Atal Narayan Sahu

Atal Narayan Sahu AI and Quantum Computing Engineer

About

I’m an incoming graduate student in Quantum Science and Technology at UCLA. My current research interests lie in computational methods for scientific discovery. I aim to develop generative and physics-informed ML models and explore hybrid quantum-classical algorithms to tackle inverse design, property prediction, and synthesis planning in drug discovery, materials design, and quantum chemistry.

Previously, at Regology, I led the development of Reggi, an LLM-based legal assistant. Earlier, at KAUST, I researched communication-efficient methods for distributed deep learning.

If you’re interested in learning more about my professional background, feel free to check out my CV.

Research

Publications

1. REFL: Resource-Efficient Federated Learning
   A. M. Abdelmoniem, A. N. Sahu, M. Canini, S. Fahmy
   In ACM EuroSys’23
   [Link]

2. On the Convergence Analysis of Asynchronous SGD for Solving Consistent Linear Systems
   A. N. Sahu, A. Dutta, A. Tiwari, & P. Richtárik
   In Linear Algebra and Applications
   [Link]

3. Rethinking gradient sparsification as total error minimization
   A. N. Sahu, A. Dutta, A. M. Abdelmoniem, T. Banerjee, M. Canini, & P. Kalnis
   In NeuRIPS’21
   [Link]

4. Efficient Sparse Collective Communication and its application to accelerate Distributed Deep Learning
   J. Fei, Chen-Yu Ho, A. N. Sahu, M. Canini, & A. Sapio
   In ACM SIGCOMM 2021
   [Link]

5. On the Discrepancy between the Theoretical Analysis and Practical Implementations of Compressed Communication for Distributed Deep Learning
   A. Dutta, E. H. Bergou, A. M. Abdelmoniem, Chen-Yu Ho, A. N. Sahu, M. Canini, & P. Kalnis
   In Thirty-Fourth AAAI Conference on Artificial Intelligence (AAAI) 2020
   [arXiv]

6. Natural Compression for Distributed Deep Learning
   S. Horváth, Chen-Yu Ho, L. Horváth, A. N. Sahu, M. Canini, & P. Richtárik
   3rd Annual Conference on Mathematical and Scientific Machine Learning, PMLR
   [Link]

Contact