Researchers at Georgia Tech have taken a critical step forward in creating efficient, useful and brain-like artificial intelligence (AI). The key? A new algorithm that results in neural networks with internal structure more like the human brain.

The study, “TopoNets: High-Performing Vision and Language Models With Brain-Like Topography,” was awarded a spotlight at this year’s International Conference on Learning Representations (ICLR), a distinction given to only 2 percent of papers. The research was led by graduate student Mayukh Deb alongside School of Psychology Assistant Professor Apurva Ratan Murty.

Thirty-two of Tech’s computing, engineering, and science faculty represented the Institute at ICLR 2025, which is globally renowned for sharing cutting-edge research. 

“We started with this idea because we saw that AI models are unstructured, while brains are exquisitely organized,” says first-author Deb. “Our models with internal structure showed more than a 20 percent boost in efficiency with almost no performance losses. And this is out-of-the-box — it’s broadly applicable to other models with no extra fine-tuning needed.”

For Murty, the research also underscores the importance of a rapidly growing field of research at the intersection of neuroscience and AI. “There's a major explosion in understanding intelligence right now,” he says. “The neuro-AI approach is exciting because it helps emulate human intelligence in machines, making AI more interpretable.”

“In addition to advancing AI, this type of research also benefits neuroscience because it informs a fundamental question: Why is our brain organized the way it is?,” Deb adds. “Making AI more interpretable helps everyone.”

Brain-inspired blueprints

In the brain, neurons form topographic maps: neurons used for comparable tasks are closer together. The researchers applied this concept to AI by organizing how internal components (like artificial neurons) connect and process information. 

This type of organization has been tried in the past but has been challenging, Murty says. “Historically, rules constraining how the AI could structure itself often resulted in lower-performing models. We realized that for this type of biophysical constraint, you simply can’t map everything — you need an algorithmic solution.”

“Our key insight was an algorithmic trick that gives the same structure as brains without enforcing things that models don't respond well to,” he adds. “That breakthrough was what Mayukh (Deb) worked on.” 

The algorithm, called TopoLoss, uses a loss function to encourage brain-like organization in artificial neural networks, and it is compatible with many AI systems capable of understanding language and images. 

“The resulting training method, TopoNets, is very flexible and broadly applicable,” Murty says. “You can apply it to contemporary models very easily, which is a critical advancement when compared to previous methods.” 

Neuro-AI innovations

Murty and Deb plan to continue refining and designing brain-inspired AI systems. “All parts of the brain have some organization — we want to expand into other domains,” Deb says. “On the neuroscience side of things, we want to discover new kinds of organization in brains using these topographic systems.”

Deb also cites possibilities in robotics, especially in situations like space exploration where resources are limited. “Imagine running a model inside a robot with limited power,” he says. “Structured models can help us achieve 80 percent of performance with just 20 percent of energy consumption, saving valuable energy and space. This is still experimental, but it's the direction we are interested in exploring.”

“This success highlights the potential of a new approach, designing systems that benefit both neuroscience and AI — and beyond,” Murty adds. “We can learn so much from the human brain, and this project shows that brain-inspired systems can help current AI be better. We hope our work stimulates this conversation.”

Solar cells account for approximately six percent of the electricity used on Earth; however, in space, they play a significantly larger role, with nearly all satellites relying on advanced solar cells for their power. That’s why Georgia Tech researchers will soon send 18 photovoltaic cells to the International Space Station for a study of how space conditions affect the devices’ operation over time.

“The main goal here is to improve power generation in space,” said Jud Ready, principal research engineer at the Georgia Tech Research Institute (GTRI) and Executive Director of Georgia Tech's Space Research Institute. “The limiting factor on the performance of a spacecraft is usually how much power you can produce. Power, size, weight, complexity, cost – all of these are tied closely to the electrical generation of the solar panels.”

Read the story in the GTRI newsroom.

Overwhelmed doctors and nurses struggling to provide adequate patient care in South Korea are getting support from Georgia Tech and Korean-based researchers through an AI-powered robotic medical assistant.

Top South Korean research institutes have enlisted Georgia Tech researchers Sehoon Ha and Jennifer G. Kim to develop artificial intelligence (AI) to help the humanoid assistant navigate hospitals and interact with doctors, nurses, and patients.

