All posts by: Catherine Meyers

Building next-gen AI chips at UMBC: A Q&A with NSF CAREER award winner Chenchen Liu

Many recent artificial intelligence (AI) breakthroughs—such as smartphone tools that recognize your friends’ faces or understand your spoken commands—are based on a computing approach that was first dreamed up nearly 80 years ago. Called neural networks, the approach loosely mimics the way biological brains work. For decades, it was a quixotic idea with results that fell far short of the power of human brains. Yet starting in the early 2000s, the technique took off. What changed? In short, computers finally got fast enough (and training data plentiful enough) for neural networks to realize their hoped-for potential.

Woman in white blouse and jacket, who studies AI chips — Chenchen Liu (Marlayna Demond ’11/UMBC)

The story of neural networks illustrates a key principle in computer science: the power of any computing technique is bound by the capabilities of the hardware that runs it. Improving that hardware is a major research focus for Chenchen Liu, an assistant professor of computer science and electrical engineering at UMBC. Liu was recently awarded a prestigious NSF CAREER award totaling nearly $540,000 that will fund her efforts over the next five years to advance the next generation of powerful computer chips.

As society continues to engage with the implications of the latest commercial iterations of AI, research efforts such as Liu’s are looking ahead to an even newer wave of applications, including self-driving cars, immersive virtual reality, and AI-assisted agriculture. These new applications often require separate neural networks—for example, an image recognition system and a collision avoidance system, to share information and work together.

Working with students and colleagues, Liu is studying how modifications to state-of-the-art AI chips could make such coordination easier.

UMBC News asked Liu about her research, the NSF CAREER award, the future of AI, and the need for responsible AI use.

UMBC News: What are some of the main challenges holding back next-generation AI applications?

Liu: In next-gen AI applications, machines are faced with a complex composition of multiple tasks, requiring them to run multiple AI models simultaneously. And not only do the models need to run at the same time, they need to share information from one to the other. The computing needs of the models may be different. Some models may need a lot of computing power, others may need a lot of memory. It becomes very difficult to coordinate these distinct resource needs on a single computing platform and to run everything in parallel and as efficiently as possible.

UMBC News: What are some ways that you plan to tackle these challenges in your NSF CAREER-funded research?

Liu: We first need to map the interactions of the different models and understand what computing resources they need and how they work together. We then plan to investigate a novel computer chip architecture designed to support these interactions. We will look for ways to flexibly schedule tasks and allocate computing power and memory so the system can operate differently for different scenarios. Eventually, we will integrate these techniques into a comprehensive framework that could be widely applied to various next-gen AI applications.

UMBC News: What are you the most excited about in your work?

Liu: I believe an incredible increase in AI complexity will be the next big thing in the world, and I am excited to become a small part of it. The increase in complexity comes with an urgent need for novel computer architecture to support it. Since the first computer was invented, people have been trying to optimize computing performance. Now with AI, current computing resources cannot meet the requirements caused by larger and larger volumes of data. A lot of researchers are working on novel computing architectures to improve performance and I am happy to be one of them.

UMBC News: Do you see a lot of opportunities for students in the field?

Liu: AI is changing everything in the world now. You can hear about job opportunities in Silicon Valley or with new AI start-ups nowadays, indicating a thriving new era of careers centered on AI computing. This trend is also shaping the interests of students on campus. They want to work on projects with AI components and we offer opportunities like that in many of our classes. It’s not only an opportunity for students, but also an opportunity for us to review our course development and career guidance.

UMBC News: What do you think is the future of AI?

Liu: The overall feeling is one of awe. Just like the first word carved on stone and the first rocket launched to the sky, AI is another milestone in human history. It marks humankind’s exploration of intelligence itself, asking questions about how it is formed and how it might evolve.

We should proceed cautiously, since AI is also strong enough now to challenge human intelligence and even deceive us. Many governments have made calls to regulate it.

I believe in a few years, we will have every perspective of our world reshaped by AI, with clear impacts on the economy and society.

As summer wildfire smoke choked Baltimore, UMBC air pollution researchers leapt into action

Starting this May, a series of wildfires in Eastern Canada sent enormous smoke clouds wafting into the U.S., triggering air quality warnings in cities from the Midwest to the Northeast. For days, orange skies backdropped landscapes clouded by acrid air. People who could hunkered inside with the doors and windows shut. Those who had to go out faced itchy eyes, burning throats, and worse.

As a resident of the Baltimore area—which was blanketed with particularly bad smoke in both early and late June—UMBC Professor Chris Hennigan looked at the haze with dismay. But as an environmental engineer who studies air pollution, he had an additional thought: “We were looking at the air quality forecasts, and we thought ‘We have to gather data,’” he says.

The public found many colorful words to describe the summer’s unwanted smoke: brutal, eerie, dystopian.

Hennigan and his team have been working to put numbers to the adjectives. On the roof of the engineering building, the researchers installed a squat, white sensor that monitors the levels of tiny particles in the air, particularly those measuring 2.5 micrometers in diameter or less—smaller than most bacteria. Called PM_2.5, these particles are released in large numbers during fires. They are dangerous to human health because they can work their way into the deepest parts of the lungs and even enter the bloodstream.

Three people stand on a roof next to equipment. Trees in distance. — Chris Hennigan, Joel Tyson, Ph.D. ’23, and Luis Rodriguez ’25 (left to right) on the roof of the engineering building next to an air quality sensor. (Marlayna Demond ’11/UMBC)

The sensor showed huge spikes in PM_2.5 when the smoke blew through, on some days reaching levels considered unhealthy for anyone to breathe.

The researchers also set up equipment to filter particles out of the air. After 24 hours, they collected the filters, which they are storing, neatly labeled, in a refrigerator in Hennigan’s lab.

A gloved hand holds a sample dish with dark contents. Another sample dish is white. — Hennigan shows samples of smoke particles collected this summer. (Marlayna Demond ’11/UMBC)

The filtered samples will advance at least two ongoing investigations, Hennigan says. In one avenue of inquiry, Joel Tyson, Ph.D. ’23, biochemical engineering, is studying how tiny particles can harm human lung cells. Before this year’s smoky summer, Tyson had been studying the toxic effects of particulate matter normally found in the Baltimore air. With the new smoke samples, he will start to investigate whether wildfire smoke particles, per unit, are more toxic than regular urban particulate matter, which comes from sources such as cars and power plants. Some studies have indicated that wildfire particulate matter is indeed more toxic, but more research is needed before any definitive conclusions can be reached.

In another line of research, Hennigan is also studying how particles in the air, including from smoke, may affect the climate. Undergraduate chemical engineering students Danielle Larios ’25 and Luis Rodriguez ’25 are assisting in the investigations.

The researchers study how particles of brown-colored carbon-containing material absorb light. Burning vegetation sends large amounts of this brown carbon into the atmosphere. It’s possible that the particles are trapping significant heat from the sun, accelerating the pace of planetary warming. Such effects are not normally included in global climate models, and better understanding of the process could improve humanity’s ability to predict, and manage, the coming years of climate upheaval.

Three people wearing gloves and lab coats talk in a laboratory. — Rodriguez, Danielle Larios ’25 and Hennigan (left to right) discuss research in the lab. (Marlayna Demond ’11/UMBC)

Two people in the lab look at liquid in a container. — Rodriguez, Danielle Larios ’25 and Hennigan (left to right) discuss research in the lab. (Marlayna Demond ’11/UMBC)

Climate change and wildfires are intimately linked. This summer was not only smoky, but also scorching. July marked the hottest month ever recorded, and scientists predict that as the world continues to warm, wildfires will continue to increase in quantity and intensity. “Smokeageddon,” as headlines put it, may become the new normal.

