Grant Funds AI Visualization Collaboration with Argonne
Grant Funds AI Visualization Collaboration with Argonne
Modern scientific tools are generating massive datasets, faster than researchers can begin to examine them – making real-time analysis practically impossible.
Associate Research Professor Luc Renambot is working with scientists at Argonne National Laboratory to develop an AI-powered scientific assistant to wrangle large amounts of data and provide a seamless way to visualize the information.
Renambot is a member of the Electronic Visualization Laboratory (EVL) a lab within the computer science department that specializes in collaborative visualization, virtual reality, visual data science, and advanced computing and networking infrastructure. The lab was established in 1973 and is a pioneer in scientific visualization.
Renambot was awarded $600,000 over a five-year period from Argonne National Laboratory for the project, Scientific Data Management and Visualization to Advance AI for Science. He is collaborating with Victor A. Mateevitsi, a computer scientist at Argonne, who is a UIC alumnus and a CS department adjunct faculty member.
The AI tool can compress scientific datasets over 100-fold without losing fidelity and provides a way to explore the information through intuitive, adaptive, and interoperable interactions. A large language model assistant can understand a researcher’s intent and guide data exploration interactively.
This AI-powered visualization and context-aware data exploration is coupled with compact, computation-ready data representations, which will dramatically accelerate
scientific discovery by enabling real-time, human-in-the-loop analysis of large-scale simulation outputs and live experimental data.
A recent brainstorming session with Argonne examined some large-scale simulations of how the tool could be used for studying the cosmos. The universe is about 95% dark energy and dark matter, which dictate the expansion and structure formation history of the universe. Scientists can use these tools to compare the observed statistical distribution of structure to the theoretical predications through numerical simulations.
“This collaboration with Argonne National Laboratory represents an important opportunity to advance methodologies for scientific data management and visualization in support of AI for science,” Renambot said. “I look forward to working with Victor and ANL’s strong team of academic and industry collaborators on these challenges over the next few years.”