Ha and Kim will partner with Neuromeka, a South Korean robotics company, on a five-year, 10 billion won (about $7.2 million US) grant from the South Korean government. Georgia Tech will receive about $1.8 million of the grant.

Ha and Kim, assistant professors in the School of Interactive Computing, will lead Tech’s efforts and also work with researchers from the Korea Advanced Institute of Science and Technology and the Electronics and Telecommunications Research Institute.

Neuromeka has built industrial robots since its founding in 2013 and recently decided to expand into humanoid service robots.

Lee, the group leader of the humanoid medical assistant project, said he fielded partnership requests from many academic researchers. Ha and Kim stood out as an ideal match because of their robotics, AI, and human-computer interaction expertise. 

For Ha, the project is an opportunity to test navigation and control algorithms he’s developed through research that earned him the National Science Foundation CAREER Award. Ha combines computer simulation and real-world training data to make robots more deployable in high-stress, chaotic environments. 

“Dr. Ha has everything we want to put into our system, including his navigation policies,” Lee said. “He works with robots and AI, and there weren’t many candidates in that space. We needed a collaborator who can create the software and has experience running it on robots.”

Ha said he is already considering how his algorithms could scale beyond hospitals and become a universal means of robot navigation in unstructured real-world environments.

“For now, we’re focusing on a customized navigation model for Korean environments, but there are ways to transfer the data set to different environments, such as the U.S. or European healthcare systems,” Ha said. 

“The final product can be deployed to other systems and industries. It can help industrial workers at factories, retail stores, any place where workers can get overwhelmed by a high volume of tasks.”

Kim will focus on making the robot’s design and interaction features more human. She’ll develop a large-language model (LLM) AI system to communicate with patients, nurses, and doctors. She’ll also develop an app that will allow users to input their commands and queries. 

“This project is not just about controlling robots, which is why Dr. Kim’s expertise in human-computer interaction design through natural language was essential.,” Lee said. 

Kim is interviewing stakeholders from three South Korean hospitals to identify service and care pain points. The issues she’s identified so far relate to doctor-patient communication, a lack of emotional support for patients, and an excessive number of small tasks that consume nurses’ time.

“Our goal is to develop this robot in a very human-centered way,” she said. “One way is to give patients a way to communicate about the quality of their care and how the robot can support their emotional well-being.

“We found that patients often hesitate to ask busy nurses for small things like getting a cup of water. We believe this is an area a robot can support.”

The robot’s hardware will be built in Korea, while Ha and Kim will develop the software in the U.S.

Jong-hoon Park, CEO of Neuromeka, said in a press release the goal is to have a commercialized product as soon as possible. 

“Through this project, we will solve problems that existing collaborative robots could not,” Park said. “We expect the medical AI humanoid robot technology being developed will contribute to reducing the daily work burden of medical and healthcare workers in the field.”

A longstanding mystery of the periodic table involves a group of unique elements called lanthanides. Also known as rare earth elements, or REEs, these silvery-white metals are challenging to isolate, given their very similar chemical and physical properties. This similarity makes it difficult to distinguish REEs from one other during extraction and purification processes. 

The world has come to depend on lanthanides’ magnetic and optical properties to drive much of modern technology — from medical imaging to missiles to smart phones. These metals also are in short supply, and because they’re found in minerals, lanthanides are difficult to mine and separate.   But that may change — thanks to a Georgia Tech-led discovery of a new oxidation state for a lanthanide element known as praseodymium.  

For the first time ever, praseodymium achieved a 5+ oxidation state. Oxidation occurs when a substance meets oxygen or another oxidizing substance. (The browning on the flesh of a cut apple, as well as rust on metal, are examples of oxidation.)
   
As far back as the 1890s, scientists suspected lanthanides might have a 5+ oxidation state, but  lanthanides in that state were too unstable to see, said Henry ”Pete“ La Pierre, an associate professor in Georgia Tech’s School of Chemistry and Biochemistry. Discovering an element’s new oxidation state is like discovering a new element. As an example, La Pierre noted how plutonium’s discovery opened up a whole new area of the periodic table. 

