NVIDIA AI Summit Japan: NVIDIA’s role in Japan’s big AI ambitions

Japan is on a mission to become a global AI powerhouse, and it’s starting with some impressive advances in AI-driven language models. Japanese technology experts are developing advanced models that grasp the unique nuances of the Japanese language and culture—essential for industries such as healthcare, finance,…

Unmasking Bias in Artificial Intelligence: Challenges and Solutions

The recent advancement of generative AI has seen an accompanying boom in enterprise applications across industries, including finance, healthcare, transportation. The development of this technology will also lead to other emerging tech such as cybersecurity defense technologies, quantum computing advancements, and breakthrough wireless communication techniques. However,…

Graph-based AI model maps the future of innovation

Imagine using artificial intelligence to compare two seemingly unrelated creations — biological tissue and Beethoven’s “Symphony No. 9.” At first glance, a living system and a musical masterpiece might appear to have no connection. However, a novel AI method developed by Markus J. Buehler, the McAfee Professor of Engineering and professor of civil and environmental engineering and mechanical engineering at MIT, bridges this gap, uncovering shared patterns of complexity and order.

“By blending generative AI with graph-based computational tools, this approach reveals entirely new ideas, concepts, and designs that were previously unimaginable. We can accelerate scientific discovery by teaching generative AI to make novel predictions about never-before-seen ideas, concepts, and designs,” says Buehler.

The open-access research, recently published in Machine Learning: Science and Technology, demonstrates an advanced AI method that integrates generative knowledge extraction, graph-based representation, and multimodal intelligent graph reasoning.

The work uses graphs developed using methods inspired by category theory as a central mechanism to teach the model to understand symbolic relationships in science. Category theory, a branch of mathematics that deals with abstract structures and relationships between them, provides a framework for understanding and unifying diverse systems through a focus on objects and their interactions, rather than their specific content. In category theory, systems are viewed in terms of objects (which could be anything, from numbers to more abstract entities like structures or processes) and morphisms (arrows or functions that define the relationships between these objects). By using this approach, Buehler was able to teach the AI model to systematically reason over complex scientific concepts and behaviors. The symbolic relationships introduced through morphisms make it clear that the AI isn’t simply drawing analogies, but is engaging in deeper reasoning that maps abstract structures across different domains.

Buehler used this new method to analyze a collection of 1,000 scientific papers about biological materials and turned them into a knowledge map in the form of a graph. The graph revealed how different pieces of information are connected and was able to find groups of related ideas and key points that link many concepts together.

“What’s really interesting is that the graph follows a scale-free nature, is highly connected, and can be used effectively for graph reasoning,” says Buehler. “In other words, we teach AI systems to think about graph-based data to help them build better world representations models and to enhance the ability to think and explore new ideas to enable discovery.”

Researchers can use this framework to answer complex questions, find gaps in current knowledge, suggest new designs for materials, and predict how materials might behave, and link concepts that had never been connected before.

The AI model found unexpected similarities between biological materials and “Symphony No. 9,” suggesting that both follow patterns of complexity. “Similar to how cells in biological materials interact in complex but organized ways to perform a function, Beethoven’s 9th symphony arranges musical notes and themes to create a complex but coherent musical experience,” says Buehler.

In another experiment, the graph-based AI model recommended creating a new biological material inspired by the abstract patterns found in Wassily Kandinsky’s painting, “Composition VII.” The AI suggested a new mycelium-based composite material. “The result of this material combines an innovative set of concepts that include a balance of chaos and order, adjustable property, porosity, mechanical strength, and complex patterned chemical functionality,” Buehler notes. By drawing inspiration from an abstract painting, the AI created a material that balances being strong and functional, while also being adaptable and capable of performing different roles. The application could lead to the development of innovative sustainable building materials, biodegradable alternatives to plastics, wearable technology, and even biomedical devices.

With this advanced AI model, scientists can draw insights from music, art, and technology to analyze data from these fields to identify hidden patterns that could spark a world of innovative possibilities for material design, research, and even music or visual art.

“Graph-based generative AI achieves a far higher degree of novelty, explorative of capacity and technical detail than conventional approaches, and establishes a widely useful framework for innovation by revealing hidden connections,” says Buehler. “This study not only contributes to the field of bio-inspired materials and mechanics, but also sets the stage for a future where interdisciplinary research powered by AI and knowledge graphs may become a tool of scientific and philosophical inquiry as we look to other future work.” 

