[Seminar]"Towards understanding the learning rules for human-like neural representations" by Asaki Kataoka

Dear all,

Neural Computation Unit (Doya Unit) would like to invite you to a seminar as follows.

Speaker: Asaki Kataoka 

1. Graduate School of Arts and Sciences, The University of Tokyo (Ph.D. course)
   - UTokyo WINGS-ABC Leading Research Assistant
   - Research Fellowship for Young Scientists (DC2), Japan Society for the Promotion of Science

Title: Towards understanding the learning rules for human-like neural representations

Abstract: Understanding how the neural networks in the brains of humans and other organisms acquire useful representations of sensory inputs and how these representations support appropriate computations is a key question in the field of computational neuroscience. In this talk, I will present several research projects I have undertaken, including computational models of colour constancy, structural analyses of visual object representations learnt via self-supervised methods, and Bayesian inference that leverages the stochastic properties of recurrent neural networks. In addition, I will discuss my current work on how categorical colour representations are formed and oin modeling colour vision deficiency. Building on these past and ongoing research experiences, I will conclude with a brief overview of the new research directions I plan to pursue as a postdoc.

We hope to see many of you at the seminar.

Sincerely,
Neural Computation Unit
Contact: ncus@oist.jp