“A new oxidation state tells us what we don’t know and gives us ideas for where to go,” he explained. “Each oxidation state of an element has distinct chemical and physical properties — so the first glimpse of a novel oxidation presents a roadmap for new possibilities.”
 
La Pierre and colleagues at University of Iowa and Washington State University recently discovered the 5+ oxidation state for lanthanides. 

“It was predicted but never seen until we found it,” said La Pierre, corresponding author of the study, “Praseodymium in the Formal +5 Oxidation State,” which was recently published in Nature Chemistry. “Lanthanides’ properties are really fantastic. We only use them commercially in one oxidation state — the 3+ oxidation state — which defines a set of magnetic and optical properties. If you can stabilize a higher oxidation state, it could lead to entirely new magnetic and optical properties.”
 
The researchers’ breakthrough will broaden the lanthanides’ technical applications in fields such as rare-earth mining and quantum technology and could lead to new electronic device architectures and applications. 

“Research in lanthanides has already yielded significant dividends for society in terms of technological development,” La Pierre added.
    
The researchers hope to discover new tools for mining critical REEs, including improving lanthanide separation and recycling processes. When mining these elements, lanthanide elements are frequently mixed together. The separation process is painstaking and inefficient, generating a significant amount of waste. But with increasing global demand for REEs, the U.S. faces a supply issue. Figuring out how to improve lanthanides separation, potentially through oxidation chemistry, will ultimately enhance the supply of these critical elements. 

— Anne Wainscott-Sargent
 
Funding: This research was supported by grants from the National Science Foundation and the U.S. Department of Energy. 
 

The Institute for Matter and Systems (IMS) at Georgia Tech has announced the Spring 2025 Core Facility Seed Grant recipients. The primary purpose of this program is to give graduate students in diverse disciplines working on original and unfunded research in micro- and nanoscale science and engineering the opportunity to access the most advanced academic cleanroom space in the Southeast. In addition to using the labs' state-of-the-art fabrication, lithography, and characterization tools, the awardees will have the opportunity to gain proficiency in cleanroom and tool methodology and access the consultation services provided by research staff members in IMS. Seed Grant awardees are also provided travel support to present their research at a scientific conference.

In addition to student research skill development, this biannual grant program gives faculty with novel research topics the ability to develop preliminary data to pursue follow-up funding sources. The Core Facility Seed Grant program is supported in part by the Southeastern Nanotechnology Infrastructure Corridor (SENIC), a member of the National Science Foundation’s National Nanotechnology Coordinated Infrastructure (NNCI).

The five winning projects in this round were awarded IMS cleanroom and lab access time to be used over the next year. 

The Spring 2025 IMS Core Facility Seed Grant recipients are:

Stretchable Power Sources for Vertically Integrated Bioelectronics
PI: Antonio Facchetti
Student: Sakshi Sharma
School of Materials Science and Engineering

Next-generation 3D Solid-state Neutron Detectors for Nuclear Nonproliferation
PI: Anna Erickson
Student: David Straub
George W. Woodruff School of Mechanical Engineering

Programmable Microchip-based Cytotoxicity Assay for Real-Time Immune Cell Profiling
PI: Fatih Sarioglu
Student: Ahmadreza Rostamzadeh
School of Electrical and Computer Engineering

Experimental Study of Neutron-induced Radiation Damage in Lunar Materials with Key Implications for the Future of Lunar Geochronology
PI: Karl Lang
Student: Shreya Mukherjee
School of Earth and Atmospheric Sciences

Enhanced Dielectrophoretic Enrichment and Removal of Microplastics from Drinking Water via Engineered Nonuniform Electric Fields on Microfluidic Chips
PI: Xing Xie
Student: Shuai Wang
School of Civil and Environmental Engineering

An algorithmic breakthrough from School of Interactive Computing researchers that earned a Meta partnershipdrew more attention at the IEEE International Conference on Robotics and Automation (ICRA).

Meta announced in February its partnership with the labs of professors Danfei Xu and Judy Hoffman on a novel computer vision-based algorithm called EgoMimic. It enables robots to learn new skills by imitating human tasks from first-person video footage captured by Meta’s Aria smart glasses. 

Xu’s Robot Learning and Reasoning Lab (RL2) displayed EgoMimic in action at ICRA May 19-23 at the World Congress Center in Atlanta.