Hennigan says recent research illuminates how much wildfire smoke has contributed to air pollution trends. He points to a paper published in September in the scientific journal Nature that estimated that since 2016, wildfire smoke in the contiguous United States has undone around 25% of the progress in air quality made between 2000 and 2016.

For the researchers in Hennigan’s lab, those effects have been felt personally.

Rodriguez recalled how in June he had to go out to buy a fresh pack of N95 masks. “The smoke was just awful,” he says. Larios says she felt a burning at the back of her throat in just 15 minutes walking to her car.

For Tyson, the effects of the smoke were so bad that at one point he struggled to breathe and had to visit the doctor. The episode, he says, drove home the importance of his toxicology research.

All three note both the complexity of the systems they are studying and the importance of discovering new knowledge that might help society handle the environmental challenges it faces.

“Our work can have real-world impact, and that’s exciting,” says Larios.

From apples to army robots, curiosity and commitment define Priya Narayanan’s career

She didn’t exactly experience a Sir Isaac Newton-like epiphany after being conked by a falling apple, but Priya Narayanan, Ph.D. ’08, mechanical engineering, spent a lot of her time at UMBC interacting with the iconic red fruit.

For her Ph.D. thesis, Narayanan worked with the U.S. Department of Agriculture to study whether a simple device made of a long, inclined track could reliably orient apples. The ultimate goal was to automate visual inspection of the fruit—using cameras to spot blemishes—and the cameras required the same view of the apple each time.

Narayanan spent thousands of hours performing experiments with rolling fruit, filming high-speed videos of its descent down the track, and developing mathematical models of the apple’s motion to explore why the fruit ended up in its final orientation at the bottom of the track.

“I was able to shape the direction of the research and got interested in modeling the contact mechanics of the apples,” Narayanan says. “I thought it might be easy, but it turned out to be very hard.”

Rouben Rostamian, a professor in the Department of Mathematics and Statistics at UMBC who worked with Narayanan on the modeling, commended her persistence.

“Priya was interested in studying the problem deeply,” he says. “She learned new areas of math so that she could really pursue her questions.”

Developing her skill set

Headshot courtesy of Priya Narayanan.

Curiosity and commitment to learning have been a constant of Narayanan’s approach to problem solving—even as the systems she studies and the roles she takes on have shifted over her career. Narayanan was recently named chief of the autonomous systems branch at the U.S. Army DEVCOM Army Research Laboratory (ARL). She leads teams working to develop intelligent robots to assist the U.S. military.

“I’m not studying my apples anymore,” she says. “But in graduate school, I broadened my skill set. That has really helped me throughout my career.”

After graduating from UMBC, Narayanan worked for a few years at the University of Maryland School of Medicine in Baltimore. She studied robotic systems that could assist patients recovering from strokes. The work gave her the opportunity to develop expertise in computer vision and image processing as well as gain experience making prototypes for robotic research and conducting clinical trials.

In 2014, Narayanan was awarded a National Research Council fellowship to work at the Naval Research Laboratory.

“It was around this time that deep learning started making waves,” she says. She shifted her focus to building artificial intelligence (AI) models. When her fellowship ended after three years, Narayanan joined the ARL, headquartered in Adelphi, Maryland, and has been there ever since.

She has ascended the ranks at the organization, moving from a research engineering position, to leader of a small team, to ultimately head of an entire branch.

Overlapping work with UMBC

Researchers under Narayanan’s leadership conduct research in robotics to help aerial, wheeled, and legged robots autonomously navigate rugged landscapes and difficult terrains, and coordinate their actions with other robots and humans.

As part of her role, Narayanan has worked with UMBC researchers affiliated with the Center for Real-time Distributed Sensing and Autonomy. The center, which is led by information systems professors Aryya Gangopadhyay and Nirmalya Roy, coordinates with ARL on research to develop AI-enabled smart robots.

A woman and two men look at a laptop screen in contemplation — Narayanan with fellow graduate students Raghavendra Angara and Jingrui Wang in the robotics lab at UMBC. (Photo courtesy of Narayanan)

Narayanan says she found her way into management by embracing leadership opportunities that came her way, such as leading an ARL seedling program that encouraged researchers to form teams and submit proposals for new high-risk, high-payoff projects. “I’ve also attended an entrepreneurship and leadership program organized by ARL in collaboration with UMBC Training Centers, so I’m still learning from the university,” she says.

She honed soft skills, such as communication and networking, that have been critical to her success. She also says her broad range of research experience has been useful in her leadership positions, helping her connect with team members who have varying areas of expertise.

an indian woman in decorate make up and outfit poses — Narayanan at a performance at UMBC. (Photo courtesy of Narayanan)

Looking back at her days at UMBC, Narayanan remembers a wonderful community and the feeling of being supported. Not long after she arrived on campus, she and a few fellow students found themselves in the same elevator as then President Freeman Hrabowski. He introduced himself and asked all the students their names and what they were studying. Later, when Narayanan and a friend performed a classical dance on campus in celebration of the Indian holiday Diwali, she remembers how President Hrabowski came backstage to complement their performance.

She encourages current students to take advantage of all that the university and the region have to offer: “Realize how many resources you have at your disposal, and really make use of them,” she says.

Uri Tasch, her Ph.D. advisor, who is now an emeritus professor of mechanical engineering, recalls how Narayanan showed up nearly every day at the lab. “You couldn’t stop her,” he said. “She did beautiful work, and she always had a smile on her face.”

UMBC’s Vandana Janeja aims to boost high-performance computing know-how to tackle environmental science challenges with a $1 million NSF grant

The discovery of the Higgs boson. The first picture of a black hole. The Covid-19 vaccine. Many recent scientific advances, such as these, owe much to a largely unsung hero: high-performance computing.

Woman smile at camera — Vandana Janeja (Marlayna Demond ’11/UMBC)

Yet the use of high-performance computing is generally limited to a niche group. “Some researchers are very sophisticated in the use of these tools, but many are not,” says Vandana Janeja, a professor of information systems at UMBC.

Janeja is on a mission to spread advanced computing know-how far beyond its current borders. “It’s not just for the elites,” she says.

She was recently awarded a nearly $1 million grant from the National Science Foundation (NSF) that will help further that mission. The grant is part of a larger NSF collaborative award with the University of Maryland Center for Environmental Sciences (UMCES).

Janeja and her UMBC and UMCES colleagues will work to connect a wide range of students and faculty with high-performance computing experts, creating and nurturing what has been called the cyberinfrastructure pipeline. The ultimate goal is to facilitate the flow of knowledge.

“Once one person gains experience, they can turn around and help someone else,” Janeja says. “It’s about supporting and connecting people with resources and not just about setting up the hardware and software.”

Building a high-performance computing workforce

The need for fast computers in research is driven in part by a modern deluge of data. For example, NASA estimates that it will soon have accumulated hundreds of petabytes of Earth science data, from sources such as satellites flying overhead and sensors installed on the ground. Making full use of all that data requires new approaches to computing. While traditional computers generally perform calculations one at a time, high-performance computing facilities break up the work between multiple computing nodes and run processes in parallel.

UMBC has its own high-performance computing facility, which houses computing equipment that can process data significantly faster than a standard laptop. University researchers have employed the fast computers to study topics as diverse as weather modeling, cancer treatment, and flight dynamics.

“Learning how to use high-performance computing is not straight-forward,” says Janeja. Users have to gain a good understanding of the infrastructure of the computers, and also hone their skills in identifying the best techniques to take full advantage of that infrastructure.

People with a good understanding of how these machines and their software work will be in high demand, Janeja adds.

