Assefaw Gebremedhin
Washington State University, WA·Computer Science
Artificial intelligence now runs inside billions of internet-connected devices, from medical equipment to cars to city sensors. That spread makes the data these devices generate hard to keep private and secure, and attackers actively hunt for weak points.
- Fields
Computer Science
Engineering
- Topics
cybersecurity
machine learning security
trustworthy AI
- Grants
- 1 known grant, through Jul 2028
What they work on
- 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.
- Design cryptographic defenses against AI-specific threats using differential privacy (adding noise to protect individuals), structure-preserving encryption and secure multi-party computation (joint computing without sharing private inputs).
- Apply these cryptography schemes to real-world data to analyze large-scale machine learning models, using high-performance computing centers in Sweden.
- Address secure data sharing among interconnected devices and real-time threat detection across diverse cyber systems through paired U.S.–Swedish student research projects.
More about the research
A central question is how to build machine learning (ML) systems that keep working correctly and protect sensitive data even when someone is trying to break them.
Why it matters
Hospitals, vehicle makers and city infrastructure operators increasingly depend on AI running on connected devices that attackers target. Privacy-preserving and attack-resistant ML methods could help protect patient data, vehicle systems and critical infrastructure. Training students in both countries in these methods could also help ease the shortage of skilled cybersecurity professionals.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Omkant Pandey
Stony Brook University, NY·Computer Science·Early-career
Construct secure multi-party computation protocols that remain secure when many sessions run concurrently and the adversary is a quantum computer, closing open questions in post-quantum secure computation.
Vishnu Boddeti
Michigan State University, MI·Computer Science
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.
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.
The school
Washington State University: 22 labs on PI Pages, $403M a year on research, 259 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.