Lawrence Zhu, Pranav Kuppili, and Patcharapong “Elmo” Aphiwetsa — students from Xu’s lab — used Egomimic to compete in a robot teleoperation contest at ICRA. The team finished second in the event titled What Bimanual Teleoperation and Learning from Demonstration Can Do Today, earning a $10,000 cash prize.

Teams were challenged to perform tasks by remotely controlling a robot gripper. The robot had to fold a tablecloth, open a vacuum-sealed container, place an object into the container, and then reseal it in succession without any errors.

Teams completed the tasks as many times as possible in 30 minutes, earning points for each successful attempt.

The competition also offered different challenge levels that increased the points awarded. Teams could directly operate the robot with a full workstation view and receive one point for each task completion. Or, as the RL2 team chose, teams could opt for the second challenge level.

The second level required an operator to control the task with no view of the workstation except for what was provided to through a video feed. The RL2 team completed the task seven times and received double points for the challenge level.

The third challenge level required teams to operate remotely from another location. At this level, teams could earn four times the number of points for each successful task completed. The fourth level challenged teams to deploy an algorithm for task performance and awarded eight points for each completion.

Using two of Meta’s Quest wireless controllers, Zhu controlled the robot under the direction of Aphiwetsa, while Kuppili monitored the coding from his laptop.

“It’s physically difficult to teleoperate for half an hour,” Zhu said. “My hands were shaking from holding the controllers in the air for that long.”

Being in constant communication with Aphiwetsa helped him stay focused throughout the contest.

“I helped him strategize the teleoperation and noticed he could skip some of the steps in the folding,” Aphiwetsa said. “There were many ways to do it, so I just told him what he could fix and how to do it faster.”

Zhu said he and his team had intended to tackle the fourth challenge level with the EgoMimic algorithm. However, due to unexpected time constraints, they decided to switch to the second level the day before the competition due to unexpected time constraints. 

“I think we realized the day before the competition training the robot on our model would take a huge amount of time,” Zhu said. “We decided to go for the teleoperation and started practicing.”

He said the team wants to tackle the highest challenge level and use a training model for next year’s ICRA competition in Vienna, Austria.

ICRA is the world’s largest robotics conference, and Atlanta hosted the event for the third time in its history, drawing a record-breaking attendance of over 7,000.

A new mission strives to take black hole imaging to space. Scientists from the Georgia Institute of Technology, the Georgia Tech Research Institute (GTRI), the National Aeronautics and Space Administration (NASA), and 12 universities from around the world recently convened for a three-day workshop to plan the launch of the Space-based Precision Millimeter Interferometry Telescope (SPRITE) project. The proposed NASA Medium-Class Explorer mission aims to revolutionize the understanding of black holes through space-based imaging.

From Earth to orbit: The next step

SPRITE builds on the groundbreaking achievements of the Event Horizon Telescope (EHT), a network of ground-based telescopes able to synchronize observations from around the globe. EHT is most well-known for capturing the first images of black holes, M87* and Sagittarius A*.

“We’ve done what we can from the ground; we’ve run out of Earth,” says Professor and Chair of the School of Physics Feryal Özel, SPRITE’s principal investigator and a well-known astrophysicist instrumental in EHT’s success and development. “SPRITE will send two telescopes into orbit – achieving better imaging than a dozen telescopes on the ground.”

By sending the telescopes into space, the mission will be able to overcome the limitations of Earth’s atmosphere, which blocks certain wavelengths of light and produces turbulence that can degrade image quality. Unlike Earth-based telescopes, which rely on the planet’s rotation to change viewing angles, SPRITE’s telescopes will rotate independently across the vastness of space with data continuously transmitted from the satellites to ground stations.

“I like to think of it as an MRI machine rotating around a patient,” explains Özel. “In space, our telescopes can perform this orbital dance from great distances – giving us multiple perspectives of a black hole and allowing us to build a much more complete image.”

Mission goals

SPRITE’s objectives are ambitious and far-reaching, specifically to:

• Create more images of previously unseen black holes at resolutions better than M87* and Sagittarius A*;
• Confirm the presence of binary black holes through visual imagery; and
• Study the hot gas dynamics around black holes.