Many modern AI algorithms, which are used by scientific researchers and big tech companies alike, require high-performance computing to operate. “While everyone is running toward the shiny object of AI, many don’t realize that there is a backbone, called cyberinfrastructure, that runs AI,” says Janeja.

Part of Janeja’s goal with the new grant is to educate students who may pursue careers building, maintaining, utilizing, and improving that cyberinfrastructure backbone. She also cares deeply about ensuring diverse groups of people have access to these roles.

“We are trying to democratize the use of cyberinfrastructure and make it accessible to people who have never used it,” Janeja says.

Jack Suess, ‘81, M.S. ‘94, UMBC’s vice president of information technology, says he is excited for his division to play a key role in the award, helping shape best practices and build a national cyberinfrastructure support structure. “I look forward to watching how this work informs not just how we support faculty, but how we train and support advanced undergraduates and graduate students to evolve into these roles,” he says.

Tackling environmental challenges

While there is no shortage of scientific questions that might be tackled with high-performance computing, Janeja and her colleagues on the new grant will test-run their collaborative initiatives with projects in environmental science.

Aerial view of rugged landscape with snow and ice. — Changes to the ice and snow in Greenland could contribute to sea level rise. (Image credit: NASA)

Janeja is also director of iHARP, an NSF-funded institute that aims to harness big data and advanced computing tools to better model how climate change will affect the polar regions. As part of the new grant, iHARP and UMCES researchers will partner with cyberinfrastructure professionals to explore environmental science questions, such as in iHARP investigating how to predict the rate of snow melting in Greenland.

Scientists increasingly rely on big data and high-performance computing to understand and predict changes in the environment. As human activities continue to put pressure on natural systems, it will be more important than ever to forge collaborations between researchers and computing experts that advance the science.

“I’m excited to leverage high-performance computing to solve real-world problems and make it accessible to students and researchers of all backgrounds,” Janeja says.

UMBC joins collaboration to create new STEM education, research center at Arecibo Observatory site in Puerto Rico

This week UMBC was named as one of four institutions chosen to work together on a National Science Foundation-funded project to establish a new science educational center at the Arecibo Observatory site in Puerto Rico. NSF will contribute more than $5 million over five years to establish the multidisciplinary center, called the Arecibo Center for Culturally Relevant and Inclusive Science Education, Computational Skills, and Community Engagement (Arecibo C3).

The center is expected to open in early 2024 and will include a research laboratory and a hands-on, interactive science center open to the public.

Patricia Ordóñez head shot — Patricia Ordóñez (Photo courtesy of Patricia Ordóñez)

Patricia Ordóñez, M.S. ’10, Ph.D. ’12, computer science, an associate professor in information systems, will lead UMBC’s contributions to the project. As part of the Arecibo C3 collaboration, Ordóñez is organizing a Women in Data Science Puerto Rico conference, as well as several workshops in data science that will be held both online and in person. One of the workshops will prepare attendees to participate in the 2024 Women in Data Science Worldwide datathon. Other workshops will recruit mothers and daughters to learn coding together, and will teach new skills to women and other participants who already code.

Several UMBC graduate and undergraduate students and staff members have been involved in developing and testing the workshop materials.

“My passion has always been to increase the number of Latinas in computing,” says Ordóñez, who, before joining the faculty of UMBC, was a member of the computer science department at the University of Puerto Rico-Río Piedras. “There are too few of us in the field—creating inclusive programs from Arecibo C3 will help us bridge that gap.”

The other grant awardees are Cold Spring Harbor Laboratory, University of Puerto Rico-Río Piedras, and the Universidad del Sagrado Corazón. In addition to data science programming, Arecibo C3 will offer a suite of activities, including explorations of biodiversity and how to identify species using short segments of DNA. The Arecibo C3 team will also advance research to explore how STEM teaching can be enhanced through the presentation of data in audio, tactile, and other sensory forms, to increase accessibility.

“The new educational center builds on the great scientific, educational, and cultural legacy of the Arecibo Observatory and is closely aligned with NSF’s goal to create STEM opportunities everywhere,” said James L. Moore III, NSF assistant director for STEM Education, in a press release. “The center aims to create new opportunities for STEM education, exploration, discovery, engagement and participation of students, scientists and researchers in various STEM disciplines ranging from astronomy and radio science to biological, computer and natural sciences in Puerto Rico and beyond.”

Ordóñez credits the talents of the whole Arecibo C3 team with turning what feels like a dream of hers into a reality. “We recognize the responsibility of what we are doing and we are going to work very hard to create the environment of inclusive excellence in Arecibo that everyone deserves, so that everyone sees they have a place in STEM,” she says.

For updates on the project, please visit the Arecibo C3 website.

Students discover the beauty of mold and mentorship in Mark Marten’s UMBC lab

Mold on your bread or bathroom tiles can be a nuisance. Mold in a scientific lab can be a marvel.

Up close, the growth of mold becomes living artwork—white, feathery shoots morphing into undulating waves of color. And molds can be amazingly useful.

“They are used to ferment food, make laundry detergent enzymes, and help produce pharmaceuticals,” explains Garrett Hill ’24, biochemistry and molecular biology, who has been working with molds in the research lab of UMBC chemical engineering professor Mark Marten for more than four years. “It’s surprising how ubiquitous they are in industry.”

Mold under the microscope shows thin filaments and round dots. — Mold under the microscope. (Image by Kathie Hodge via Flickr. CC BY-NC-SA 2.0.)

Molds (along with mushrooms) belong to a group of organisms more technically known as filamentous fungi. One of the oldest and largest living organisms in the world is a filamentous fungus, nicknamed the “Humongous Fungus,” that has likely been spreading across the Blue Mountains in eastern Oregon for more than 2,000 years, and has come to occupy an area of more than 2,000 acres.

The Humongous Fungus grew (and grew) through an enormous network of interconnected thread-like structures called hyphae that gather and share vital nutrients.

Networks—of a different sort—are also vitally important to the students studying fungi in Marten’s lab. Lab members, from high schoolers to Ph.D. students, work together on projects. Marten offers advice not only on research questions, but also on skills such as communication. Support flows in from the university, in the form of research awards, scholar programs, and more. It’s tied together with a simple philosophy that helps everyone flourish: “Mentorship is the magic ingredient,” Marten says.

High school research leads to UMBC

Hill found his way to Marten’s lab through a program in nearby Howard County Public Schools called the Biotechnology Career Academy. As part of the program, he earned credit for conducting research in the lab. After high school graduation, he enrolled at UMBC and has continued his work in Marten’s lab every year since.

“UMBC has a culture that emphasizes supporting students,” says Hill. “This was something that initially attracted me to the school, and something that I’ve absolutely experienced during my time here.”

In Marten’s lab, Hill has been working on research that investigates how a fungus called Aspergillus nidulans repairs damaged cell walls. Armed with a better understanding of the complicated cascade of biochemical reactions triggered when the fungal cell wall is damaged, scientists could possibly manipulate the process to use molds more effectively (or in the case of harmful molds, eradicate them more effectively.)

Man in lab coat holds petri dish with mold. — Mark Marten (left) and Garrett Hill examine mold cultures. (Marlayna Demond ’11/UMBC)

To untangle the hidden and complicated inner workings of the fungus, the researchers deploy an arsenal of analytical tools and methods.

“When I started in the lab, I spent a lot of time learning the background and standard lab techniques,” Hill says. “But now that I’ve had a few years to acclimate, I have the foundation to support my own project.”

As part of that project, Hill has been investigating how to use the Nobel Prize-winning gene editing tool called CRISPR to create fungal cells with inner elements that light up. The light provides a beacon for researchers to track how those elements move as the cell experiences stress or initiates repairs.

