Vishnu Boddeti
Michigan State University, MI·Computer Science
Artificial intelligence (AI) now helps diagnose disease from medical images and manage patient care in emergency rooms. But these systems can carry over errors from historical or hand-labeled data.
- Fields
Computer Science
Engineering
Medicine & Clinical Research
- Topics
machine learning
artificial intelligence
AI fairness
- Grants
- 1 known grant, through Jul 2029
What they work on
- Derive theoretical limits on the trade-offs among predictive accuracy, robustness across patient populations and privacy in healthcare AI models, and evaluate them in practice.
- Analyze these trade-offs in models that diagnose skin and soft tissue infections, comparing centrally trained models with personalized models tailored to patient characteristics.
- Develop algorithms that reach optimal performance under accuracy, fairness and privacy trade-offs by integrating varied, distributed, noisy and self-supervised data sources.
More about the research
As a result, they may work less well for some patient groups or leak sensitive patient information. This project asks a basic question: how much accuracy, fairness across patient groups and privacy can a healthcare model have at the same time, and what limits the trade-offs between them?
Why it matters
Patients, especially those in groups that AI tools now serve poorly, could benefit from diagnostic models that are both accurate and protect their private data. Clear limits and benchmarks could guide hospitals and developers in building safer, more reliable AI for tasks such as diagnosing skin and soft tissue infections.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Guang Wang
Florida State University, FL·Computer Science
Build cross-domain data curation tools for MobilityNet that combine real mobility datasets from different sources to reduce data bias.
Assefaw Gebremedhin
Washington State University, WA·Computer Science
Develop robust and trustworthy machine learning algorithms with anomaly-detection features that flag malicious activity in AI systems used in healthcare, autonomous vehicles and smart cities.
Sanjay Purushotham
University of Maryland-Baltimore County, MD·Computer Science·Early-career
Develop multi-view federated learning algorithms that train shared models across hospitals holding different kinds of patient data, without moving the data between sites.
The school
Michigan State University: 28 labs on PI Pages, $932M a year on research, 555 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.