Adobe Firefly Video Model: How AI is Changing the Future of Video Editing

Video production has always depended on human expertise and manual processes to create captivating content. From film studios to social media influencers, the need for high-quality videos has grown rapidly. However, with this demand comes the challenge of producing content quickly without losing quality. This is…

David Lee, COO of WEBTOON Entertainment – Interview Series

David J. Lee is Chief Operating Officer (COO) and Chief Financial Officer (CFO) for WEBTOON Entertainment, where he oversees global corporate, financial and operations teams for WEBTOON, Wattpad, and Wattpad WEBTOON Studios. Mr. Lee is also serving on the boards of WEBTOON Entertainment. A seasoned consumer…

Anchoreum: A New Game for Learning Anchor Positioning

You’ve played Flexbox Froggy before, right? Or maybe Grid Garden? They’re both absolute musts for learning the basics of modern CSS layout using Flexbox and CSS Grid. Thomas Park made those and he’s back with another game: Anchoreum.

Anchoreum: A New Game for Learning Anchor Positioning originally…

OpenAI faces diminishing returns with latest AI model

OpenAI is facing diminishing returns with its latest AI model while navigating the pressures of recent investments. According to The Information, OpenAI’s next AI model – codenamed Orion – is delivering smaller performance gains compared to its predecessors. In employee testing, Orion reportedly achieved the performance…

When muscles work out, they help neurons to grow, a new study shows

There’s no doubt that exercise does a body good. Regular activity not only strengthens muscles but can bolster our bones, blood vessels, and immune system.

Now, MIT engineers have found that exercise can also have benefits at the level of individual neurons. They observed that when muscles contract during exercise, they release a soup of biochemical signals called myokines. In the presence of these muscle-generated signals, neurons grew four times farther compared to neurons that were not exposed to myokines. These cellular-level experiments suggest that exercise can have a significant biochemical effect on nerve growth.

Surprisingly, the researchers also found that neurons respond not only to the biochemical signals of exercise but also to its physical impacts. The team observed that when neurons are repeatedly pulled back and forth, similarly to how muscles contract and expand during exercise, the neurons grow just as much as when they are exposed to a muscle’s myokines.

While previous studies have indicated a potential biochemical link between muscle activity and nerve growth, this study is the first to show that physical effects can be just as important, the researchers say. The results, which are published today in the journal Advanced Healthcare Materials, shed light on the connection between muscles and nerves during exercise, and could inform exercise-related therapies for repairing damaged and deteriorating nerves.

“Now that we know this muscle-nerve crosstalk exists, it can be useful for treating things like nerve injury, where communication between nerve and muscle is cut off,” says Ritu Raman, the Eugene Bell Career Development Assistant Professor of Mechanical Engineering at MIT. “Maybe if we stimulate the muscle, we could encourage the nerve to heal, and restore mobility to those who have lost it due to traumatic injury or neurodegenerative diseases.”

Raman is the senior author of the new study, which includes Angel Bu, Ferdows Afghah, Nicolas Castro, Maheera Bawa, Sonika Kohli, Karina Shah, and Brandon Rios of MIT’s Department of Mechanical Engineering, and Vincent Butty of MIT’s Koch Institute for Integrative Cancer Research.

Muscle talk

In 2023, Raman and her colleagues reported that they could restore mobility in mice that had experienced a traumatic muscle injury, by first implanting muscle tissue at the site of injury, then exercising the new tissue by stimulating it repeatedly with light. Over time, they found that the exercised graft helped mice to regain their motor function, reaching activity levels comparable to those of healthy mice.

When the researchers analyzed the graft itself, it appeared that regular exercise stimulated the grafted muscle to produce certain biochemical signals that are known to promote nerve and blood vessel growth.

“That was interesting because we always think that nerves control muscle, but we don’t think of muscles talking back to nerves,” Raman says. “So, we started to think stimulating muscle was encouraging nerve growth. And people replied that maybe that’s the case, but there’s hundreds of other cell types in an animal, and it’s really hard to prove that the nerve is growing more because of the muscle, rather than the immune system or something else playing a role.”

In their new study, the team set out to determine whether exercising muscles has any direct effect on how nerves grow, by focusing solely on muscle and nerve tissue. The researchers grew mouse muscle cells into long fibers that then fused to form a small sheet of mature muscle tissue about the size of a quarter.

The team genetically modified the muscle to contract in response to light. With this modification, the team could flash a light repeatedly, causing the muscle to squeeze in response, in a way that mimicked the act of exercise. Raman previously developed a novel gel mat on which to grow and exercise muscle tissue. The gel’s properties are such that it can support muscle tissue and prevent it from peeling away as the researchers stimulated the muscle to exercise.

The team then collected samples of the surrounding solution in which the muscle tissue was exercised, thinking that the solution should hold myokines, including growth factors, RNA, and a mix of other proteins.

