Brown CS News

Working to move LLM evaluations into the hands of the communities that actually use them, Victor will be part of a cohort dedicated to researching how we can program AI to use data that doesn’t breach our privacy.

It’s tempting, Adriana explains, to think that generative models will absorb the challenge of these complex, highly constrained spaces without the need for explicit optimization, but she argues the opposite: optimization is not merely still relevant, it’s becoming more powerful.

Brown CS community members are returning from the event with five accepted papers, and the work spans research areas ranging from activism to data maintenance to AI and mental health.

Brown CS is glad to announce that applications are open for the Randy F. Pausch '82 Computer Science Undergraduate Summer Research Award, which provides $15,000 annually to support an undergraduate engaged in an intensive faculty-student summer research partnership with the Department of Computer Science. 

It’s been a meteoric rise through CS academia for two Brown CS PhD alums this summer. Alexandra Papoutsaki has just become department chair at Pomona College’s computer science department and Danfeng (Daphne) Yao was appointed the computer science department head at NC State.

Maurice Herlihy is headed off to Paris on a fully funded, six-month research and teaching opportunity to explore distributed computing and finally master French verbs in the imperfect subjunctive voice.

Code analysis tools are crucial in preventing security vulnerabilities, but they’re difficult to develop and maintain because they must often deeply integrate with monolithic compilers. Enter: Malte’s work to “catalyze innovation” with compilers.

It focuses on transformorphic computing, a fundamentally new approach to automatically and correctly improving widely deployed systems without burdening their users and with strong guarantees about their behavior. 

In her solo paper, written when Kathi was a faculty member at Worcester Polytechnic Institute, she returns to and makes progress on Soloway’s Rainfall problem, often used to assess student programming skills in introductory courses. 

The ultimate goal is to produce new systems that enable practitioners to focus solely on creating and improving the accuracy of modern AI systems, without having to worry about details related to efficient execution and resource management.