My research is broadly focused on using software-driven techniques to make hardware more reliable. Specifically, I have explored approximate computing and software testing methodologies to improve resiliency analysis; I developed a software-based technique for GPU instruction replication; and I am currently exploring the reliability of neural networks for domain-specific resiliency.
- Computer Architecture
- Software Testing
- Machine Learning
- Approximate Computing
- New: Received the Certificate of Mentorship from the Graduate College at UIUC for mentoring two undergraduate students in research!
- “Approximate Checkers” is accepted at WAX 2019!
- Released PyTorchFI, a runtime error injection tool for PyTorch!
Go ahead and
pip install pytorchfi!
- Received the Lynn Conway Research Award for Best Technical Demonstration at ADA!
- Received the Mavis Future Faculty Fellowship Award for 2019-2020!
- Presented our paper titled, “Minotaur: Adapting Software Testing Techniques for Hardware Errors” at ASPLOS 2019!
- Invited to attend the 7th Heidelberg Laureate Forum, 1 of 200 young researchers invited worldwide!
- Presented our paper titled, “Optimizing Software-Directed Instruction Replication for GPU Error Detection” at SC 2018!