“The sense of project ownership has been one of the most rewarding parts of my research experience,” Hill says. “When my graduate mentor and I first discussed the possibility of leading my own project, I had a brief moment of doubt. But I chose to not listen to that voice.”

Networks nourish growth

Hill says he was attracted to the beauty and power of science from a young age, even though no one in his family had a scientific career. UMBC has provided the resources and counsel to help him chart his path.

Within Marten’s lab, Hill says more experienced researchers were always willing to help. When Hill first joined as a high schooler, fellow lab member Ryland Spence ’19, biological sciences, who is now a medical student at Brown University, trained him on techniques.

“Mentorship has been so important for me, so I am always happy to provide mentorship to others whenever I get the chance,” Spence says. “A supportive environment that values diversity is very much a part of UMBC.”

Marten himself also provided enormous guidance and support. “Dr. Marten has been the strongest mentor I’ve had, helping me even before I came to UMBC,” Hill says. “He’s taught me not only about fungus, but also about how to think like a researcher, how to present research, and how to be a good student.”

University-wide programs provided Hill with additional research support. He was awarded multiple Undergraduate Research Awards (URAs), which provide financial assistance for research projects and opportunities to practice skills such as presenting research.

Hill is also part of the nationally renowned Meyerhoff Scholar Program, which seeks to increase diversity among future leaders in science, technology, engineering, and mathematics by supporting students who intend to pursue a Ph.D. or combined M.D./Ph.D. in STEM and are interested in the advancement of minorities in those fields.

“A lot of my best friends are from the program,” says Hill. “It’s been great to have that community.”

Now that he is applying to graduate schools, Hill says the Meyerhoff program provides detailed guidance through the process. “They have been an invaluable resource, and a rock really. They make sure I know what I need to do now to ensure I have opportunities in the future.”

A closing chapter and a new beginning

As the fall semester of his senior year kicks off, Hill says that he’s feeling confident and excited.

“I find myself frequently looking back on how I felt on my first day in the lab, during my first lab meeting presentation, or during my first day of classes, and realizing how much I’ve grown these past few years,” he says. “What once used to really shake me, I am now able to do with confidence—that tells me a lot about what I’ve gained from my research experience.”

As Hill gears up for grad school, he is passing the baton to other UMBC students like Matthew Quintanilla ’27, chemical engineering, a first-year student whose journey shares many similarities with Hill’s. The first in his family to pursue a scientific career, Quintanilla also started work in Marten’s lab as a high-schooler and decided to enroll at UMBC. In Marten’s lab, Quintanilla is working with Ph.D. student Alex Doan, who attended the same high school and embraced the opportunity to mentor fellow students.

“It’s been great catching up on high school news, but more importantly helping students grow,” says Doan.

A student in a lab coat holds a glass rod in the flames of a Bunsen burner — Matthew Quintanilla works in the lab. (Marlayna Demond ’11/UMBC)

Quintanilla is also a URA-recipient and Meyerhoff Scholar, and says he is excited for the new school year.

“I am eager to start my academic career at UMBC, meet many others, and integrate my knowledge from courses into my lab work,” he says.

“Matthew and Garrett are both really talented individuals,” says Marten. “Having them in the lab has been a win-win situation.”

April Householder ’95, the director of undergraduate research and prestigious scholarships at UMBC, looks at Marten’s lab as a microcosm of the vibrant UMBC research environment. “These two students—Garrett and Matthew—represent two ends of the research spectrum. One is just getting started, and the other is a four-time URA Scholar,” she says.

“The support these students receive from the mentorship in their lab is invaluable, but also as important is the peer-to-peer support they will get from one another. It’s this type of academic community building that gets student researchers excited about being a part of UMBC.”

Cybersecurity expert Richard Forno appointed an honorary international professor

Richard Forno, assistant director of UMBC’s Center for Cybersecurity and a principal lecturer in the department of computer science and electrical engineering, has been appointed an honorary international professor in the School of Science and Engineering at the Universidad Autónoma del Estado de Hidalgo (UAEH), one of Mexico’s oldest universities.

Forno joins a select group of academics who have been named honorary international professors at UAEH. The recipients, some of whom have received some of the highest honors in their fields, are chosen for their overall academic and professional contributions.

The investiture ceremony coincided with the university’s two-week-long international book fair, held in late August and organized this year around the theme of cybersecurity. Over two days, Forno delivered five talks, met students and faculty, and spoke with university leaders about areas of possible collaboration between UMBC and UAEH.

A man in a suit stands at a podium. — Richard Forno delivers a lecture on cybersecurity. (Photo courtesy of UAEH)

Countries around the world share the challenge of maintaining cybersecurity in an interconnected and ever-evolving digital landscape. In his main lecture, Forno emphasized the importance of education and interdisciplinary collaboration to combat cybersecurity threats.

Forno says the formal investiture ceremony, held in one of Hidalgo’s oldest structures, was a particularly meaningful event for him. “It was a humbling personal experience and certainly a professional honor that I will not soon forget,” he says.

Harnessing AI to improve healthcare: Sanjay Purushotham wins $590,000+ NSF CAREER award

Today, a plethora of technologies, from image recognition tools to chatbots, are powered by machine learning. A key component to the technique’s success is data—and lots of it. For doctors and hospitals who hope to use machine learning to improve healthcare, that need for copious data presents a problem: medical data is protected by privacy laws and often exists in incomplete or diverse forms—from doctor’s notes to medical scans—that make it difficult for machine learning models to use it effectively.

“Using data better to aid medical decisions is a grand challenge of the 21st century,” says Sanjay Purushotham, an assistant professor in information systems at UMBC. “We need innovations in existing techniques to take full advantage of artificial intelligence in healthcare.”

Together with his students, Purushotham is tackling that challenge. He recently received a prestigious NSF CAREER award to support his team’s efforts to develop new ways to train health-focused machine learning models.

Purushotham has been contributing his expertise in computer science to collaborations with doctors and hospitals for almost ten years. The NSF CAREER award will help Purushotham further that research in a new direction, and ultimately, his team hopes their work will improve medical treatments and reduce costs, benefiting patients around the world.

The need for data-driven healthcare

Portrait photo of man in suit jacket in front of a brick building. — Sanjay Purushotham (Marlayna Demond ’11/UMBC)

When you visit the doctor, you may be asked about your general health, your family medical history, and your current symptoms. The doctor may order a series of tests and scans. All of this data could yield valuable insights into your health and the best course of medical treatment.

Bringing the data together in ways that make it accessible and illuminating is an ongoing challenge in health care. Synthesizing and presenting health information in new ways could improve individual medical care. Computer systems could ensure busy doctors see the most relevant information at the right time, aiding their decision making. It could also prove valuable for addressing public health challenges, such as the spread of infectious diseases.

“During the COVID pandemic, the health community realized the importance of having access to good data,” says Purushotham. Better data could help public officials make better decisions about when to take precautions, such as closing schools. It could also help individuals better understand the risks when choosing to partake in activities like attending large social gatherings or traveling.

Over the next five years, the award will support Purushotham and his students as they investigate ways to advance a technique called federated learning, which could allow hospitals and doctors to jointly build and evaluate a machine learning model without sharing sensitive medical data.

A new way to collaborate

Most machine learning models are trained centrally, on vast quantities of data that are often scrapped from the internet. However, a lot of data exists that cannot be sent to a central processing facility, for cost or privacy reasons. The term ‘federated learning’ was coined in 2016 to describe a decentralized machine learning technique that could train on data that never leaves users’ mobile phones. Instead, individual devices would take turns downloading the model, training it on their own data, and then sending the updated model parameters back to a central location.

