April 28, 2021
By Rob Mitchum
After a year that many would like to forget, it’s understandable that the focus of Mind Bytes 2021 was on the future. But despite the challenges of 2020, the University of Chicago Research Computing Center (RCC) and the faculty, students, and staff it supports stayed active, constructing a powerful foundation for the post-pandemic world. The annual RCC research expo and symposium, held in a virtual format this year, showcased these new resources and multiple branches of UChicago research, while also looking farther into the future at quantum computing, AI, and other innovations.
An apt metaphor was chosen by keynote speaker Margaret Martonosi, Assistant Director, Directorate for Computer & Information Science & Engineering (CISE) at the National Science Foundation, who used the imagery of planting trees to describe her long-horizon view of research. A series of “tire track” diagrams showed how the technological conveniences of today — including the video conferencing software used for this very conference — were seeded by the basic science of years and decades before. Thus, the investments of the past year in new infrastructure, partnerships and people, on both the national scale of CISE and the local level of RCC, set the stage for a post-pandemic burst of discovery.
RCC director Hakizumwami Birali Runesha opened the day with a description of the new resources available to the UChicago community, headlined by the debut of Midway3. The next generation of the campus high-performance computing cluster features a new emphasis on architecture optimized for deep learning and artificial intelligence applications. With its 10,000 CPU and GPU cores, 200 nodes, and 2.3 petabytes of high-performance solid-state storage, the cluster will help the more than 6,000 users of RCC resources reach the next level of their research.
“Research today requires, in many cases, an array of computational skills that are not taught in the classroom and these skills are becoming more and more essential in conducting transformative research,” Runesha said. “To support discovery, the RCC enables researchers to pilot and explore the feasibility of higher-risk experimental projects, and work with software and hardware vendors to bring novel computing architecture to Chicago for early access and evaluation, and to facilitate analysis of data.”
Midway3 is also joined by additional new RCC offerings such as Skyway, which helps connect UChicago researchers with commercial cloud resources, new GIS services, and the Data Lifecycle Instrument (DaLI). The power of DaLI, an NSF funded project, was demonstrated in the first of four research lightning talks at the symposium, delivered by Callum Ross, Professor of Organismal Biology and Anatomy. By connecting the instrument to the XROMM high-speed X-ray facility, Ross and his laboratory have accelerated their anatomical studies of how animals such as lungfish and possums chew and swallow. A web portal connected to DaLi automatically archives the massive video data collected in each experiment, making it easily accessible for later analysis on the researchers’ local computers or through AI algorithms running on HPC resources.
Other lightning talks demonstrated the breadth of scientific questions that can be answered with modern computational tools. Anjali Adukia, Assistant Professor, Harris School of Public Policy and the College, discussed her project using computer vision to study gender and racial diversity in the illustrations and text of children’s books. Tiffany Shaw, Associate Professor in the Department of Geophysical Sciences, talked about the current state of global climate models and how scientists are refining their simulations to detect emergent signals of climate change.
James Evans, Professor of Sociology and Director of Knowledge Lab, tackled the even bigger question of how scientists and the computational “mind” of AI can work together most effectively in the future. His argument: that AI assistants can augment intelligence by offering an “alien,” outside-the-box perspective that inspires new promising and fruitful research directions.
“To generate robust insight and rapidly advanced technology, you need an ensemble of diverse intelligences,” Evans said. “AI can make a big difference, especially in the context of science. We need a new AI that really radically augments intelligence by understanding and programming past the limits of human reason.”
In her overview of CISE activities, Martonosi highlighted “the transcendence of AI” as one of the top NSF priorities, with $500 million invested annually across several National AI Research Institutes. Other focus areas include cyber-human interactions, the search for new hardware architectures in a post-Moore’s-Law landscape, and quantum computing — a topic of much interest at the University of Chicago, between the Enabling Practical-Scale Quantum Computing (EPiQC) project and the Chicago Quantum Exchange (CQE).
Appropriately, Mind Bytes closed with a discussion between the director of the Chicago Quantum Exchange, Liew Family Professor of Molecular Engineering David Awschalom, and Runesha about the potential of quantum technology. While quantum computers may still be years away from beginning to solve previously intractable scientific problems in drug design, energy, and finance, Awschalom predicted that these technologies will eventually be akin to film going from black and white into color in the implications for data and communication.
“One quantum bit, instead of just a zero or a one and two pieces of information, could be a billion pieces of information, or 100 billion, and that's just one atom or one electron. The volume of information you can hold is literally astronomical,” Awschalom said. “It also means that data can be entangled, you can share information independent of distances, everything becomes multiply connected all at the same time. It's a different way of thinking, and it will take time for us to embrace and exploit this new technology.”
Perhaps even more critical for this young field’s growth than scientific progress is the workforce demand — Awschalom estimated that one million quantum engineers will be needed in the next decade. From its home at the University of Chicago, the Chicago Quantum Exchange brings together academic institutions, national laboratories, and industry partners to help advance both research and education in the rapidly growing area.
“Chicago has emerged as one of the largest hubs of quantum science and engineering the nation, and will continue to grow," Awschalom said.
In the future, some of that quantum education and experimentation may take place at the RCC. For now, the availability of cutting-edge resources for high-performance and cloud computing, artificial intelligence, data science, and other computational approaches will continue to support the research of today and innovations of tomorrow.
“We rely a lot on the feedback of our faculty and the researchers as we refine our programs,” Runesha said. “RCC will continue to work with you to understand and support your research by providing access to high-end computing and advanced technical user support in order to enable you to undertake research and analysis that you will not otherwise be able to do.”