Embedded Files
What we do
We are a group of curious researchers dedicated to understanding machine intelligence, with a focus on building intelligent models that generalize effectively and are reliable. To achieve this, we study a wide array of deep learning topics, including foundation models, survival analysis, and out-of-distribution (OOD) detection. Our research extends into critical domains, such as healthcare and material science, where these models can have significant impact. Through these efforts, we aim to push the boundaries of what machine intelligence can achieve in solving real-world problems.
Recent publications
In the media
Prashnna Gyawali and doctoral student Alina Devkota at West Virginia University developed AI models trained on electrocardiogram data from over 55,000 rural Appalachian patients to better detect heart failure, a condition especially prevalent in the region. Their study, published in Scientific Reports, showed that deep learning models like ResNet outperformed others in predicting ejection fraction, offering a low-cost alternative to echocardiograms and improving diagnostic equity for underserved rural populations.
WVU experts discuss artificial intelligence at recent WV Press Association convention
At the West Virginia Press Convention, experts from West Virginia University discussed the pervasive influence of artificial intelligence (AI) and its potential risks. Dr. Prashnna Gyawali highlighted concerns about AI-generated deep fakes and the challenges of distinguishing real from fake content, emphasizing the need for stronger policies and AI ethics education.
Art History, Computer Science programs collaborate to launch Artful Algorithms project
The Artful Algorithms project, led by art history professor Megan Leight and computer science professor Prashnna Gyawali at West Virginia University, seeks to enhance AI's ability to decipher ancient Maya glyphs, a crucial step in preserving and understanding ancient civilizations. Supported by a grant from the WVU Humanities Center, the interdisciplinary project addresses the limitations of current AI systems in recognizing Maya script.
Student awards and honors
2nd Place - International Disease Diagnosis Challenge@MICCAI | 2025
Best Poster - Jacob Thrasher | WVU AI Symposium | 2025
2nd Place (poster) - Alina Devkota | WVU AI Symposium | 2025
Best Poster - Shivam | Lane Department Graduate Reseach Symposium | 2024
2nd Place (poster) - Jacob Thrasher | Lane Department Graduate Research Symposium | 2024
Page updated
Google Sites
Report abuse