One person sits in front of a computer and microphone, while two standing people look at the screen. — Students Sultan Ahmed (left) and Zahid Hassan Tushar (center) work in the lab with Sanjay Purushotham (right). (Marlayna Demond ’11/UMBC)

The general approach could be beneficial in many industries, from finance to manufacturing. In health care, federated learning has clear appeal because privacy laws strictly limit how health data can be shared.

For example, federated learning could allow hospitals to train a joint model on their medical data, such as CT scans, without sharing the scans with each other. The jointly trained model might reveal new ways to detect disease.

Yet, many hurdles remain to successfully deploying the technique, including the non-uniformity and limited nature of some patient data, the varying computational resources available to different medical practitioners, and the threat of bad actors seeking access to private information.

Advancing federated learning

In the coming years, Purushotham and his team will pursue three main avenues of research to address the obstacles of deploying federated learning in healthcare. The project will develop new algorithms, methodologies, and software to make data-driven federated learning for healthcare more robust and trustworthy.

The first focus of the team will be to develop new ways to handle the diversity of health data. The team will then study potential methods of attacks on the federated learning systems, and develop defenses against such attacks. The researchers will also investigate ways that the system can “un-learn” in situations where users request the influence of their data be removed from the model. In all cases, the researchers will focus on developing fair and interpretable algorithms. Finally, the researchers will study ways to generate synthetic health data, which can be used to augment or replace real data to improve the AI models.

The scope of the work is ambitious, but Purushotham is confident in his team.

“I have great students and collaborators,” he says. “I’m really excited to make federated learning in healthcare work.”

A group of 6 people stand in a line and smile at the camera. — Sanjay Purushotham (third from left) with members of his research team (from left to right): Shaswati Shah, Catherine Ordun, Sultan Ahmed, Md Mahmudur Rahman, and Zahid Hassan Tushar. (Marlayna Demond ’11/UMBC)

UMBC teams with the Navy and the University of Arizona to develop new capabilities for hypersonic flight

When the Wright brothers first launched their famous plane off the tall sand dunes near Kitty Hawk, North Carolina, it flew slower than a person can run. Now, military fighter jets routinely rip through the air at supersonic speeds of 1,000 miles per hour or more. Uncrewed experimental aircraft have even gone hypersonic, traveling more than five times faster than the speed of sound.

Flying at such breakneck speeds presents an array of engineering challenges, from the stresses on the materials to the struggle to control the aircraft.

“When flying above the speed of sound, the operating environment can degrade extremely quickly and there is very little time to react,” says Ankit Goel, an assistant professor of mechanical engineering at UMBC. “If a correcting control signal is not applied quickly enough, catastrophic failure is almost always guaranteed.”

Left side shows 1903 Wright Flyer, right side shows F-22 fighter jet. — On left, the Wright Flyer on display at the Smithsonian Air and Space Museum in Washington, D.C. (Smithsonian Institution). On the right, an F-22 Raptor (U.S. Air Force photo by 2nd Lt. Samuel Eckholm.)

Goel has been investigating better ways to control aircraft flying at hypersonic speeds, by primarily focusing on the vehicle’s engine. He recently received more than $850,000 in funding from the Office of Naval Research to further the investigations. Over the next three years he will partner with Kyle Hanquist at the University of Arizona and researchers from the Naval Air Warfare Center (NAWC) to develop improved engine control strategies and assess their performance in ground experiments conducted at the NAWC facility at China Lake, California.

Powering superfast flight

The first airplane to break the sound barrier—the Bell X-1 piloted by Chuck Yeager in 1947—was dropped from the bomb bay of a Boeing B-29 and fired rocket engines to accelerate to its top speed.

Pilot stands in front of airplane. — Chuck Yeager in front of the X-1. (U.S. Air Force)

Rockets are incredibly powerful, but they also guzzle fuel. A more efficient alternative for fast flight is an air-breathing engine called a ramjet. Ramjets, which work best above the speed of sound, exploit the fast forward motion of the plane to effectively “ram” air into the engine and compress it. Squeezing the air heats it up, and the hot air then spontaneously ignites the fuel. As the burning mixture is channeled out the back of the engine, it pushes the aircraft forward.

Ramjets can burn either liquid or solid fuel. In the solid fuel version, which Goel and his collaborators are concentrating on in this latest project, sand-like grains of solid fuel are pressed together and embedded in the sides of the engine. This eliminates the need for pumps and other equipment to inject liquid fuel. It also means the fuel can more easily be transported and stored.

A solid fuel ramjet engine’s design is simple, but its operation is finicky. Structural vibrations, changes in airflow, and too much or too little heat can all cause the engine to stop working suddenly. Sometimes the engine will “buzz” in a rapid series of undesirable starts and stops, a state known as “engine unstart.”

“More reliable control of the engine could enable faster flight, longer range, and better maneuverability,” says Goel. To get that better control, Goel and his collaborators must grapple with the complex and chaotic environment inside the engine.

A tricky controls problem

Imagine a car driving along a hilly road. If the driver wants to maintain a constant speed, they must press the gas pedal harder going up the hills. If the car has cruise control, a computer can do the work of the driver. This is an example of a relatively simple control problem.

Ankit Goel (Marlayna Demond ’11/UMBC)

The inside of a solid fuel ramjet engine presents an example of a not-so-simple control problem.

A number of factors make the problem especially difficult. For starters, conditions inside the engine are constantly and rapidly changing. As the solid fuel burns away, the shape of the combustion zone inside the engine changes, which changes the airflow, which affects the rate of burning. Conditions shift dramatically in less than one-thousandth of a second. The system is also very sensitive to slight perturbations. A small change in flow conditions upstream can lead to big changes downstream.

It’s nearly impossible to completely understand and model what is going on.

In the face of such complexity, Goel is turning to a control technique that’s relatively novel in aerospace applications: machine learning. The beauty of machine learning is that it can solve problems without needing a conceptual understanding of them. The downside is that most machine learning requires enormous datasets and large amounts of computational power to work, two resources that aerospace applications typically lack. However, Goel and his collaborators have found a potential solution.

Real-time learning for aerospace

The researchers are experimenting with a technique that can learn, quite literally, on the fly, rather than being trained ahead of time. This distinguishes the technique from typical machine learning. While the ramjet is operating, the learning algorithms will continually re-evaluate the relationship between two simple factors—a measure of the air that’s being let into the engine and the thrust that the engine is generating—and use that relationship to drive the generated thrust to a desired value. By implementing a learning-based control scheme, the technique can control the output of the engine while ignoring the complexity of what’s actually happening inside.

The researchers plan to test their model first on computer simulations of a ramjet engine, and then on the real thing.

The ultimate goal is to embed the learning controller in a system that would take pressure sensor measurements from an engine in flight and use them to rapidly adjust the airflow into the engine—perhaps making thousands of small changes a second—to control the thrust. This would mean that flight operators could count on getting the requested thrust from the engine, even if flight conditions are changing.

Goel says the proposed learning technique could also be used on systems far removed from superfast aircraft. “The key insights into what makes the technique work in highly unmodeled and uncertain systems will allow us to generalize it to a large body of interesting dynamic problems,” he says. The team is already thinking about applying it to other types of aircraft, such as vehicles that use flapping wings or that take off and land vertically.

One man points to equations on a white board and discusses with another man. — Ankit Goel (left) and Parham Oveissi (right) are working to develop better control techniques for ramjet engines. (Marlayna Demond ’11/UMBC)

Parham Oveissi, a Ph.D student in mechanical engineering at UMBC who will be involved in the research, says he is motivated by a deep interest in developing control techniques for aircraft, an interest he has nurtured since childhood, when his parents gave him a toy quadcopter. “This early encounter ignited an enduring curiosity within me, driving me to unravel the mysteries of flight and explore methods of controlling these machines,” he says.

