S. Korea Woman's school syncs w/ Theta EdgeCloud
What do we know about Ewha Womans University
It was founded in 1886 in Seoul, is the world’s largest women’s university and one of South Korea’s most prestigious private research institutions. Ewha now enrolls over 20,000 students across undergraduate and graduate programs. The university is renowned for academic excellence across disciplines including AI, life sciences, international studies, and engineering. Additionally EWHA is now home to the innovative Department of Artificial Intelligence (AI). This new lab is known as PAI Lab.
Why did PAI Lab select Theta EdgeCloud?
Being an avid follower you already know that Theta's EdgeCloud is built using high-end NVIDIA GPUs such as A100s and H100s, as well as AWS next-generation Inferentia and Trainium AI chips. The PAI Lab is looking to leverage EdgeCloud for many use cases including the usual suspects...training LLMs, reinforcing learning simulations. But expect to deploy their GenAI workloads to save cost and time by exploiting the existing Theta infrastructure. Their main use case will be around object recognition, robust learning and multi-model learning. By exploiting EdgeCloud they can further develop use cases and push the AI boundaries all while remaining flexible to the dynamics in the developing AI field.
The PAI Lab led by Assistant Professor Junhyug Noh based in the Department of Artificial Intelligence at Ewha Womans University, is focused on building AI technologies that are practical, adaptable, and effective in diverse real-world scenarios. The lab’s research spans object recognition, including detection and segmentation, as well as weakly and semi-supervised learning approaches that reduce the reliance on labeled data. Additional areas of focus include domain adaptation and robust learning methods that enhance model performance under distribution shifts, along with active learning strategies that improve data efficiency. The lab also works extensively on multi-modal learning and AI + X applications, such as medicine, surveillance, and process automation.
Summary
Theta's South Korean academic dominance continues, by onboarding EWHA WOMANS University they have locked up top tier education providers. This does have me wondering even with use of time slicing and schedulers, what is the utilization rate of Theta's GPU's? I ask as I am curious about how many large customers they can onboard and still provide shared GPU value?
Credits - BLOG Image - Theta Network = ALL RIGHTS REMAIN!
