27-31 October 2025 | Dublin, Ireland | ACM Multimedia 2025
While multimodal systems excel at basic emotion recognition, they struggle to understand why we feel and how emotions evolve. This workshop pioneers cognitive AI that interprets human affect through multimodal context and causal reasoning. Join us in redefining emotional intelligence for healthcare robots, empathetic chatbots, and beyond.
Stay updated: https://CogMAEC.github.io/MM2025
Welcome to the 1st CogMAEC Workshop, proudly co-located with ACM Multimedia 2025!
As human-computer interaction evolves, emotional intelligence and empathy are becoming essential capabilities of intelligent systems. The CogMAEC Workshop (Cognition-oriented Multimodal Affective and Empathetic Computing) aims to push the boundaries of traditional affective computing by exploring the next frontier: cognitive emotional understanding.
While previous work in multimodal affective computing has focused on recognizing basic emotions from facial expressions, speech, and text, this workshop sets its sights on deeper challenges — understanding the "why" behind emotions, reasoning over context, and simulating human-like empathetic responses. With the recent advances in Multimodal Large Language Models (MLLMs), the time is ripe to rethink how machines perceive, reason, and respond to human emotions.
CogMAEC'25 brings together researchers and practitioners working on:
The workshop will cover both traditional multimodal emotion recognition techniques and cutting-edge cognition-driven methodologies. We aim to foster meaningful discussion and collaboration at the intersection of affective computing, cognitive modeling, and multimodal AI.
Join us as we collectively reimagine what emotional AI can become — not just smarter, but more human.
All workshop details, schedules, and updates can be found on our website.
We invite contributions in three categories:
1. Novel Position or Perspective Papers (4–8 pages, excl. references, archival). Forward-looking works that propose new ideas, conceptual frameworks, or identify open challenges aligned with the workshop themes. Accepted papers will appear in the CogMAEC 2025 Proceedings (co-located with MM '25).
2. Non-archival Featured Papers (title + abstract + original manuscript). Influential papers already published in top venues, or well-curated summaries of substantial prior work. These submissions are presentation-only and will not be included in the proceedings.
3. Demonstration Papers (≤ 2 pages, excl. references, archival). Short papers describing prototypes, tools, or systems that showcase practical implementations or evaluation methodologies. Accepted demos will be published in the CogMAEC 2025 Proceedings (co-located with MM '25).
All accepted submissions will be invited to present their work at the workshop.
A Best Paper Award will be announced during the workshop.
The workshop welcomes submissions on the following topics (but not limited to):
1) Traditional Multimodal Affective Computing
2) MLLM-based Multimodal Affective Computing
3) Cognition-oriented Multimodal Affective Computing
October 27-28, 2025 (AoE)
August 23, 2025 (AoE)
July 20, 2025 (AoE)
June 30, 2025 (AoE)
April 15, 2025 (AoE)
March 30, 2025 (AoE)
All submissions must be written in English and follow the current ACM two-column conference format. Page limits are inclusive of all content, including figures and appendices. Submissions must be anonymized by the authors for review.
Authors should use the appropriate ACM templates: the "sigconf" LaTeX template or the Interim Word Template, both available on the ACM Proceedings Template page. Alternatively, authors can prepare their submissions using Overleaf's official ACM templates.
Please use \documentclass[sigconf, screen, review, anonymous]{acmart} when preparing your LaTeX manuscript for submission and review.
For any questions about the workshop, please contact us through:
Email:
Google Group: https://groups.google.com/g/cogmaec