| title | About |
|---|---|
| permalink | /about/ |
We are the Machine Intelligence and Neural Technologies (MINT) Lab at Vanderbilt University. Our overarching focus is on mathematical machine learning and (geometric) deep learning. We strive for Next-Generation Machine Learning algorithms (Next-Gen ML) that address current deficiencies in our technologies regarding sample/label efficiency, explainability, brittleness, and lifelong/continual learning.
Our research is motivated by the ever-increasing demand for automated understanding of data and predictive/generative modeling in real-world problems with a focus on biomedical application. We generally focus on developing practical Machine Learning (ML) and Computer Vision (CV) solutions for challenging problems. Our current and past research towards NextGen-ML fall under the following broad research directions:
- Label Efficient Learning
- zero/few-shot transfer learning, domain adaptation, and self-supervised learning
- Explainable ML
- invertible models, explainable graph/convolutional neural networks
- Lifelong/Continual Learning
- Overcoming catastrophic forgetting, memory replay, selective transfer
- Adversarial Attacks and Defenses
- Real-world object level adversarial attacks, Trojan attacks and defenses
Please refer to our publications for a complete list of papers covering these topics.
We strive to build a remarkably interdisciplinary research group spanning computer science, statistics, applied mathematics, biomedical sciences, and many other disciplines. Visit our people page to see more information on each person who works in the lab (publications, contact information, photos).