Emir Ceyani
Ph.D. Student at vITAL Lab


Recent News

SpreadGNN accepted to FL-ICML’21 Workshop!

July 2021

SpreadGNN accepted to DLG-KDD’21 Workshop!

June 2021

Accepted to LOGML summer school!

June 2021

Short Bio

I am a Ph.D. student in Prof. Salman Avestimehr’s Information Theory and Machine Learning (vITAL) research lab at the University of Southern California (USC) in Los Angeles . I pursue research on the foundations and applications of federated learning, graph neural networks, probabilistic deep learning, and information theory. I seek to design learning algorithms that can effectively handle the non-I.I.D. nature of real-life data in federated setting. I try to actively contribute to FedML, a promising research library for federated learning as I believe that federated learning systems should be accessible for everyone. Previously, I was a Master’s student at Bilkent University, Ankara, Turkey, and an R&D engineer at Turkcell Technology under my master’s advisor, Dr. Salih Ergut with a prestigious 5G & Beyond Graduate Support Program.


Ph.D. in Electrical & Computer Engineering
University of Southern California | Los Angeles, CA

Advisor: Prof. Salman Avestimehr

Jan 2021 - Present
M.Sc. in Electrical & Electronics Engineering
Bilkent University | Ankara,TR

Thesis Title: Spatio-temporal Forecasting Over Graphs with Deep Learning
Advisor: Dr. Salih Ergut

2018 - 2020
B.Sc. in Electrical & Electronics Engineering
Bilkent University | Ankara,TR
2013 - 2018


Google Scholar


SpreadGNN: Serverless Multi-task Federated Learning for Graph Neural Networks
C. He*, E. Ceyani*, K. and Balasubramanian*, M. Annavaram, and S. Avestimehr
Accepted to FL-ICML'21 & DLG-KDD'21 workshops .Three co-1st authors have equal contribution (alphabetical order) 2021
[1] [abs] [pdf] [code] [slide] [poster]
FedGraphNN: A Federated Learning System and Benchmark for Graph Neural Networks
C. He*, K. Balasubramanian*, E. Ceyani*, Y. Rong, P. Zhao, J. Huang, M. Annavaram, and S. Avestimehr
Accepted to ICLR - DPML 2021 & MLSys - GNNSys'21 workshops (equal contribution) 2021
[2] [abs] [pdf] [code] [slide] [poster] [video]


Spatio-temporal forecasting over graphs with deep learning
E. Ceyani
Bilkent 2020
[3] [abs] [pdf]


A highly efficient recurrent neural network architecture for data regression
T. Ergen and E. Ceyani
26th Signal Processing and Communications Applications Conference, SIU 2018, Izmir, Turkey 2018
[4] [pdf]

Industry Experience

5G VATS R&D Engineer | Turkcell Technology | Istanbul, TR
Developed deep learning algorithms for forecasting spatio-temporal grid and graph-structured mobile traffic series collected from 4G/5G base stations under non-stationary environments.

Oct 2018 - Dec 2020

Honors & Awards

Accepted to LOGML Summer School 2021
Deep|Bayes & PRAIRIE AI ML summer schools.

Travel & Accommodation Grant

5G & Beyond Graduate Support Program and Bilkent Graduate Scholarship

Funded during M.Sc. studies by Bilkent University and ICT(Awarded for the first time).

2018 - 2020
Research Excellence Award

Awarded by Bilkent University Electrical and Electronics Engineering Department.


Invited Talks

MIT - FL Reading Group 2021

Professional Activities

Reviewing: IEEE TNNLS, TBDCS 2019 -
Teaching Assistant at ProbAI 2021 Summer School 2021


FedML-AI/FedML | 703 | FedML - A Research-oriented Federated Learning Library 2020
FedML-AI/FedGraphNN | 51 | FedGraphNN - A Research-oriented Federated Learning Library for Graph Neural Networks 2021
FedML-AI/SpreadGNN | 3 | SpreadGNN - Serverless Multi-Task Federated Learning for Graph Neural Networks 2021


Languages C, C++, Java, MATLAB, Julia, Python
Frameworks JAX, NumPy, Pandas, PyTorch, Pyro, SciPy, Keras
Tools Linux, vim, git, tmux, zsh

All credit goes to Brandon Amos. Last updated on 2021-07-22