The research project will be an exciting opportunity to develop his knowledge and skills. “I’m excited to collaborate with professionals, gain valuable research experience, and see the impact of my contributions,” he says.

The success of the project will hinge on the joint efforts of researchers with a variety of skill sets, from a variety of institutions.

“This work requires an interdisciplinary team,” says Kyle Hanquist, an assistant professor in the Department of Aerospace and Mechanical Engineering at the University of Arizona who will contribute his expertise in reactive flow modeling to the project. “We are working together to tackle a difficult problem that none of us could tackle on our own.”

Building AI We Can Trust

The AI apocalypse is coming. Or it isn’t. Depending on what you read, you might get confused.

One thing is certain: Humans are fired up about smart machines. Much of the attention has focused on ChatGPT, an “artificial intelligence language model designed to generate human-like responses to natural language prompts” (in its own words).

ChatGPT gets coy if you ask whether its existence should be cause for human concern. “It’s important to recognize that I am a tool and not inherently good or bad. It’s how people choose to use me that can have positive or negative consequences,” it says.

Many researchers, however, are not so noncommittal. They see inherent flaws in the machine learning technology that forms the foundation of tools such as ChatGPT, and they would like to make it better.

While ChatGPT advises that “it’s always a good idea to double-check any important information I provide,” some UMBC researchers are working to build better safeguards into the AI systems themselves—AI the public can trust.

Colorful abstract imagery of hands coming out of a magician's hat. One holds a magnifying glass, the other a feathered pen. The hat has bunny ears and a single eye.

On March 22 of this year, a group including prominent artificial intelligence researchers and tech entrepreneurs released an open letter calling for a six-month pause on the training of powerful AI systems.

“AI systems with human-competitive intelligence can pose profound risks to society and humanity,” the letter argued. “Powerful AI systems should be developed only once we are confident that their effects will be positive and their risks will be manageable.”

The letter signers, including two UMBC faculty, expressed alarm at an AI arms race unleashed with the November 2022 public debut of ChatGPT, a celebrity chatbot that answers almost any question or prompt with humanlike ease. In a mere two months, the bot attracted 100 million users, and big tech companies began sprinting to deploy similar technology in their products.

Yet a general unease is accompanying this latest rush for AI gold.

ChatGPT can dazzle users with its eloquent prose (and poetry!), but it sometimes delivers complete falsehoods. People fret that such technology will eliminate jobs and empower scammers and dictators. And beneath it all, many researchers worry that we do not fully understand—nor can we reliably control—how creations such as ChatGPT work.

“At the core of many powerful AI systems today are what are called ‘blackbox’ models,” says Manas Gaur, an assistant professor in the Department of Computer Science and Electrical Engineering (CSEE) at UMBC. The models percolate data through layers of calculations so dense and complex that researchers struggle to track what’s happening inside. The models may excel at certain tasks—like writing sentences in ChatGPT’s case—but they cannot explain why they make the decisions they do. Sometimes they do perplexing, and erratic, things.

“Some people see ChatGPT and similar technology as a progressive tool while others fear it is dangerous,” says Nancy Tyagi, a master’s student in computer science at UMBC who is also working as a researcher in Gaur’s lab. “In my opinion, such tools are inherently risky and need further analysis. If these models are to be used in sensitive areas such as mental health or defense systems, then more work is required to make them safe, controllable, and trustworthy.”

Tyagi is working on a project to build an AI mental health assistant capable of initiating safe and appropriate conversations based on clinical guidelines in mental health. Her project is one of many that Gaur and other AI researchers at UMBC are launching with the aim of ensuring AI tools are accurate, transparent, and safe.

To better understand these researchers’ quest for trustworthy AI, it helps to take a step back and consider how the latest AI trend fits into the big picture.

Abstract illustration by David Habben, depicting robotic hands and an AI creature.

A Brief History of
Thinking Machines

When the field of artificial intelligence launched in the 1950s, computers were feeble compared to the muscular monsters that power systems such as ChatGPT today. Yet researchers were intrigued by the possibility of teaching them to think like humans. What followed was a roller coaster of booms and busts.

“The history of AI has been marked by periods of hype, followed by some level of disillusionment,” says Tim Finin, CSEE professor and a researcher at UMBC who has been studying AI problems for more than 50 years.

Driving the ups and downs were three interrelated factors: the power (and limits) of the hardware that formed computers’ brains, the data available to train those brains, and the “thinking strategies” AI researchers devised.

In the beginning, researchers taught machines to play games, learn language, and solve mathematical puzzles using a variety of “thinking” approaches. Yet the field hit a wall in the 1970s: Computers couldn’t store enough information or process it fast enough to tackle real-world problems. This was the first “AI winter,” when funding dried up and the topic faded from public view.

The birth of the microprocessor at the end of the decade revitalized AI research. Riding the shoots of this new life, a certain approach to machine thinking rose to prominence—that of the expert system. These AI programs were based on pre-programmed knowledge and logic meant to mirror the reasoning of human experts. Perhaps the most famous expert system was IBM’s Deep Blue, which beat the Russian chess champion Garry Kasparov in a chess match in 1997.

Expert systems could shine when solving narrowly defined problems (such as winning a game of chess), but they were brittle, says Finin. The systems struggled to adapt to fuzzy and fluid real-world situations, and it was cumbersome to program all the rules that an expert might use to evaluate a problem.

As the limits of expert systems became clear in the 1990s, AI felt the chill of a second winter.

It was another advance in computing hardware that thawed the field again after the turn of the 21st century. The graphic processing units developed to enhance video games supercharged computers’ speed and power at low cost. This, coupled with a flood of free data from the internet, propelled a new type of AI to the forefront: machine learning.

With loads of computing power and heaps of examples to learn from, researchers found surprising success getting computers to teach themselves how to think. The computers start with a question, perform some calculations, and guess the answer. They then compare it to the actual answer. If they are wrong, (which they usually are at first), they fiddle with the calculations and try again. After running billions of calculations, such systems can become quite proficient at tasks such as identifying images of cats and predicting the next word in a sentence.

THE SEASONS OF AI

The growth of the modern field of AI has been marked by a series of rapid spurts, followed by more dormant periods. People often liken these ups and downs to the seasons. During AI summers, public attention shines hot on the field. Yet the bountiful fruit of the season has often grown from seeds of ideas planted during quiet AI winters.

Summer 1: Expert systems

AI programs based on knowledge and logic flourished in the 1980s. Examples include systems that can identify unknown chemicals, diagnose diseases, and play chess. The systems are safe and explainable, but fail to adapt to fluid and complex situations.

Summer 2: Machine learning

Starting in the early 2010s, the potent combination of supercharged computing and heaps of free internet data powered AI’s second summer: the golden era of machine learning—an era that we are arguably still in.
AI systems started to recognize images, transcribe and translate language, and create text and art almost like humans do. These systems have surprised even their own creators with their range of abilities, but they are hard to understand, reason with, and control.

Into the future: Hybrid AI

It’s not yet clear when or if the second AI summer will turn to fall. But researchers are already planting the seeds for future advances. Combining the fruits of past summers, researchers hope to make future AI systems that are adaptable and safe, self-taught and able to explain their decisions.

Abstract illustration by David Habben, depicting a sun and some flowers.

FUN FACT:

UMBC’s first Ph.D. graduate in computer science, Sanjeev Bhushan Ahuja, earned his degree when expert systems dominated AI. His dissertation, published in 1985 and titled “An Artificial Intelligence Environment for the Analysis and Classification of Errors in Discrete Sequential Processes,” advances techniques popular during this time.

This approach to machine learning is called a neural network, so named because it was originally inspired by the way neurons in the brain work. Neural networks lie at the heart of most famous AI applications today, including image classification tools, voice recognition, and text and image generators.