“I would think of myokines as a biochemical soup of things that muscles secrete, some of which could be good for nerves and others that might have nothing to do with nerves,” Raman says. “Muscles are pretty much always secreting myokines, but when you exercise them, they make more.”

“Exercise as medicine”

The team transferred the myokine solution to a separate dish containing motor neurons — nerves found in the spinal cord that control muscles involved in voluntary movement. The researchers grew the neurons from stem cells derived from mice. As with the muscle tissue, the neurons were grown on a similar gel mat. After the neurons were exposed to the myokine mixture, the team observed that they quickly began to grow, four times faster than neurons that did not receive the biochemical solution.

Animation of green circles expanding during two stimulations: biochemical and mechanical.

MIT scientists find that motor neuron growth increased significantly over 5 days in response to biochemical and mechanical signals related to exercise. The green ball represents cluster of neurons that grow outward in long tails, or axons.

Credit: Angel Bu


“They grow much farther and faster, and the effect is pretty immediate,” Raman notes.

For a closer look at how neurons changed in response to the exercise-induced myokines, the team ran a genetic analysis, extracting RNA from the neurons to see whether the myokines induced any change in the expression of certain neuronal genes.

“We saw that many of the genes up-regulated in the exercise-stimulated neurons was not only related to neuron growth, but also neuron maturation, how well they talk to muscles and other nerves, and how mature the axons are,” Raman says. “Exercise seems to impact not just neuron growth but also how mature and well-functioning they are.”

The results suggest that biochemical effects of exercise can promote neuron growth. Then the group wondered: Could exercise’s purely physical impacts have a similar benefit?

“Neurons are physically attached to muscles, so they are also stretching and moving with the muscle,” Raman says. “We also wanted to see, even in the absence of biochemical cues from muscle, could we stretch the neurons back and forth, mimicking the mechanical forces (of exercise), and could that have an impact on growth as well?”

To answer this, the researchers grew a different set of motor neurons on a gel mat that they embedded with tiny magnets. They then used an external magnet to jiggle the mat — and the neurons — back and forth. In this way, they “exercised” the neurons, for 30 minutes a day. To their surprise, they found that this mechanical exercise stimulated the neurons to grow just as much as the myokine-induced neurons, growing significantly farther than neurons that received no form of exercise.

“That’s a good sign because it tells us both biochemical and physical effects of exercise are equally important,” Raman says.

Now that the group has shown that exercising muscle can promote nerve growth at the cellular level, they plan to study how targeted muscle stimulation can be used to grow and heal damaged nerves, and restore mobility for people who are living with a neurodegenerative disease such as ALS.

“This is just our first step toward understanding and controlling exercise as medicine,” Raman says. 

Admir Masic: Using lessons from the past to build a better future

As a teenager living in a small village in what was then Yugoslavia, Admir Masic witnessed the collapse of his home country and the outbreak of the Bosnian war. When his childhood home was destroyed by a tank, his family was forced to flee the violence, leaving their remaining possessions to enter a refugee camp in northern Croatia.

It was in Croatia that Masic found what he calls his “magic.”

“Chemistry really forcefully entered my life,” recalls Masic, who is now an associate professor in MIT’s Department of Civil and Environmental Engineering. “I’d leave school to go back to my refugee camp, and you could either play ping-pong or do chemistry homework, so I did a lot of homework, and I began to focus on the subject.”

Masic has never let go of his magic. Long after chemistry led him out of Croatia, he’s come to understand that the past holds crucial lessons for building a better future. That’s why he started the MIT Refugee Action Hub (now MIT Emerging Talent) to provide educational opportunities to students displaced by war. It’s also what led him to study ancient materials, whose secrets he believes have potential to solve some of the modern world’s most pressing problems.

“We’re leading this concept of paleo-inspired design: that there are some ideas behind these ancient materials that are useful today,” Masic says. “We should think of these materials as a source of valuable information that we can try to translate to today. These concepts have the potential to revolutionize how we think about these materials.”

One key research focus for Masic is cement. His lab is working on ways to transform the ubiquitous material into a carbon sink, a medium for energy storage, and more. Part of that work involves studying ancient Roman concrete, whose self-healing properties he has helped to illuminate.

At the core of each of Masic’s research endeavors is a desire to translate a better understanding of materials into improvements in how we make things around the world.

“Roman concrete to me is fascinating: It’s still standing after all this time and constantly repairing,” Masic says. “It’s clear there’s something special about this material, so what is it? Can we translate part of it into modern analogues? That’s what I love about MIT. We are put in a position to do cutting-edge research and then quickly translate that research into the real world. Impact for me is everything.”

