Arto Maranjyan

prof_pic.jpg

I’m a PhD student at KAUST, advised by Prof. Peter Richtárik. My research focuses on optimization for machine learning (ML) and federated learning (FL), where I contribute to the development of distributed and randomized optimization algorithms. I’m currently focused on addressing system heterogeneity issues in distributed ML and FL, with an emphasis on asynchronous methods.

Before starting my PhD, I earned my MSc and BSc from Yerevan State University. During my bachelor’s, I co-authored several papers in Harmonic Analysis under the guidance of Prof. Martin Grigoryan.

Outside of academics, I enjoy dancing bachata, playing board games, ultimate frisbee, and foosball.

Recent News

May 01, 2025 Honored to receive the CEMSE Dean’s List Award for academic and research excellence at KAUST — with a $2,500 prize.
May 01, 2025 A paper has been accepted to UAI 2025.
MindFlayer SGD: Efficient Parallel SGD in the Presence of Heterogeneous and Random Worker Compute Times.
May 01, 2025 Two papers accepted to ICML
  • Ringmaster ASGD: The First Asynchronous SGD with Optimal Time Complexity
  • ATA: Adaptive Task Allocation for Efficient Resource Management in Distributed Machine Learning
Mar 23, 2025 I spent two weeks visiting Yi-Shuai Niu at the Beijing Institute of Mathematical Sciences and Applications (BIMSA), collaborating on a project on Server-Assisted Federated Learning. During the visit, I also gave talks at three universities: PKU, BUAA, and BIMSA.
Feb 10, 2025 I’ll be giving a talk at the AMCS/STAT graduate seminar at KAUST on February 27, presenting our paper, Ringmaster ASGD: The First Asynchronous SGD with Optimal Time Complexity.
Feb 04, 2025 New paper out ATA: Adaptive Task Allocation for Efficient Resource Management in Distributed Machine Learning. Co-authored with El Mehdi Saad, Peter Richtárik, Francesco Orabona. [LinkedIn post]
Jan 28, 2025 New paper out: Ringmaster ASGD: The First Asynchronous SGD with Optimal Time Complexity. Co-authored with Alexander Tyurin, Peter Richtárik. [LinkedIn post]

Selected Publications

  1. Ringmaster ASGD: The First Asynchronous SGD with Optimal Time Complexity
    Artavazd Maranjyan, Alexander Tyurin, and Peter Richtárik
    arXiv:2501.16168, 2025
  2. ATA: Adaptive Task Allocation for Efficient Resource Management in Distributed Machine Learning
    Artavazd Maranjyan, El Mehdi Saad, Peter Richtárik, and Francesco Orabona
    arXiv:2502.00775, 2025
  3. MindFlayer SGD: Efficient Parallel SGD in the Presence of Heterogeneous and Random Worker Compute Times
    Artavazd Maranjyan, Omar Shaikh Omar, and Peter Richtárik
    In The 41st Conference on Uncertainty in Artificial Intelligence , 2025
  4. GradSkip: Communication-Accelerated Local Gradient Methods with Better Computational Complexity
    Artavazd Maranjyan, Mher Safaryan, and Peter Richtárik
    Transactions on Machine Learning Research, 2025