The power (and limits) of
machine learning

When many of the new machine learning systems debuted, their powers seemed almost miraculous. But soon enough, drawbacks emerged. The machines require enormous data sets (and enormous amounts of energy) to learn. They will adopt biases from their training data and sometimes from their interactions with humans. A chatbot named Tay was quickly scuppered after its 2016 release, when users pushed it into spewing racist and sexist ideas.

Machine learning systems can also fail spectacularly in individual instances (even if they get answers correct most of the time). For example, a driver was killed in 2016 when the autopilot in a Tesla car failed to recognize the side of a white trailer truck against a bright sky.

The blackbox nature of state-of-the-art machine learning means the systems are unable to explain or justify their conclusions, giving users—and even their own creators—little insight into their thinking. For the most part, the systems struggle to build consistent worldviews or reason logically.

The weaknesses of learning models also leave them susceptible to malicious manipulation. Adversaries might “poison” the data used to train the models or exploit the model’s opaqueness to hide an attack.

“It is time we fall back from trusting these models,” says Gaur, whose personal push to make AI systems more explainable, robust, and safe is part of a growing international movement.

Another UMBC researcher joining the push is Houbing Song, a professor in the Department of Information Systems at UMBC. Song says that transportation, defense, medicine, and the law are some areas where explainable and safe AI systems are needed the most.

As researchers tackle the challenge of making current AI systems better, they are often returning to ideas from an earlier era of AI.

Hybrid systems to merge logic and learning

If the AI systems of the 1980s married the AI systems of the 2010s, their baby might be the type of system Gaur, Song, and others are working to develop.

These systems look to deliver the learning capabilities of neural networks alongside the safeguards of knowledge and rule-based systems.

In the field of mental health, Gaur points out that current chatbot systems are not well suited to answering patients’ questions since they can give unsafe or off-the-wall responses.

“Guaranteeing these systems’ safety calls for more than just improving their overall performance” he says. “We must also make sure the systems are prevented from giving risky answers.”

Working with Karen Chen, an assistant professor from the Department of Information Systems, Gaur has written a paper highlighting the properties that AI-powered virtual mental health assistants should exhibit to be considered safe and effective.

Creating “AI Scientists” at UMBC

New scientific discoveries often lay the groundwork for significant advances in human well being. Think of medical treatments that spring from a better understanding of the human body or labor-saving devices we fashion using our knowledge of material properties.

Tyler Josephson, an assistant professor in the Department of Chemical, Biochemical, and Environmental Engineering, hopes to turbocharge science’s discovery engine, with a little help from AI.

Josephson has started a new project to translate chemical theories into a machine-readable mathematical language. Once the computers have access to the foundations of science, Josephson believes they could be tasked with logically manipulating that information to reveal new discoveries.

You might wonder if Josephson has any worries about creating his own AI-powered replacement. But he doesn’t think AI scientists will displace the human kind.

“I think scientists have so many different problems to solve. And if we solve them faster with AI, they just open up brand new questions for us to go after next,” Josephson said in an interview about his work with the Canadian radio program Quirks & Quarks.

Abstract illustration by David Habben, depicting a figure wearing glasses and holding out an atom in one hand.

Together with his students, he is also working to create such systems. Using an approach called knowledge-infused learning, the researchers are looking to anchor their AI systems in clinically approved guidelines. They are also pushing their systems to reveal their thinking so that the approaches can be checked by mental health experts. Sometimes the results reveal that even when a system arrives at a correct conclusion, the information it used to reach that conclusion may be irrelevant to a human doctor’s thinking.

Song has also been coaxing AI learning models to open up. In a recent paper, he and his co-authors developed a tool to identify attacks on an image-recognition program by figuring out which parts of its neural network are most susceptible to manipulation.

In the fall of 2023, he will be teaching a new graduate-level course on a broad category of hybrid AI called neurosymbolic AI. UMBC will be only the second university in the world to offer such a course, he says.

Song arrived at UMBC in January on the heels of winning major honors for his research in computing and engineering and is looking forward to turning more of his attention to this emerging frontier in AI research. He says he eventually hopes to build a world-class AI research institute at his new academic home, focused on delivering learning machines that can be confidently used when safety is a top priority.

“I recognize the need for trustworthy AI,” Song says. “I believe that this field of research is where I can make unique contributions and take on responsibilities for my professional communities and my home institution.”

Abstract illustration by David Habben, depicting artificial intelligence.

Technology to benefit society

The initial goal of AI, as defined by a group of researchers credited with launching the field at a 1956 workshop, was to “make machines use language, form abstractions and concepts, solve [the] kinds of problems now reserved for humans, and improve themselves.” But if the aim is human-like thinking, it naturally raises the question: How do humans think?

In a bestselling book titled “Thinking, Fast and Slow,” world-renowned psychologist Daniel Kahneman posits that humans have two thinking systems: a fast one and a slow one. The fast one is the thinking that comes to mind almost without effort, and we use this thinking most of the time. Yet it is prone to errors. The deliberative slow thinking system catches mistakes and enables breakthroughs in understanding.

Finin compares machine learning models to the fast-thinking system while knowledge and logic-based systems are more like the slow-thinking system. To make today’s faddish, fast-thinking models more competent, researchers such as Gaur, Song, and their students are extending them with slow-thinking capabilities.

We may still be decades away from AI systems approaching the full range of human intelligence. There are many ethical questions to grapple with before we reach a Hollywood-esque future of self-flying cars and android coworkers. Yet the decades of AI research up to this point have already transformed the world. AI concepts underpin the ways we search the web, shop online, and otherwise interact with the digital world.

AI has enormous potential to improve human lives, but we must proceed wisely. UMBC researchers are at the frontiers of AI research, pushing the limits of knowledge and theory, and striving to make the technology better for the benefit of society.

Training Your Robot Assistants

If you hope the AI revolution will bestow humanity with machine “Jeeves” capable of meeting your every need, Cynthia Matuszek has some bad news. “I’m always being asked: ‘When will we have robot butlers?’ I have to say—not any time soon,” says Matuszek, an associate professor in the Department of Computer Science and Electrical Engineering.

Matuszek researches how to build robots that understand human commands in complex and chaotic natural environments. She has successfully trained a robot hand that can respond to written prompts such as “Grab the apple.” She is also exploring how to teach robots to understand spoken language and to learn new concepts, such as how to dice a vegetable, if a human shows them how.

Part of what motivates Matuszek’s work is the huge unmet demand for caregivers to assist people as they age. Robots might fill the gap. Matuszek says we likely won’t have “Jack of all trades” helpers, but robots could specialize in certain tasks, such as preparing food or folding laundry.

Another part of what motivates Matuszek is the thrill of being the first person to discover how to do something new. “It’s really just so much fun,” she says.

UMBC leads research into light-based timing and navigation technologies for DOD-funded consortium

Every day, radio signals from GPS satellites help millions of people figure out what time it is and where they are. Yet the system is vulnerable to disruptions and attacks. Sometimes users are unable to access critical information. Other times, adversaries may try to fool users into thinking they are somewhere they aren’t.

For this reason, researchers at UMBC are working to develop alternative timing and navigation technologies. The university recently received almost $2 million in initial funding from the Department of Defense (DOD) to further this important research. UMBC will collaborate with the Army Research Laboratory (ARL) in Adelphi, Maryland, and other members of a national consortium, managed by the National Center for Manufacturing Sciences (NCMS). UMBC will conduct fundamental research to develop the knowledge base that is needed to design, test, and build clocks and communication protocols that could deliver critical information in the event of a disruption to GPS service.

