policy and professional ethics require that referees treat the contents of papers under review as privileged information not to be disclosed to others before publication. It is expected that no one with access to a paper under review will make any inappropriate use of the special knowledge, which that access provides. Contents of abstracts submitted to conference program committees should be regarded as privileged as well, and handled in the same manner. The Conference Publications Chair shall ensure that referees adhere to this practice.

Organizers of IEEE conferences are expected to provide an appropriate forum for the oral presentation and discussion of all accepted papers. An author, in offering a paper for presentation at an IEEE conference, or accepting an invitation to present a paper, is expected to be present at the meeting to deliver the paper. In the event that circumstances unknown at the time of submission of a paper preclude its presentation by an author, the program chair should be informed on time, and appropriate substitute arrangements should be made. In some cases it may help reduce no-shows for the Conference to require advance registration together with the submission of the final manuscript.

---

Blue Sky/Vision Track Papers (By Invitation Only)

We would like to solicit papers that promote visionary ideas and blue sky thinking in areas aligned with the conference themes. These papers are expected to spark intense discussions and newer research directions/insights through potentially disruptive, controversial, or highly cross-disciplinary ideas that look forward to collaboration and internet computing space for the next 10 years and beyond. Ideas that are just being conceived, not fully developed, far from experimentally evaluated, or out-of-the box are highly encouraged. The papers should follow the same format as the regular conference papers and can be up to 10 pages.

---

Industry Track Papers

We would like to solicit papers that focus on the design, implementation, and deployment of solutions related to CogMI within industrial or government environments. The papers submitted to this track are expected to advance practical and applied research focused on the use of CogMI technologies and systems applications.

The Industry/Government Track will include papers selected through a separate program committee. Authors must clearly indicate sub-areas their papers are to be evaluated in because distinct criteria may be used for reviewing different category of submissions:

• Deployed

  Deployed systems that aim to provide real practical value to industry, Government, or other organizations, or communities. The papers should point out how the deployed system explicitly addresses CogMI issues or describe either qualitatively (lessons learned, deployment experiences, etc.) or quantitatively how these issues are addressed in systems and applications.

• Emerging

  Newer models and mechanisms or innovative solutions related to CogMI. The authors should clearly demonstrate value and interest to Industry, Government of society (e.g., scientific or medical professions; critical infrastructures). Papers that describe infrastructure development and deployment that enable the large-scale deployment of related technologies/applications or their validation are also in these areas.

---

Awards

IEEE CogMI will feature a Best Paper award and a Best Student Paper award (to be selected by the program committee/best paper award team). A paper is eligible for the Best Student Paper award if the first author is a full-time student at the time of submission. A partial travel grant or cash award may be offered to the winner student depending on fund availability.

---

Program Co-Chairs

• Jerry Jialie Shen, University of London, UK
• Amarda Shehu, George Mason University, USA
• Reza Zafarani, Syracuse University, USA