Pyone MaungMaung

I am a postdoctoral researcher at Echizen Laboratory, National Institute of Informatics (NII), Japan.

In NII, I work on security of foundation models. I did my PhD at Tokyo Metropolitan University, where I was advised by Prof. Hitoshi Kiya.

Email  /  CV  /  Scholar  /  X  /  Github

Research

I am interested in evaluation of neural network models in terms of utility, security, and privacy.

Block-wise Image Transformation with Secret Key for Adversarially Robust Defense
AprilPyone MaungMaung, Hitoshi Kiya
TIFS, 2021
[code] [arXiv]

Hiding a model's decision boundary to prevent from adversarial attacks by training the model with block-wise transformed images with a secret key.


Template from Jon Barron