7/22/2024 Jenny Applequist
Written by Jenny Applequist
The two were identified by organizers as “exceptional” early-career researchers.
The National Academy of Engineering (NAE) has announced the names of 76 early-career researchers “performing exceptional research and technical work” whom it has invited to participate in the upcoming 2024 The Grainger Foundation Frontiers of Engineering Symposium, and two of the 76 hail from Illinois' Grainger College of Engineering: Phillip J. Ansell and Yuxiong Wang.
Ansell is an associate professor in Aerospace Engineering, and Wang is an assistant professor in Computer Science.
The annual symposium—which is “a signature activity” of the NAE, according to the NAE’s website—has been held since 1995. It brings together a group of outstanding early-career engineers for intensive exploration of a set of key themes, which change each year.
This year’s event, to be held in September, will revolve around the themes of water-air-surface connections for indoor microbiology and health; the future of AI; connections between the gut and brain; and the impending digital twin revolution.
Ansell said that the symposium’s digital twin theme is the one that relates strongly to his own work.
“My principal research area these days is in the broad category of sustainable aviation,” he said. “‘Digital twin’ is a bit of a buzzword for model-based approaches to representing complex physical systems. This is one of the ways in which we have contributed towards envisioning designs of future zero-emissions aircraft configurations, for example.”
Ansell’s team has had great success in using models to investigate the use of hydrogen energy sources for aircraft. “We’ve been able to use these digital approaches to really understand where the sweet spot is in hydrogen aircraft design, and how this sustainable energy system requires a radical rethinking of how aircraft are configured,” he said. “That’s one of the things that we’ve been able to do with some of these model-based tools: we can create digital versions of these aircraft which fly virtual missions, so to speak.”
Wang is working on AI, primarily for computer vision and machine learning. Specifically, he’s tackling the so-called “open world” challenge of improving AI’s ability to understand novel things it encounters.
Existing AI perception systems can recognize objects they already know—operating in a “closed world”—but cannot easily learn new things. Wang said that a child can see a single image of an unfamiliar animal and instantly generalize that impression to recognize the same type of animal elsewhere—but today, a machine must be trained on thousands of images before it too can recognize a new example of something.
“So what I’m trying to do is basically enabling AI agents to achieve this kind of human-level generalization, so that they’re able to recognize unseen objects, understand new environments, and learn novel skills from very limited data,” Wang explained.
Small-data AI is important because critical information is often limited; for example, healthcare data are typically scarce. To overcome such challenges, Wang’s team has recently made significant progress in advancing generative AI capable of imagining things it has never observed at all, effectively serving as world models that empower autonomous agents to perceive, interact and act in the open world.
Among his various other small-data AI successes in fields like agriculture and materials science, Wang has developed an AI system that can detect early stages of Parkinson’s disease based on characteristics like facial expressions, movements of key points on the face, and speech.
Ansell described his selection for the NAE symposium as “a tremendous opportunity” and said he’s “thinking that there’s going to be some really, really cool people to engage with there.” He added that he’s appreciative of the impactful work the NAE does and that he’s “very excited” that he might be able to start playing a role in it.
Wang expressed similar enthusiasm about the symposium. “I’m very excited to leverage this opportunity to enrich my understanding of engineering and its intersections with the broader world,” he said. “To me, the lack of data is a pervasive challenge in nearly every field of science and engineering, and it has increasingly become a bottleneck for many practical applications. The hope is really to establish a groundbreaking interdisciplinary collaboration to develop a unified open-world, small-data AI framework that scales up across various application domains for broader societal impact.”