Xia Wang
Ohio State University, OH·Computer Science
Academic supercomputing centers serve tens of thousands of scientists and engineers, and their small user-support teams field a constant stream of technical questions. Answers depend on scattered information: past support tickets, documentation, the live status of the machines, and specialized expertise.
- Fields
Computer Science
Engineering
- Topics
high performance computing
generative AI
large language models
- Grants
- 1 known grant, through Jul 2027
What they work on
- Build hpcGPT, a question-answering service for academic high-performance computing centers, by combining large language model fine-tuning with retrieval augmented generation so answers reflect center-specific documentation and past support experience.
- Integrate heterogeneous data sources that update at different frequencies, including historical support knowledge, documentation, external technical expertise, and real-time system status of versatile computing platforms.
- Reduce the effort needed to align questions, answers, and supporting information and to identify dependencies among them, so centers with limited staff can deploy the service.
- Evaluate whether hpcGPT improves support precision and response time for HPC users at partner centers such as the Ohio Supercomputer Center and the Texas Advanced Computing Center.
More about the research
The question driving this project is whether a generative artificial intelligence assistant can pull those sources together and answer accurately enough to speed up support without adding staff.
Why it matters
Researchers who rely on national supercomputers often lose time waiting for help with technical problems. A domain-tuned question-answering service could resolve many issues immediately, improving support quality and capacity at academic computing centers without additional staff and freeing support teams for novel issues. Centers with small staffs and diverse applications may benefit most from the reusable components the project releases.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Chi Jin
Princeton University, NJ·Computer Science·Early-career
Identify the general structures of function classes, such as those represented by neural networks, under which model-free reinforcement learning is tractable, and design provably efficient algorithms that also handle model misspecification.
Leif Nelson
Boise State University, ID·Computer Science
Deploy and integrate six compute nodes, each with four NVIDIA L40S GPUs, into the existing campus research computing cluster to support data science, machine learning, and AI workloads.
Neranjan Edirisinghe Pathirannehe
Georgia State University, GA·Computer Science
Deploy a campus computing cluster with 12 NVIDIA H100 GPUs, BlueField data processing units, 682 CPU cores and 228 TB of storage to support 11 research projects spanning wildfire management, solar prediction, neuroscience, finance and policy.
The school
Ohio State University: 71 labs on PI Pages, $1.6B a year on research, 882 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.