Theta Labs, Kangwon National University joined hands to push AI with Edgekloud

Thea Labs expands the AI ​​Research Partnership as Kangwon National University is included in the decentralized GPU network of Thea Edgekloud.

San Francisco, CA – March 5, 2025 – Thea Labs, a global leader in decentralized cloud infrastructure for AI, media and entertainment, announces his latest educational customer, Kangwon national university Data Analytics and Machine Intelligence (DAMI) LabUnder the leadership of Professor Kyongpil Kang,

Bringing Kangwone National University as Theta Edgekloud’s 23rd AI and video customer is an important milestone, collectively closed by more than 85% GPU usage rates in its worldwide network and more than 230 million Elite Booster nodes than TFUL over 230 million.

Theta Edgekloud is proud to give strength to many major enterprises such as NHL’s Las Vegas Nights and Global Asports Teams, with flycuven and Evil Genius, and top-tier genius, and top-tier educational institutions, Seoul National University, Cast, Oregon University, Michigan State University, and Singapore, and Singapore have played their role.

About Professor Kyonpil Kang and Dummy Lab

Professor Kyonpil Kangwon is an assistant professor in the Department of Computer Science and Engineering at Kangwon National University, specializing in Natural Language Processing (NLP), machine learning and data mining. He did his Ph.D. In Computer Science from Korea University, where he studied under Professor Jagul Chu, and later worked as a machine learning researcher in SK Telecom AI Center, Scatter Lab and Liner before joining Academia.

Data Analytics and Machine Intelligence (DAMI) lab at Kangwon National University has focused on a large scale AI research with specialization in natural language processing, multimodal AI and knowledge extraction.

Focus on the lab:

• Natural Language Processing and Large Language Model (LLMS): Research on LLM Fine-Tuning, Chen of Thoughts (COT), Retrieval-Ended Generation (RAG), and AI alignment. DAMI LAB LLM agent also examines performance adaptation and simulation-based learning.
• Machine learning and deep learning: addressing domain-specific AI challenges by creating training and developing deep teaching solutions for real-world applications.
• Data mining and AI-operated insights: Applying machine learning and LLM-operated analytics in areas such as social science and medical AIs, highlighting insight to deal with complex problems.

AI Research with Theta Edgekloud forward

Dummy Lab Natural Language Processing (NLP), to run innovation in machine learning and data mining, will avail the first hybrid cloud-edge GPU computing platform, Thea EdgeCloud. Through this collaboration, researchers at the DAMI lab will also get access to the decentralized GPU infrastructure of Theta, which will provide cost-affect, scalable AI computing power to carry forward LLM training, multi-agent AI system and multimodal learning applications.

• Increase NLP and LLM models: Scalable, train advanced language models with high-demonstration GPU computing.
• Develop a strong multi-agent AI system: Use with multi-agent debate models and adverse AI.
• Power AI at the Age: deploy the AI ​​model in a real-time environment using Thea’s hybrid cloud-ease architecture.
• Adapt the AI ​​training efficiency: Leverages the distributed computing structure of leverage thea to run cost-efficient, massive AI experiments.

Addressing these developments, Professor Kyonpil Kang says, “We have already been evaluating the ageclouds for several weeks, and we are deeply influenced by its flexibility and scalance. This will serve us well as our students, researchers and employees will continue to grow over the next few years. We are excited to take advantage of the distributed GPU architecture of Edgecloud and see how DAMI Lab NLP and Multimodal AI can fasten research by optimizing model performances for real -world applications. ,

Mich Liu, co-founder and CEO of Theta Labs, says, “It is exciting to bring the Kangwon National University on a ship as our 23rd customer, but is also in our development challenges. By extending the rate of more than 85% of usage in our GPU network, we need to continuously expand our GPU infrastructure, maximizing cost effectiveness. Our target in 2025 is to our customer base 10x, drive up more than 95% use rates and lock more than 2 billion tafuel through our elite edge node operators. ,