Our group aims to help understand the brain and mind by building computational models. We focus on building artificial neural network models for cognition. Our research style is influenced by neural circuit models from neuroscience, deep neural networks from machine learning, and normative models from cognitive science.
Our work centers around higher cognitive capabilities that are difficult to study with hand-designed models. We focus on scientific questions and insights, with a lesser emphasis on the usual machine learning benchmarks. By combining bottom-up and top-down modeling approaches, we can better integrate knowledge of the brain from different levels, e.g. molecular, cellular, circuit, system, and behavior.
Our long-term goals are to
- understand fundamental questions about the human brain and mind
- build better artificial intelligence and life using insights from the brain
- help develop treatment of psychiatric disorders based on multi-scale understanding of the brain