Faster, more secure approach to machine learning helps safeguard private information

Keeping sensitive data safe has sometimes come at the expense of speed when training machines to perform automated tasks like biometric authentication and financial fraud detection. Now, Berkeley researchers have solved this issue by devising a practical way to keep data secure while training neural networks. In a study presented at the 2022 USENIX Security Symposium, Raluca Ada Popa, associate professor of electrical engineering and computer sciences, and her Ph.D. student, Jean-Luc Watson, described their innovative privacy-preserving approach to machine learning. They introduced a new platform, dubbed Piranha, that harnesses the speed of graphics processing units (GPUs) to train a realistic neural network on encrypted data for the first time.

Fellowship recipient aims to affect change in education for underrepresented students

Noor-Ul-Ain Ali, ’22, practically grew up in the shadow of the Campanile, in the neighboring east bay city of El Cerrito, and says that her first glimpse of what it might be like to go to college was as a participant in the University of California’s Early Academic Outreach Program, which assists underserved high school students with access to higher education. She didn’t even consider exploring STEM classes when she arrived at Berkeley as a first year student suffering from imposter syndrome, and she certainly had no plans to attend data science classes.

UC Berkeley researchers win $5M to launch center that advances decentralization technology

UC Berkeley and Imperial College London have won $5 million to create an international, interdisciplinary center aimed at advancing decentralization technology that offers users more control of their data, the Algorand Foundation announced today. The Berkeley Center for Responsible Decentralized Intelligence-Imperial London College Algorand Center of Excellence will – through research, education and events – help democratize access to data and ensure that data remains secure. These tools can be applied to digital ledgers like the blockchain Algorand created for cryptocurrency.

Just for you

Scrolling through content selected “just for you” on social media feeds seems like a harmless pastime. But new research shows that some algorithms go from recommending content that matches our preferences to recommending content that shapes our preferences in potentially harmful ways. In a study presented at the 2022 International Conference on Machine Learning, a UC Berkeley-led research team revealed that certain recommender systems try to manipulate user preferences, beliefs, mood and psychological state. In response, the researchers proposed a way for companies to choose algorithms that more closely follow a user’s natural preference evolution.

Craig Newmark gives Berkeley $2M for university cybersecurity clinics

Craig Newmark Philanthropies (CNP), the grantmaking organization launched by the founder of craigslist, has announced a three-year, $1.725M commitment to support the newly established Consortium of Cybersecurity Clinics, a network of organizations working to expand cybersecurity clinics at colleges and universities in the U.S. and around the world. Similar to clinics in schools of law and medicine, cybersecurity clinics provide students with hands-on experience as they work to strengthen the digital defenses of nonprofits, hospitals, municipalities, small businesses, small critical infrastructure providers, and other under-resourced organizations.

I School scholar shares what you need to know about data privacy

The U.S. constitution grants citizens a right to privacy. But in this new era where our lives increasingly include computers and data capture, legislators, regulators and others are grappling to define how the public’s right to privacy translates to these mediums. We spoke to Sophia Baik, a postdoctoral scholar at UC Berkeley School of Information’s Center for Long-Term Cybersecurity, about data privacy and why it matters to the public.

Jennifer Chayes elected as honorary member of London Mathematical Society

Jennifer Chayes has been elected an honorary member of the United Kingdom’s premier organization advancing, disseminating and promoting mathematical knowledge. Chayes, associate provost for UC Berkeley’s Division of Computing, Data Science, and Society, is one of two honorary members confirmed by the London Mathematical Society this year, the group said in its July 1 announcement. The society highlighted Chayes’s “fundamental contributions” to mathematics, computation and other related fields in its citation.

Workshop for data science educators welcomes national and international perspectives

UC Berkeley’s fifth annual National Workshop on Data Science Education brought together national and international perspectives on teaching undergraduate data science. Topics ranged from curriculum development to human contexts and ethics to student-led data science groups and programs. This was the first year the workshop was carried out in a hybrid format, with in-person presentations and training the first two days, online panels the next two days and technology demonstrations on the final day.

UC Berkeley and Tuskegee University announce data science partnership

Tuskegee University and UC Berkeley recently announced the Berkeley-Tuskegee Data Science Initiative, a multi-year partnership to develop curriculum and collaborative research opportunities for students and faculty at both institutions. Charlotte Morris, president of Tuskegee University, recently met with Berkeley Chancellor Carol Christ to discuss the new initiative. Faculty from both universities also discussed the partnership at the National Workshop on Data Science Education.

UC Berkeley named #1 online master’s in cybersecurity

The Master of Information and Cybersecurity (MICS) program at UC Berkeley ranked #1 among online cybersecurity master’s degree programs, according to an analysis by Fortune magazine released on June 30, 2022.