Workshop @ICDL2025 - 16 September 2025
Humans are perhaps the most adaptable species on Earth—not because we are inherently the best at any one thing, but because we excel at learning. We don’t come equipped with a perfect model of the world; instead, we continuously update our partial understanding through experience. We can’t read minds, but we learn to interpret others’ behaviors and intentions through interaction and empathy.
As researchers, we build models that capture some of the brain’s inner workings, describe its functional components, and replicate isolated behaviors in controlled environments. But we must go beyond static systems to bring these models into the complex, ever-changing real world. We must design adaptive, developmental architectures—systems that grow, learn, and evolve, much like human infants do.
This workshop seeks to revitalize the conversation around Developmental Cognitive Architectures, creating a space for both established experts and emerging researchers to exchange ideas, share insights, and explore how we can shape the future of intelligent, autonomous systems.
Explore how developmental principles—such as sensorimotor learning, curiosity-driven exploration, and social interaction—can inform the design of intelligent robotic systems.
Highlight cutting-edge work in cognitive robotics that draws inspiration from developmental processes.
Identify key challenges and open research questions that must be addressed to advance the field.
Foster interdisciplinary collaboration across AI, robotics, psychology, neuroscience, and engineering.
We warmly invite all those curious about the intersection of learning, cognition, and robotics to join us in this important discussion.
The level of complexity to design such developing architectures is tremendous and requires an interdisciplinary approach: however, too often we risk getting stuck in our bubble, as even dipping our toes in a new promising field is a daunting task, due to the sheer amount of knowledge published day by day. The aim of AI:CogDev is to provide an opportunity for researchers with the same goal, to make agents more autonomous, to share their ideas and results across their many fields of study. In particular, the workshop will focus on the design of Cognitive Architectures that make Developmental Learning and Embodiment their core pillars.
Topics of interest include, but are not limited to:
Cognitive Architectures applied to real or simulated embodiments
Developmental Learning and Developmental Robotics
Symbol grounding and concept emergence in autonomous systems
Active and Interactive Learning
The role of embodiment in learning
Affordance Learning
Open-ended Reinforcement Learning
Intrinsically Motivated Learning
Affection and Cognition in Development
Ethical and trust considerations of developmental agents in real-world scenarios
Brain/Psychologically -inspired development/ computational intelligence
These topics are meant to be suggestive and not exhaustive. Contributions to other relevant topics will also be considered, as long as they align with the general theme of the workshop.
The workshop organizers are committed to creating an inclusive, diverse, and equitable environment for all attendees. We recognize and value the diversity of our attendees and understand that people have different needs and preferences. To accommodate the different time zones of attendees, the workshop will be held in a hybrid format and will be recorded for later viewing. To ensure that all attendees have reliable access to the internet, we will provide technical support and troubleshooting assistance as needed. We encourage attendees from underrepresented groups to submit their work and participate in the workshop. We are committed to creating an equitable and diverse environment, and we are open to suggestions on how to improve the workshop in terms of inclusion and accessibility.
Images generated with DALL-E3