I am recruiting graduate RAs and postdocs for 2021. If you are passionate about Machine Learning and have strong analytical skills, please contact me. Students with degrees from other disciplines such as Math, Physics and Stats are also encouraged.
- Akshay will be presenting our paper titled "How Robust are Randomized Smoothing based Defenses to Data Poisoning?" at the Computer Vision and Pattern Recognition (CVPR) 2021. We will release the camera-ready version soon.
- Chanho will be presenting our current work on "Machine Learning Approaches in Digital Health" at the 14th Annual Western Afib Symposium.
- I gave a tutorial talk on Adversarial Machine Learning (slides are here) at the Pattern Recognition and Machine Learning Winter School 2021.
- Byunggill will be presenting "Machine Learning with Electronic Health Records is vulnerable to Backdoor Trigger Attacks" at the AAAI 2021 Workshop on Trustworthy AI for Healthcare.
- Akshay will be presenting "Stealthy Poisoning Attack on Certified Robustness" at the NeurIPS'20 Workshop on Dataset Curation and Security. There is also the poster and the full arXiv version.
- I'm the Tulane lead of the Louisiana Material Design Alliance (LAMDA). We have recently been awared $20 milion NSF RII grant.
- Welcome Chanho Lim and Farzad Azarmi to the Machine Learning Group !!!
- The lecture slides on Deep Generative Models for CMPS6720 (Machine Learning) has been uploaded.
- The slides for the Nov 4th talk on Adversarial Machine Learning given at Tulane Medical School has been uploaded.
- We have a new paper "Penalty Method for Inversion-Free Deep Bilevel Optimization" uploaded on arXiv.
- The accompanying tensorflow scripts can be found on github.
All aspects of Machine Learning. Specific topics include the following:
- Foundations of Deep Learning
- Adversarial Learning and Optimization
- Private and Secure Machine Learning
- Medical Image Analysis and Computational Anatomy
- Nonlinear Dimensionality Reduction and Manifold Learning
You can find the list of publications from Google scholar.