数理科学チームセミナー「Reconstruction of training samples from loss functions」

This is an information for AIP math-sci team's seminar on 6/11(Mon.)

speaker: Akiyoshi Sannai
title: Reconstruction of training samples from loss functions
abstract:
I will talk about weak reconstruction theorem of training samples from loss functions. This theorem shows that for deep neural networks with square loss functions, if we know all input and out put of the loss function, then we can reconstruct training samples upto scalar multiplication.

time: 13:00
place: Yagami campus, Keio Univ.

room: Building 14, Room 631A

If you are interested in, please feel free to join.