Finding a purpose

Masic’s family fled to Croatia in 1992, just as he was set to begin high school. Despite excellent grades, Masic was told Bosnian refugees couldn’t enroll in the local school. It was only after a school psychologist advocated for Masic that he was allowed to sit in on classes as a nonmatriculating student.

Masic did his best to be a ghost in the back of classrooms, silently absorbing everything he could. But in one subject he stood out. Within six months of joining the school, in January of 1993, a teacher suggested Masic compete in a local chemistry competition.

“It was kind of the Olympiads of chemistry, and I won,” Masic recalls. “I literally floated onto the stage. It was this ‘Aha’ moment. I thought, ‘Oh my god, I’m good at chemistry!’”

In 1994, Masic’s parents immigrated to Germany in search of a better life, but he decided to stay behind to finish high school, moving into a friend’s basement and receiving food and support from local families as well as a group of volunteers from Italy.

“I just knew I had to stay,” Masic says. “With all the highs and lows of life to that point, I knew I had this talent and I had to make the most of it. I realized early on that knowledge was the one thing no one could take away from me.”

Masic continued competing in chemistry competitions — and continued winning. Eventually, after a change to a national law, the high school he was attending agreed to give him a diploma. With the help of the Italian volunteers, he moved to Italy to attend the University of Turin, where he entered a five-year joint program that earned him a master’s degree in inorganic chemistry. Masic stayed at the university for his PhD, where he studied parchment, a writing material that’s been used for centuries to record some of humanity’s most sacred texts.

With a classmate, Masic started a company that helped restore ancient documents. The work took him to Germany to work on a project studying the Dead Sea Scrolls, a set of manuscripts that date as far back as the third century BCE. In 2008, Masic joined the Max Planck Institute in Germany, where he also began to work with biological materials, studying water’s interaction with collagen at the nanoscale.

Through that work, Masic became an expert in Raman spectroscopy, a type of chemical imaging that uses lasers to record the vibrations of molecules without leaving a trace, which he still uses to characterize materials.

“Raman became a tool for me to contribute in the field of biological materials and bioinspired materials,” Masic says. “At the same time, I became the ‘Raman guy.’ It was a remarkable period for me professionally, as these tools provided unparalleled information and I published a lot of papers.”

After seven years at Max Planck, Masic joined the Department of Civil and Environmental Engineering (CEE) at MIT.

“At MIT, I felt I could truly be myself and define the research I wanted to do,” Masic says. “Especially in CEE, I could connect my work in heritage science and this tool, Raman spectroscopy, to tackle our society’s big challenges.”

From labs to the world

Raman spectroscopy is a relatively new approach to studying cement, a material that contributes significantly to carbon dioxide emissions worldwide. At MIT, Masic has explored ways cement could be used to store carbon dioxide and act as an energy-storing supercapacitor. He has also solved ancient mysteries about the lasting strength of ancient Roman concrete, with lessons for the $400 billion cement industry today.

“We really don’t think we should replace ordinary Portland cement completely, because it’s an extraordinary material that everyone knows how to work with, and industry produces so much of it. We need to introduce new functionalities into our concrete that will compensate for cement’s sustainability issues through avoided emissions,” Masic explains. “The concept we call ‘multifunctional concrete’ was inspired by our work with biological materials. Bones, for instance, sacrifice mechanical performance to be able to do things like self-healing and energy storage. That’s how you should imagine construction over next 10 years or 20 years. There could be concrete columns and walls that primarily offer support but also do things like store energy and continuously repair themselves.”

Masic’s work across academia and industry allows him to apply his multifunctional concrete research at scale. He serves as a co-director of the MIT ec3 hub, a principal investigator within MIT Concrete Sustainability Hub, and a co-founder and advisor at the technology development company DMAT.

“It’s great to be at the forefront of sustainability but also to be directly interacting with key industry players that can change the world,” Masic says. “What I appreciate about MIT is how you can engage in fundamental science and engineering while also translating that work into practical applications. The CSHub and ec3 hub are great examples of this. Industry is eager for us to develop solutions that they can help support.”

And Masic will never forget where he came from. He now lives in Somerville, Massachusetts, with his wife Emina, a fellow former refugee, and their son, Benjamin, and the family shares a deep commitment to supporting displaced and underserved communities. Seven years ago, Masic founded the MIT Refugee Action Hub (ReACT), which provides computer and data science education programs for refugees and displaced communities. Today thousands of refugees apply to the program every year, and graduates have gone on to successful careers at places like Microsoft and Meta. The ReACT program was absorbed by MIT’s Emerging Talent program earlier this year to further its reach.

“It’s really a life-changing experience for them,” Masic says. “It’s an amazing opportunity for MIT to nurture talented refugees around the world through this simple certification program. The more people we can involve, the more impact we will have on the lives of these truly underserved communities.”