I am a fourth year PhD student at Tulane University, working with Prof. Jihun Hamm. My current research is related to understanding the robustness of machine learning under distribution shifts. I also work on developing methods for solving large-scale bilevel optimization problems that appear in popular machine learning such as hyperparameter optimization, few-shot learning, importance learning and training-data poisoning.

Previously, I have completed my Masters in Computer Science from The Ohio State University under the guidance of Prof. Jihun Hamm and Prof. Mikhail Belkin where I worked on active learning and semi-supervised learning problems.

Research Interest

To deepen the understanding of robustness of machine learning models to different types of distribution shifts such as shifts induced by corrupted data, shifts induced by adversarial attacks, etc..

Publications

• “On Certifying and Improving Generalization to Unseen Domains”.
  Code
  Akshay Mehra, Bhavya Kailkhura, Pin-Yu Chen and Jihun Hamm.
• “Certified Adversarial Defenses Meet Out-of-Distribution Corruptions: Benchmarking Robustness and Simple Baselines”.
  Jiachen Sun, Akshay Mehra, Bhavya Kailkhura, Pin-Yu Chen, Dan Hendrycks, Jihun Hamm and Z. Morley Mao.
  European Conference on Computer Vision (ECCV) 2022.
• “Understanding the Limits of Unsupervised Domain Adaptation via Data Poisoning”.
  Code, Poster
  Akshay Mehra, Bhavya Kailkhura, Pin-Yu Chen and Jihun Hamm.
  Neural Information Processing Systems (NeurIPS) 2021.
• “How Robust are Randomized Smoothing based Defenses to Data Poisoning?”.
  Code, Poster
  Akshay Mehra, Bhavya Kailkhura, Pin-Yu Chen and Jihun Hamm.
  Computer Vision and Pattern Recognition (CVPR) 2021.
• “Penalty Method for Inversion-Free Deep Bilevel Optimization”.
  Code, Poster
  Akshay Mehra, Jihun Hamm.
  Asian Conference on Machine Learning (ACML) 2021.
• “Fast Interactive Image Retrieval using large-scale unlabeled data”.
  Akshay Mehra, Jihun Hamm and Mikhail Belkin.
• “Machine vs Machine: Minimax-Optimal Defense Against Adversarial Examples”.
  Jihun Hamm and Akshay Mehra.