This class of mission requires a three-year operational lifetime to achieve its main science goals – although planners estimate the project will be able to operate considerably longer.

Preparing for launch

SPRITE is being organized to reflect Georgia Tech’s commitment to advancing space science through interdisciplinary collaboration and innovation, and will work closely with the Institute’s new Space Research Initiative. Locating SPRITE at Georgia Tech allows the mission to benefit from the knowledge of leading experts from the Colleges of Sciences, Engineering, and Computing; and GTRI. 

The recent kickoff meeting marked SPRITE’s first large-scale gathering of contributors from around the world.

“We had smaller meetings before, but this was the first time the full team came together to share expertise and collaboratively shape the mission,” says Özel. “Most importantly, this meeting showed us that we have a strong scientific case for our mission and its design.”

Over the next two to three years, the team will work to validate key technologies and prepare a compelling proposal for NASA. If selected, SPRITE is expected to launch in the mid-2030s, marking the beginning of a new era in space imaging.

Lithium-ion batteries power everything from electric cars to laptops to leaf blowers. Despite their widespread adoption, lithium-ion batteries carry limited amounts of energy, and rare overheating can lead to safety concerns. Consequently, for decades, researchers have sought a more reliable battery. 

Solid-state batteries are less flammable and can hold more energy, but they often require intense pressure to function. This requirement has made them difficult to use in applications, but new research from Georgia Tech could change that. 

The research group of Matthew McDowell, professor and Carter N. Paden Jr. Distinguished Chair in the George W. Woodruff School of Mechanical Engineering and the School of Materials Science and Engineering, has designed a new metal for solid-state batteries that enables operation at lower pressures. While lithium metal is often used in these batteries, McDowell’s group discovered that combining lithium with softer sodium metal results in improved performance and novel behavior.

McDowell and his collaborators presented their findings in the paper, “Interface Morphogenesis with a Deformable Secondary Phase in Solid-State Lithium Batteries,” published in Science on June 5.

Stackable Solution

Lithium-ion batteries have been the industry standard because they combine compact size, reliability, and longevity. However, they contain a liquid “electrolyte,” which helps lithium ions move in the battery but is also flammable. In solid-state batteries, this electrolyte is a solid material that is less flammable. The challenge is that when the battery is used, the lithium metal in the battery changes its shape, potentially losing contact with the solid electrolyte, which degrades performance. A common way to ensure the metal doesn’t lose contact is to apply high pressure to these batteries.

“A solid-state battery usually requires metal plates to apply this high pressure, and those plates can be bigger than the battery itself,” McDowell said. “This makes the battery too heavy and bulky to be effective.”

The researchers, led by Georgia Tech research scientist Sun Geun Yoon, sought a solution. The solid-state batteries would still require some pressure to function, but they found that by also using a softer metal, less pressure is required. The researchers decided to pair the commonly used lithium metal with a surprising element: sodium. 

“Adding sodium metal is the breakthrough,” McDowell noted. “It seems counterintuitive because sodium is not active in the battery system, but it’s very soft, which helps improve the performance of the lithium.”

How soft can sodium be? In a controlled environment, a person could stick their gloved finger into sodium metal and leave an imprint. 

From Biology to Battery

To understand the enhanced performance of their battery, the researchers borrowed a concept from biology called morphogenesis. This concept explains how tissues or other biological structures evolve based on local stimuli. Morphogenesis is rarely seen in materials science, but the researchers found that the combination of lithium and sodium behaves according to this concept. 

McDowell’s research group has been working on applying morphogenesis to battery materials as part of a project funded by the Defense Advanced Research Projects Agency in collaboration with several other universities. Their battery is among the first viable demonstrations of this concept — effectively, the sodium deforms readily at the low pressures needed for solid-state batteries to function. 

Battery Boon

The possibilities of a viable, smaller solid-state battery are vast. Imagine a phone battery that could last much longer or an electric vehicle that could drive 500 miles between charges. With this in mind, McDowell and his team have filed for a patent for this battery system.

While solid-state batteries still have some way to go before commercial use, results like these could mean that solid-state batteries can compete with lithium-ion. McDowell’s lab continues to experiment with other materials to further improve performance. 

Funding from the Defense Advanced Research Projects Agency.