Johns Hopkins University Computer Science

06/12/2026

The FDA has approved Bayesian Health’s AI-powered early warning system to detect sepsis, one of the deadliest infections for hospital patients.

“FDA approval is a regulatory first that shifts what the standard of care can be for a condition associated with roughly 1 in 3 in-hospital deaths," says lead researcher Suchi Saria, an associate professor of computer science at Johns Hopkins University. "This represents decades of clinical AI research at Johns Hopkins translated into practice—not just models built in the lab, but technology delivered where it matters: at the bedside.”

Every hour that sepsis detection is delayed significantly reduces a patient's chance for survival

06/11/2026

“I have confidence that if you want to talk to an AI without anyone else seeing your conversation, including Meta, this will do the job,” says Johns Hopkins University’s Matthew Green.

The company says its new Incognito Chat allows you to use its AI chatbot without anyone else—including Meta—being able to access your conversations.

06/11/2026

“The Canvas breach is a reminder that no platform is immune: There are countless widely used systems that remain attractive targets for sophisticated bad actors, including nation-states,” says JHU CS’ Anton Dahbura, the executive director of the Johns Hopkins University Information Security Institute. “Educational platforms are particularly rich targets given the concentration of personal, financial, and international student data.”

‘PAY OR LEAK’: Hackers Target Big Higher Ed Vendor

The criminal extortion group ShinyHunters breached Instructure last week. The hackers, who have also attacked individual universities, demanded the ed-tech giant pay up or face a data leak. https://bit.ly/4nah8Cm

06/11/2026

Congratulations, Oliver!

Meet Hopkins Engineering's 2026 Goldwater Scholars—one of the oldest and most prestigious national scholarships in the United States!

Henry Le Chang, a neuroscience and applied mathematics and statistics major, plans to pursue a PhD in biomedical sciences focused on chronic pain and itch. He works in Prof Xinzhong Dong's lab and leads an independent project on bacterial meningitis as a University Undergraduate Research Fellow. He is also active in Brazilian Jiu Jitsu, Nu Rho Psi, the Immunology and Immunoengineering Club, and Dragon Boat Club.

Roma Desai, majoring in biomedical engineering, plans to pursue a PhD in biomedical engineering to conduct translational tissue engineering research and contribute to safer and more reliable laboratory research models. She has researched with Prof Deok-Ho Kim, Asst Prof John Miller, and Assoc Prof Nicholas Durr on drug-testing models, Lyme disease, and point-of-care diagnostics. She also holds leadership roles with the Biomedical Engineering Society, Maryland Science Olympiad, and Med Tech Network, and serves on the JHU Student Conduct Board.

Sameer Gabbita is majoring in biomedical engineering and plans to pursue an MD-PhD studying gene expression and disease. He works in Tej Azad’s lab on artificial intelligence applications for neurocritical care. He is president of the Hopkins AI Society, volunteers with the Johns Hopkins Outpatient Center and Crisis Text Line, mentors student researchers, and co-founded Baltimore Benefits.

Oliver Nizet is majoring in chemical and biomolecular engineering and computer science and plans to pursue a PhD in bioengineering focused on cancer and infectious disease therapeutics. He works in Prof Denis Wirtz’s lab on gynecologic and prostate cancer research. He also serves as a PILOT peer leader and an elementary school tutor with the Johns Hopkins Tutorial Project.

This year, 454 Scholars were selected from a pool of more than 5,000 college sophomores and juniors demonstrating exceptional promise in the natural sciences, engineering, and mathematics. Each Goldwater Scholar receives up to $7,500 toward the cost of tuition, mandatory fees, books, and room and board. Sophomore recipients receive a second year of funding.

Established by Congress in 1986 to honor the legacy of soldier and statesman Barry Goldwater, the Goldwater Scholarship is one of the earliest significant national scholarships focusing on STEM fields.

06/10/2026

It’s always a par-tee 🎉⛳️ with the CS staff! 🥰 We had a great 2026 retreat at Topgolf Baltimore!

06/10/2026

New work by Johns Hopkins University computer scientists proves that in-context learning is not tied to language.

Their work, which proves that in-context learning is not tied to language, appears in Transactions on Machine Learning Research.

06/09/2026

Yinzhi Cao, the technical director of the Johns Hopkins University Information Security Institute and an associate professor of computer science, says universities can increase protections from attacks by raising cyber awareness, providing only necessary information to third parties such as Canvas, and conducting routine internal tests to ensure that everything is secure.

A ransomware attack disrupted Canvas, an online platform used by schools and colleges across Maryland to manage grades, assignments and classwork, with hackers claiming access to billions of records.

06/09/2026

The AI system detects deadly infections faster than doctors, saving thousands of lives from a condition that claims more than 250,000 lives each year in the U.S.

06/08/2026

“Using the same username across platforms can make it easier for hackers to conceptually link a person to multiple accounts, although it’s not guaranteed,” says JHU CS’ Anton Dahbura, the executive director of the Johns Hopkins University Information Security Institute.

He emphasizes that password practices, not usernames, are far more consequential in real-world attacks: “Common password hashing across different platforms combined with dictionary attacks on passwords would seem to be a more meaningful attack. People should not use passwords that have been successfully cracked elsewhere, and should not use the same password for different platforms.”

Experts say common online habits, such as reusing identifiers like "spiderkash," can leave even senior government officials vulnerable to exposure.

06/08/2026

“We’re able to identify [sepsis cases] early, as a continuous monitor, and we’re doing this in a two, three, four percent prevalence population,” says JHU CS and Bayesian Health’s Suchi Saria. “There is no other solution today that exists that has lead time.”

As new AI models for predicting sepsis hit the market, experts say performance alone will not drive adoption by hospitals.