Heshan Fernando

Heshan Fernando Ph.D. Student Department of Electrical, Computer, and Systems Engineering Rensselaer Polytechnic Institute E-mail: fernah@rpi.edu Google Scholar

About me

My background is in Signal Processing and Deep learning

My current research focuses on the theory and applications of Optimization and Machine learning.

I am fortunate to be advised by Prof. Tianyi Chen.

Education

• Rensselaer Polytechnic Institute, Troy, NY: Ph.D., Electrical Engineering, 2021 -

• University of Moratuwa, Sri Lanka: B.Sc., Electronic and Telecommunication Engineering, 2014 - 2018

Highlights

• Dec 2023: Our paper was accepted in the 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)!:

  • Variance Reduction Can Improve Trade-off In Multi-Objective Learning.

• Sep 2023: Our paper was accepted in the 2023 Conference on Neural Information Processing Systems (NeurIPS)!:

  • Three-Way Trade-Off in Multi-Objective Learning: Optimization, Generalization and Conflict-Avoidance.

• May 2023: Presented our work at ICLR 2023! (presentation and slides)

• Jan 2023: Our paper was accepted (notable-top-5%) in the 2023 International Conference on Learning Representations (ICLR)!:

  • Mitigating Gradient Bias in Multi-objective Learning: A Provably Convergent Stochastic Approach.

      Another paper was accepted in the 2023 American Control Conference (ACC)!:

  • On the Stability Analysis of Open Federated Learning Systems.