Recent News & Accomplishments
2024
As large language models (LLMs) continue to take on complex tasks previously done by humans—analyzing RNA for vaccines, writing software code, generating news articles, and much more—how does the technology fare at preventing cyberattacks on critical infrastructures like financial institutions or energy grids? LLMs are already being used to launch cyberattacks, with cybercriminals exploiting malicious inputs to LLMs to generate malware, create phishing emails and phishing sites. Other AI-infused technologies, known as LLM agents, are even more powerful than just LLMs alone, with the large... read more
Ph.D. student John Kirchenbauer and his team at the University of Maryland are among those advancing watermarking technology to identify AI-generated text, building on prior work in the field.
Seldom has a tool or technology erupted so suddenly from the research world and into the public consciousness — and widespread use — as generative artificial intelligence (AI). The ability of large language models (LLMs) to create text and images almost indistinguishable from those created by humans is disrupting, if not revolutionizing, countless fields of human activity. Yet the potential for misuse is already manifest, from academic plagiarism to the mass generation of misinformation. The fear is that AI is developing so rapidly that, without guard rails, it could soon be too late to... read more
Leveraging emotion recognition technology, Aayush Gupta, Akash Moorching and Eric Chen's platform converts live-stream highlights into short-form social media content.
In the world of streaming, where around seven million creators vie for viewer attention across platforms, the ability to repurpose content efficiently can make the difference between obscurity and viral success. For Twitch streamers, translating hours of live footage into digestible, shareable clips is daunting. A group of University of Maryland students has taken on this challenge with their startup, Saved . It uses machine learning to identify and repurpose key moments from streams, turning them into ready-to-post highlights for TikTok, Instagram and YouTube Shorts. Saved, one of four teams... read more
In this era of artificial intelligence (AI), software developers increasingly rely on large language models (LLMs) like ChatGPT and GitHub Copilot to streamline their coding processes. Companies like Meta have reported that nearly all their developers use an internal LLM to enhance productivity, highlighting the essential role AI now plays in software development. Despite making huge strides, AI’s capabilities in high-performance computing (HPC)—which involves executing complex parallel programs and processing massive datasets across hundreds to thousands of GPUs—still face limitations. Two... read more
Computer science major Rodrigo Sandon and team member Marie Brodsky use AI-driven insights to bridge communication gaps in large lecture settings.
At the University of Maryland, computer science major Rodrigo Sandon (B.S. ’24, computer science) and alumna Marie Brodsky (B.S. ’24, mathematics) have developed a solution to address a longstanding challenge: enhancing communication between students and instructors in large lecture halls. Their startup, Engauge , is designed to give students a platform for real-time feedback, helping instructors better gauge understanding in a large group setting. Currently participating in the Mokhtarzada Hatchery program, Engauge aims to expand its reach on campus and potentially to other institutions. The... read more
In the race for constant growth, many large language models (LLMs) focus on producing newer and better outputs. Sathvik Nair, a third-year linguistics doctoral student working in the Computational Linguistics and Information Processing (CLIP) Lab at the University of Maryland is taking different approach, examining how this technology operates and what it reveals about human cognition. Nair is deeply fascinated by the human mind. His research involves understanding and explaining human language processing and usage through advances in LLMs. By integrating advances in natural language... read more
The world's largest hackathon for underrepresented genders marks a decade of innovation at the University of Maryland.
The Reckord Armory at the University of Maryland transformed into a bustling hub of creativity on October 26 and 27, 2024, as Technica celebrated its 10th anniversary. The event, renowned as the world's largest hackathon for underrepresented genders, drew participants from around the region to collaborate, innovate and showcase their technological skills. Since its inception in 2015, Technica has grown from a student-run event into a significant platform for promoting diversity and inclusion in tech. This year, the hackathon attracted over 1,000 participants with the support of 22 sponsors... read more
Environmentally friendly, sustainable pesticides are popular with consumers, who are more likely to purchase sweet corn from their local market that hasn’t been sprayed with potentially harmful chemicals. Now, supported by a $650,000 grant from the U.S. Department of Agriculture (USDA), two University of Maryland researchers have teamed up to help ensure these pest-fighting methods remain on the table, in part by understanding why a common pest known as the corn earworm, or H. zea , has developed resistance to pesticides widely used in farming. The researchers— Megan Fritz , an associate... read more
University of Maryland alums, students, faculty and staff met up at the world’s largest conference for women and nonbinary technologists.
In the last 10 years, the number of undergraduate women in computing majors at the University of Maryland climbed from less than 100 to more than 2,400. This growth extends to graduate students, where women now make up 38% of graduate students in computing programs. As these numbers increased, so too has UMD’s community of women in tech. In October 2024, UMD convened that community of alums, students and staff at the Grace Hopper Celebration of Women in Computing (GHC)—a conference that brings the research and career interests of women in computing to the forefront. The Department of Computer... read more
James Purtilo emphasizes how the CrowdStrike incident reveals national security vulnerabilities, with attackers using AI to outpace cybersecurity defenses.
This past summer’s “CrowdStrike incident” – which began as a “minor” error – came as close to a total system crash as the world has yet seen. It grounded airlines, hospitals were forced to cancel surgeries, and retailers couldn’t process digital payments. The issue was made worse because though a patch was quickly released, it required every single computer impacted to be manually restarted. However, just months later there are signs that it was a wakeup call for IT and cybersecurity specialists. Instead of ignoring the issue and hoping it was an isolated incident – as has happened following... read more