Disclosure: This text doesn’t characterize funding recommendation. The content material and supplies featured on this web page are for instructional functions solely.
Theta Labs will develop its AI analysis partnership as Kangwon Nationwide College joins Theta Edgecloud's distributed GPU community.
desk of contents
San Francisco, CA – March 5, 2025 – Theta Labs, a worldwide chief in distributed cloud infrastructure for AI, media and leisure, has introduced its Kangwon Nationwide College Information Analytics & Machine Intelligence (DAMI) Lab, its newest educational consumer, led by Professor Kyeongpil Kang.
Outfitted with Kangwon Nationwide College as Theta Edgecloud's twenty third AI and video buyer, it has achieved over 85% GPU utilization throughout the worldwide community, incomes over 230 million TFUELs locked by elite booster node operators.
Theta EdgeCloud is proud to energy a lot of main enterprises reminiscent of NHL's Las Vegas Knights and international esports groups, together with FlyQuest and Evil Geniuses, and top-tier educational establishments, Seoul Nationwide College, KAIST, College of Oregon, Michigan State College, and Singapore's famend analysis establishment NTU, increasing its function in advancing AI analysis and real-world purposes.
About Kyeongpil Kang and Professor Dami Lab
Professor Kyeongpil Kang is an assistant professor within the Division of Laptop Science and Engineering at Kangwon Nationwide College, specializing in pure language machining (NLP), machine studying, and information mining. He acquired his PhD. At Laptop Science at Korea College, he studied below Professor Jaegul Choo and later joined academia after working as a machine studying researcher for SK Telecom AI Heart, Spatter Lab and Liner.
Kangwon Nationwide College's Information Evaluation and Machine Intelligence (DAMI) Lab focuses on large-scale AI analysis with experience in pure language processing, multimodal AI, and information extraction.
The lab focuses on:
- Pure Language Processing and Massive-Scale Language Fashions (LLMS): Investigation of LLM fine-tuning, Considering Chain (COT), Search Complete Era (RAG), and AI Alignment. Dami Lab additionally explores efficiency optimization and simulation-based studying for LLM brokers.
- Machine Studying and Deep Studying: Deal with domain-specific AI challenges by constructing coaching datasets and creating deep studying options for real-world purposes.
- Information Mining and AI-Pushed Insights: Apply machine studying and LLM-Pushed analyses to fields reminiscent of social science and medical AI to uncover insights to deal with complicated issues.
Progress of AI analysis utilizing Theta Edgecloud
Dami Lab leverages Theta EdgeCloud, the primary hybrid cloud edge GPU computing platform, to drive innovation in pure language processing (NLP), machine studying and information mining. By this collaboration, researchers at DAMI Lab may also entry Theta's distributed GPU infrastructure, offering cost-effective, scalable AI computing energy to advance LLM coaching, multi-agent AI techniques, and multimodal studying purposes.
- Enhanced NLP and LLM Fashions: Practice superior language fashions utilizing scalable, high-performance GPU computing.
- Develop a strong multi-agent AI system: experiment with multi-agent dialogue fashions and hostile AI.
- Edge's Energy AI: Deploy AI fashions in real-time environments utilizing Theta's hybrid cloud edge structure.
- Optimizing AI Coaching Effectivity: Exploit Theta's distributed computing framework to run large-scale, cost-effective AI experiments.
Engaged on these developments, Kyeongpil Kang mentioned: It will assist us as our college students, researchers and employees proceed to develop over the subsequent few years. We sit up for leveraging Edgecloud's distributed GPU structure and seeing how DAMI Lab can speed up NLP and multimodal AI analysis and optimize mannequin efficiency for actual purposes. ”
Mitch Liu, co-founder and CEO of Theta Labs, added: By pushing over 85% utilization throughout your GPU community, it’s worthwhile to continuously scale your GPU infrastructure whereas maximizing cost-effectiveness. Our purpose for 2025 is to drive over 95% utilization and lock over 2 billion TFUEL by way of elite edge node operators. ”
learn extra: Constructing the next-generation creator economic system with AI brokers | Opinions
Disclosure: This content material is supplied by a 3rd social gathering. Crypto.Information doesn’t advocate the merchandise listed on this web page. Customers ought to conduct their very own analysis earlier than taking any motion associated to the corporate.
