How Neuroscience Inspires Artificial Intelligence?
- Kai Du, Karolinska Institute, Sweden
- Time: 2018-04-02 10:00
- Host: Prof. Baoquan Chen
- Venue: Room 101, Courtyard No.5, Jingyuan
Neuroscience has a long history in inspiring research in the field of artificial intelligence (AI), such as deep artificial neural network, PCA, ICA and sparse coding. These supervised or unsupervised learning algorithms, however, were built upon the conceptions of neuroscience mainly developed 20 years ago. During past decades, much progress have been made in understanding information representation in real brains. In this talk, I will mainly introduce some of recent progress in both experimental and theoretical neuroscience, and relate their importance to AI. In particular, neurons can no longer be viewed as a "point-device". Neuronal dendrites are able to perform powerful computations.
I have ten years of experience in simulating "realistic" neuronal models. My main research interest is an emerging field called "dendritic computation", in particular, how dendrites encode signals in our brains and how dendritic computation forms will inspire artificial intelligence.