Theta EdgeCloud Powers Human-Centered AI Research at Soongsil University’s HUMANE Lab

We are proud to welcome the HUMANE Lab (Human-centered Language and Data Intelligence Lab) at Soongsil University to its growing network of academic research partners. Led by Professor Kunwoo Park within Soongsil’s College of AI, HUMANE Lab is leveraging Theta EdgeCloud’s decentralized GPU infrastructure to advance research at the frontier of large language models , with a mission rooted in solving real societal challenges.

Research That Tackles Misinformation and Bias

HUMANE Lab studies language and multimodal AI technologies built on large language models, with a clear human-centered focus: addressing societal problems such as misinformation, bias, and fairness through AI grounded in real-world data. Their work spans several active areas:

• Computational Social Science and Cultural Analytics
• Ethics, Bias, and Fairness
• LLM Agents
• NLP for Social Good
• Resources and Evaluation

HUMANE Lab’s first-place FEVER 2026 system verifies image–text claims in four stages — evidence retrieval, cross-modal analysis, question generation, and verdict prediction — coordinating multiple language and embedding models in parallel, exactly the kind of GPU-intensive workload Theta EdgeCloud is built to power.

This is research with measurable impact. The lab has built a remarkable track record on the world stage — most notably at the FEVER (Fact Extraction and VERification) workshop challenges, where their teams secured 1st place in 2026 (EACL), and back-to-back 2nd place finishes in 2025 (ACL) and 2024 (EMNLP). Their findings have been published at premier venues including ACL and EMNLP, often in collaboration with industry research groups.

Why Theta EdgeCloud

Modern fact-checking, multimodal verification, and LLM-agent research are compute-intensive — demanding scalable GPU resources for training, fine-tuning, and large-scale evaluation. Theta EdgeCloud delivers exactly that: distributed GPU power drawn from a global network of 30,000+ nodes, giving researchers cost-efficient, on-demand access to the horsepower their experiments require.

There’s also a natural philosophical fit. HUMANE Lab’s award-winning systems are built on open large language models — an approach that prizes transparency, reproducibility, and democratized access. That ethos aligns closely with Theta’s decentralized infrastructure, which distributes compute power rather than concentrating it.

Open models on open infrastructure: a partnership that makes sense from the ground up.

“Our research depends on access to substantial compute for training and evaluating large language models at scale. Theta EdgeCloud gives our team the flexible, cost-effective GPU resources we need to push our work forward — and its open, decentralized approach resonates with our own commitment to building transparent, human-centered AI.” — Professor Kunwoo Park, HUMANE Lab, Soongsil University

Building the Future of Trustworthy AI

As LLMs become embedded in how the world consumes information, research into verification, bias, and fairness has never mattered more. By providing the compute backbone for HUMANE Lab’s work, Theta EdgeCloud is helping accelerate the development of AI systems that are not just powerful, but trustworthy and human-centered.

We’re excited to support Professor Park and his team as they continue pushing the boundaries of what responsible AI can achieve.

If you are a university or research team looking for access to affordable compute via Theta EdgeCloud, reach us at partners@thetalabs.org

Theta EdgeCloud Powers Human-Centered AI Research at Soongsil University’s HUMANE Lab was originally published in Theta Network on Medium, where people are continuing the conversation by highlighting and responding to this story.