Neural architectures: what are they good for anyway?
Computational Neuroscience, Cognition, Mathematical Modeling and seminar by Dr Dan Goodman, Imperial College London
Hosted by NeuroAI UCL
Wednesday 13:00–14:10 London (GMT+0)
Ended
Abstract
The brain has a highly complex structure in terms of cell types and wiring between different regions. What is it for, if anything? I'll start this talk by asking what might an answer to this question even look like given that we can't run an alternative universe where our brains are structured differently. (Preview: we can do this with models!) I'll then talk about some of our work in two areas: (1) does the modular structure of the brain contribute to specialisation of function? (2) how do different cell types and architectures contribute to multimodal sensory processing?
Topics
More from NeuroAI UCL
All talks- Jul 9, 2025Kim Stachenfeld
Neurobiological constraints on learning: bug or feature?
Jun 11, 2025Cian O’DonellUnderstanding reward-guided learning using large-scale datasets
May 14, 2025Kim StachenfeldRelating circuit dynamics to computation: robustness and dimension-specific computation in cortical dynamics
Apr 23, 2025Shaul Druckmann
More with Dan Goodman
All their workSNUFA 2026
Workshop · Nov 4–5, 2026Adversarial-inspired autoencoder framework for salient sensory feature extraction
ePoster · Sep 29, 2024Bernstein Conference 2024Bootstrapping the auditory space map via an innate circuit
ePoster · Sep 29, 2024Bernstein Conference 2024Dynamics of specialization in neural modules under resource constraints
ePoster · Sep 29, 2024Bernstein Conference 2024
Upcoming in Computational Neuroscience
All seminarsConferences in Computational Neuroscience soon
All conferencesNeuroAlliance 2026
Nov 29 – Dec 2, 2026Sydney, Australia