Skip to main content

Prashnna Gyawali

Join us on Wednesday, February 28 at 2:30pm in White Hall G09 for a colloquium presented by Prashnna Gyawali, Assistant Professor in the WVU Lane Department of Computer Science and Electrical Engineering. He will speak on Generative AI beyond vision and language. Continue reading for his abstract and biography.

Abstract

Generative AI has significantly advanced in vision and language fields, bringing innovations that have captured attention within the technical community and the broader public. Models like DALL·E and StyleGAN have transformed image generation and manipulation, allowing the creation of artwork and photorealistic images from textual descriptions. Meanwhile, ChatGPT and other large language models have revolutionized natural language processing, enabling tasks from writing coherent articles to conducting human-like conversations. These developments highlight generative AI's vast potential and lay a foundation for its application in areas beyond vision and language. This talk will explore the frontier of Generative AI, examining its novel and transformative uses across various fields. It will discuss deep generative models, including Variational Autoencoder (VAE), Generative Adversarial Network (GAN), and recent diffusion models, with a focus on VAE and its connection to variational inference. The presentation will showcase examples from my research in healthcare, particularly using VAE for electrocardiogram signals and medical imaging, and delve into our group's ongoing projects in material science. Here, we'll illustrate how generative models are being applied to simulate structures and expedite new material design.

Biography

Prashnna Gyawali

Dr. Prashnna K Gyawali serves as an Assistant Professor in the Lane Department of Computer Science and Electrical Engineering at West Virginia University. Before this role, he was a postdoctoral scholar at the School of Medicine at Stanford University. Dr. Gyawali obtained his Ph.D. in Computing and Information Sciences from the Rochester Institute of Technology (RIT) and completed his Bachelor's in Engineering at IOE Pulchowk Campus in Nepal. His research intersects machine learning and its applications, focusing on developing robust and fair AI models for real-world challenges. During his Ph.D., he interned at Google Health and Verisk Analytics. Dr. Gyawali has authored over 20 peer-reviewed research articles, presented at leading AI conferences such as ICLR, ICDM, MICCAI, and published in prestigious journals including Nature Medicine, Nature Communications, and Nature Communication Biology. His work has received funding from significant sources such as WVHPEC, DARPA, and NSF CITeR.