Ahmed Khaled

akregeb@gmail.com

About Me

Hello! I’m a B.Sc. student at Cairo University expecting graduation in July 2020.

My research interests are in stochastic and distributed optimization, federated learning and the theory of reinforcement learning. I particularly like optimization algorithms whose flavor is both stochastic and approximate. My CV is here.

I was fortunate to intern in the group of Peter Richtárik at KAUST in the summer of 2019 where I worked on the theory of distributed & stochastic optimization. Prior to this, I did some (applied) research on accelerating the training of neural networks under the supervision of Professor Amir Atiya.


Papers

Better Communication Complexity for Local SGD with Konstantin Mishschenko and Peter Richtárik. Accepted to the NeurIPS 2019 Federated Learning Workshop.

Gradient Descent with Compressed Iterates with Peter Richtárik. Accepted to the NeurIPS 2019 Federated Learning Workshop.

First Analysis of Local GD on Heterogenous Data with Konstantin Mishschenko and Peter Richtárik. Accepted to the NeurIPS 2019 Federated Learning Workshop.


Other

Some school projects I’ve worked on include an implementation of style transfer without neural networks, a multiplayer PacMan-like game in x86 assembly, and others. These are available on my GitHub account. I like reading fantasy fiction and sci-fi: Some of my favorites include Steven Erickson’s Malazan Book of the Fallen, Barry Hughart’s Bridge of Birds, and Frank Herbert’s Dune. I also like drawing and some of my favorite artists include Ilya Repin and Jean Giraud (Moebius).

Last Updated: October 2019.