Four people sit around table in conversation. — From left to right, Professor Curtis Menyuk, graduate student Logan Courtright, Professor Gary Carter and graduate student Pradyoth Shandilya discuss research plans. (Marlayna Demond ’11/UMBC)

The work will be carried out within the newly launched Center for Navigation, Timing and Frequency Research (Centaνr) at UMBC, led by Curtis Menyuk, professor of computer science and electrical engineering. Centaνr is the second significant research partnership with the ARL that UMBC has launched in recent years. It joins the Center for Real-time Distributed Sensing and Autonomy (CARDS), which opened in 2021 and aims to develop smart robots that can better navigate difficult terrain and coordinate their actions with other robots and humans.

Harnessing the power of light

Centaνr is part of a wider, 10-member consortium that brings together partners from the government, academia, and industry to advance photonic technologies that harness light to process and send information. The partners seek to develop solutions that utilize light in environments where radio frequency solutions do not work. These solutions will be chip-based—taking advantage of modern advances in integrating optical and electronic technology on a single semiconductor chip—in order to achieve low size, weight, power and cost.

A yellow waveform on a grid background — To analyze photonic systems, the team will gather and study data such as this optical spectrum of an optical pulse called a soliton. (Image courtesy of Alioune Niang.)

The UMBC team will lead research on the design and manufacture of photonic technology for positioning, navigation, and timing, which is one of the three main research thrusts for the consortium.

To meet high performance requirements, the photonic elements must be manufactured with extreme precision. UMBC will partner with Worcester Polytechnic Institute and AIM Photonics, one of the U.S. Department of Defense Manufacturing Innovation Institutes, as well as the Army Research Laboratory to design these devices.

Researchers at UMBC will also work to help develop a system that uses light waves to transfer a time signal between two devices, through the open air.

The UMBC team includes several computer science and electrical engineering faculty: Professor Gary Carter, Professor Fow-Sen Choa, Associate Professor Tinoosh Mohsenin and Assistant Professor Ergun Simsek.

High-tech equipment meets motivated students

To further Centaνr’s research aims, UMBC will build a new, high-tech laboratory where photonic components can be tested and characterized. The university will also serve its educational mission by recruiting and training diverse students in the concepts of timing and navigation technology and photonics.

Two people look at electronic equipment. — Gary Carter and Research Associate Alioune Niang look at equipment used to study photonic components. The equipment will move to a new lab space soon. (Marlayna Demond ’11/UMBC)

In collaboration with consortium partners, UMBC will develop new course materials and internship programs, and will recruit students from groups underrepresented in this field to participate in these research and learning opportunities.

“There is a real need for good educational material in these areas and I’m excited by the opportunity to build and distribute it,” says Menyuk. He is also excited to build a new experimental facility and partner with other institutions to make U.S.-manufactured high-end photonic components more widely available.

Menyuk says the consortium partners are already discussing additional years of funding. He is also thinking long-term and planning ways that Centaνr can continue to be a source of frontier photonics research for years to come.

UMBC’s vibrant learning community helps students discover careers to fit their passions

Performers of the music piece “Corporel,” by the French-Slovenian composer Vinko Globokar, must use their own body as a percussion instrument. They beat, scratch, smack, and tap themselves. They chatter their teeth, snore, and cluck their tongue. The composition’s “patterns of sound and gesture” are arresting, “keeping us transfixed even as we flinch,” the L.A. Philharmonic website says in its description of the piece.

To Brandon Gouin ’23, music performance, learning the work was a highlight of his time at UMBC. “This work is a discovery of self and musical potential that resonates deeply within me,” he says. He performed the piece at his senior recital this year, when members of the music community at UMBC were celebrating the opportunity to once again gather in practice spaces and concert halls after the isolation of COVID-19.

Gouin credits his teachers and mentors, especially Tom Goldstein, associate professor of music, and Patrick Crossland, affiliate artist, with helping him reach that moment on stage—as well as with helping him find his path as an artist.

Gouin’s experience is reflective of the experience of many members of the class of 2023. Although each individual has unique talents, passions, and goals, they are brought together by a feeling of gratitude to the UMBC community for helping them find their way. Faculty, staff and fellow students helped them open their minds to new possibilities, and mentored them on their journey to discovering a career path that fits.

Challenging preconceptions

When Gouin arrived at UMBC, he thought performance jobs for percussionists were mostly limited to professional orchestras—a highly competitive and difficult career path for any musician to pursue. However, as he dove into the local music scene, his eyes were opened to other possibilities.

“I began to see opportunities as a contemporary performer by attending the Livewire festival of New Music that UMBC holds every year and watching many performances of music I had never heard of before,” he says. He also attended local experimental music shows at the Red Room in Baltimore and the Rhizome in D.C.

“Audiences at these shows are very engaged, and that is exciting to me as a performer,” he says.

Man sit on a stage, performing, wearing orange pants and no shirt, snapping his fingers — Brandon Gouin performing Vinko Globokar’s 1985 work “Corporel.” (Image courtesy of Brandon Gouin)

Mentors illuminate new possibilities

Gouin’s professors introduced him to new styles of music and showed him the vibrancy of the contemporary music world. Tom Goldstein directs the UMBC Percussion Ensemble of which Gouin was a member. Goldstein says he often programs works by lesser-known composers, and sometimes pieces by UMBC faculty and student composers.

“I think if the students perform a piece composed by someone they know, they may start to think ‘Hey, I could do something like that, too,’” he says.

Gouin appreciated the exposure to a variety of music and says he has adopted the inclusive mindset of his music professors as he pursues career opportunities after graduation. “My teachers aren’t only in love with one or the other kind of music but commit themselves fully to working with all kinds of music,” he reflects. “I think that kind of mindset is healthier and has more longevity for an artist’s life.”

A community of performers strikes various poses on the stage. — Members of the UMBC Percussion Ensemble and the Salisbury University Percussion Ensemble. Every year the two groups perform a joint concert. (Image courtesy of Brandon Gouin)

Other Class of 2023 graduates say they encountered similar opportunities for growth and expression at UMBC. Shaniah Reece ’23, information systems, discovered a love of research and a way to connect her technical skills to her passion for social justice. Hala Algrain, M.P.S. ’23, health information technology, reconnected with a love of teaching, and switched her career plans from industry to academia. Elijah Mugabe ’23, chemistry, threw himself into lab work and a quest to investigate unanswered scientific questions. And Connor McPherson ’23, history, found a way to connect his interest in the humanities with a career in the Navy.

Community builds confidence

In addition to finding their path at UMBC, these students also found communities that encouraged them to excel on their academic and personal journeys. They found student groups, faculty mentors, peer advisors, and scholars’ programs that pushed them to succeed and provided the support they needed.

“Undoubtedly, the most enriching part of my time at UMBC has been the remarkable sense of community I have experienced here,” Reece says “It has provided me with opportunities for personal growth, enabling me to evolve holistically. As I reflect upon my experiences, I feel confident and prepared to take on any obstacles that may lie ahead.”

5 people pose for camera, 3 stand in the middle, 2 are seated on either side. — Reece (second from left, UMBC shirt) and other CWIT scholars and students at the Grace Hopper Conference in Florida 2022. (Courtesy of Shaniah Reece.)

Algrain agrees that the culture of UMBC and the level of support were the best parts of her experience at UMBC.

For Gouin, the chance to connect with mentors who were dedicated performers, as well as enthusiastic teachers, was life changing. He even joined a percussion quartet with fellow and former students that they named “Hi Tom,” in honor of Tom Goldstein.

Goldstein shares that for him the honor goes in the opposite direction.

“I love working with the students—it’s a fantastic part of my life,” he says. “It’s an honor and a privilege to get to know them.”

Read more Class of 2023 stories.