Neuromorphic Tug-Of-War v2.0: Neuroscience and AI at different timescales

Neuroscience has played a foundational role in shaping Artificial Intelligence (AI), particularly by inspiring the development of Neuromorphic hardware. However, recent breakthroughs in AI, driven by Machine Learning (ML), have achieved high performance while straying from biological principles. This shift raises ongoing questions about the importance of biological realism in Neuromorphic Computing. In last year’s edition of this workshop, we initiated a community-wide conversation to examine this tension. Following the workshop, we conducted a broad online survey to gather current perspectives from the Neuromorphic Engineering community. The responses revealed a new kind of “tug-of-war”: between short-term and long-term goals, and between top-down theoretical approaches and bottom-up engineering solutions. This year’s workshop builds directly on those findings and explores their implications for the future of the field. We have invited speakers with diverse backgrounds, including Neuroscience, Machine Learning, and Neuromorphic Engineering, who are actively engaged in addressing these questions. They will present recent Neuroscience discoveries that have not yet been fully integrated into Neuromorphic systems, reflect on important lessons from the development of ML, and critically reassess the foundational goals of the field. We will also present quantitative insights from the community survey to anchor the discussion and foster a dialogue aimed at identifying concrete stepping stones for the future of Neuromorphic Computing, both in the short term and in the long term.