Divya Sharma
About Me
Hi, I am a Postdoctoral Researcher, working with Prof. Yoshua Bengio and Prof. Alex Hernandez-Gracia at MILA and UdeM. My current work focuses on developing and applying machine learning models that can be used to acclerate discovery and experimentation on materials aimed at tackling the challenges that humanity faces today. Broadly, I am interested in combining domain expertise / physics into machine learning techniques as a tool to supplement scientific discovery.
Current Projects
- Developing property prediction proxies for materials using multimodal data
- Training computer-vision model using multimodal data in low-data regime to inform active learning decisions for carbon-capture catalysts
Updates
| Dec 2024 | Graduated from Johns Hopkins University with PhD in Chemical Engineering |
| Nov 2023 | Presentation at MRS Fall in Boston |
| Jul 2023 | CIFAR - DLRL Summer School |
| May 2022 | Poster Presentation at MRS Spring 2022 |
| Nov 2020 | Poster presentation at AIChE Annual Conference |
| Aug 2019 | Graduated with MS in Chemical Engineering from Cornell University |
| Aug 2017 | Graduated from IIT Bombay with B. Tech in Chemical Engineering |
You should checkout
- Curated dataset and benchmarks for solid-state electrolytes: Paper
- Generating new periodic materials using GFlowNets, accepted at Neurip AI4Mat workshop: Paper
- Nudged elastic band benchmarking and best practices: Paper
- Trained a crystal graph convolutional neural network to conduct screening for promising HOIP candidates: GitHub Repo
- Simple recreation of this Trust Game using Python: GitHub Repo
- A simple Tic Tac Toe Game written using Python: Github Repo