中央研究院 資訊科學研究所

活動訊息

友善列印

列印可使用瀏覽器提供的(Ctrl+P)功能

TIGP (BIO)—How Does the Brain Represent Multidimensional Information of Environments, and How Could It Be Learned?

:::

Today!

TIGP (BIO)—How Does the Brain Represent Multidimensional Information of Environments, and How Could It Be Learned?

  • 講者徐經倫 博士 (中央研究院 生物醫學科學研究所)
    邀請人:TIGP (BIO)
  • 時間2021-12-09 (Thu.) 14:00 – 16:00
  • 地點資訊所新館101演講廳
摘要

I am a physiologist interested in the evolutionary solutions for addressing the needs of storing and processing complex information in the brain. Not a bioinformatician, the goal of this talk is hopefully to provide a primer for a dialogue between the two very different fields, under the theme of how complex information is represented and transformed in an intelligent system. Importantly, a flexible system would require a more dynamic, learnable architecture. Recent advances in artificial neural networks and dimensional reduction applied to neurophysiology have inspired new ways of looking at how computational frameworks based on real neural circuits store and process information. I will discuss my experience in studying neuron computation and the biological mechanisms for learning multidimensional neural representations. Moreover, outlook on novel approach to unravel the mechanistic basis of neural computation, challenges of complex data analysis, and some implication on deep learning may also be discussed.

BIO

Dr. Ching-Lung Hsu completed his postdoc training and worked as a Research Scientist at Janelia Research Campus at Howard Hughes Medical Institute (HHMI). Prior to that, he received a Bachelor degree with a major in Zoology/Life Science, a minor in Electrical Engineering and a Ph.D. from National Taiwan University. Focused on memory and spatial navigation, he is interested in understanding how computational/algorithmic properties of neurons, circuits and behavior are causally linked to each other. In his developing lab, the group is integrating several different approaches ranging from single-cell electrophysiology, imaging to computational modeling.