A paper released on arXiv last week by a team of researchers including Prof. Dawn Song and Ph.D. student Nicholas Carlini (B.A. CS/Math '13), reveals just how vulnerable deep learning is to information leakage.  The researchers labelled the problem “unintended memorization” and explained it happens if miscreants can access to the model’s code and apply a variety of search algorithms. That's not an unrealistic scenario considering the code for many models are available online, and it means that text messages, location histories, emails or medical data can be leaked.  The team doesn't “really know why neural networks memorize these secrets right now, ” Carlini says.  “At least in part, it is a direct response to the fact that we train neural networks by repeatedly showing them the same training inputs over and over and asking them to remember these facts."   The best way to avoid all problems is to never feed secrets as training data. But if it’s unavoidable then developers will have to apply differentially private learning mechanisms, to bolster security, Carlini concluded.

CS graduate student Ling-Qi Yan (advisors: Ravi Ramamoorthi/Ren Ng) and researchers at U.C. San Diego are the subject of an article in TechXplore titled "Scientists improve computer rendering of animal fur."  He is part of a team that developed a method for dramatically improving the way computers simulate fur, and more specifically, the way light bounces within an animal's pelt.  The researchers are using a neural network to apply the properties of a concept called subsurface scattering to quickly approximate how light bounces around fur fibers.  The neural network only needs to be trained with one scene before it can apply subsurface scattering to all the different scenes with which it is presented. This results in simulations running 10 times faster than current state of the art.  "We are converting the properties of subsurface scattering to fur fibers," said Yan. "There is no explicit physical or mathematical way to make this conversion. So we needed to use a neural network to connect these two different worlds."  The researchers recently presented their findings at the SIGGRAPH Asia conference in Thailand.

For the second year running, UC Berkeley has won the Fiesta Bowl Overwatch Collegiate National Championship, sweeping UC Irvine 3-0.  CS majors Kevin "SlurpeeThief" Royston and Gandira “Syeikh” Prahandika were on the Berkeley team, which battled in front of a sold-out crowd at the game in Tempe, Arizona, the first partnership between a collegiate bowl game and eSports tournament.  Royston, who was profiled in on Overwatch Wire article before the tournament, is in the top 1% of all players worldwide, has had a peak skill rating of 4626, and was on the winning team last year.  He notes that the balance between esports and school has been a tough road.  “It’s definitely rough, this week is the first time I had a homework assignment slip because of Overwatch. (Laughs) Hear me out! It’s for Machine Learning which is the hardest class at Berkeley. We had to travel (for the Fiesta Bowl) and I worked sixteen hours on an assignment and didn’t even get halfway through it.”  Overwatch is a team-based shooter game that was created by Blizzard Entertainment and released in 2016.  Teams are made up of six players, who select different characters, known in the game as “heroes,” to complete different objectives. The Berkeley team won $42,000 of the $100,000 total in scholarships and other prizes.

The  Cal Alumni Association and the Black Alumni Club are hosting an event Celebrating Diversity in Engineering and Science at Cal on February 10, 2018.  It will honor the 50th anniversary of the Black Engineering and Science Student Association (BESSA) and the 30th anniversary of the Black Graduate Engineering and Science Students (BGESS) group at Berkeley, and will feature both alumni and current students.  The keynote speaker is EECS alumnus Gary S. May (M.S. '88 and Ph.D. '92), now the Chancellor of the University of California, Davis.  In addition to the evening event, BESSA and BGESS alumni are planning an afternoon symposium with panels of engineering alumni to discuss their careers in industry and academia with undergrad, graduate and high school students

Profs. Dawn Song, Ion Stoica, Kurt Keutzer, Michael Jordan, Pieter Abbeel, and Trevor Darrell have teamed up with EECS alumnus Cameron Baradar (B.S. '15) and startup institute The House to run a new "global center-of-gravity of AI activity" called AI@The House. The new program will offer technical guidance, mentorship, free graphic processing units and financial support, among other resources, to startups focused on AI.  Their first core initiative is an accelerator for startups who are leveraging AI to build industry-defining products.

A new initiative, Graduates for Engaged and Extended Scholarship around Computing & Engineering (GEESE), aims to address growing concerns about the rapid advancement and integration of technologies in the global arena by building a coalition of engineers and social science scholars across campus to engage in issues vital to society and technology.  GEESE, launched this semester as one of CITRIS's Tech for Good initiatives,  plans to build a campus community of grad students and postdocs who will bring together disciplines and perspectives from fields like law, public policy, economics, anthropology, sociology, and philosophy, to promote cross-disciplinary scholarship on issues that cannot be wholly addressed from the silos of individual fields.  They will hold roundtables on relevant issues to gauge students' interests, and organize seminars with thought leaders to reflect and redefine their mission and acitivities.

All four EECS students nominated for this year's Computing Research Association (CRA) Undergraduate Researcher Award were recognized by the selection committee.  Senior CS/Math major Garrett Thomas (nominated by Pieter Abbeel) and EECS junior Peter Manohar (nominated by Alessandro Chiesa) were named as finalists.  Senior CS major Siqi Liu (nominated by Sanjam Garg) and CS/Statistics/Math senior Tianhe Yu (nominated by Sergey Levine and Pieter Abbeel) merited honorable mentions.  This award program recognizes undergraduate students in North American universities who show outstanding research potential in an area of computing research.

Math and L&S CS major Sameera Vemulapalli has been named Runner-up for the 2018 Alice T. Schafer Prize for Excellence in Mathematics by an Undergraduate Woman. The Schafer Prize is awarded annually by the Association for Women in Mathematics (AWM) to the most outstanding woman mathematics undergraduate in the United States.  Vemulpalli, who is currently finishing her senior year, was judged on  the quality of her performance in advanced mathematics courses and special programs,  her demonstrated real interest in mathematics ability for independent work in mathematics, and her performance in mathematical competitions at the local or national level.

EECS Master's student Allen Tang (also alumnus, B.A. CS/Statistics/ORMS) and his Berkeley teammates have won the Data Open Championship at the New York Stock Exchange.  The winners receive a $100,000 cash prize and possible job interviews with Citadel, a Chicago-based hedge fund firm. The competition was comprised of 20 one-day competitions from Stanford to MIT to Oxford, with the best performers competing in the week-long finale.  The Berkeley team of four applied data science to a meaningful problem in education--the impact of opening charter schools--to find where more funding would have the biggest effect. They worked 16-hour days during the week and produced a 20-page report and presentation on how charter schools have a negative impact in the short-term but outperform public schools in the long-term because of a survivorship bias. Only good charters stay in the system while bad ones close.

Prof. Ana Claudia Arias, Prof. Miki Lustig, and graduate student Joseph Corea, are featured in a STAT article titled "Electronics ‘like a second skin’ make wearables more practical and MRIs safer for kids."  The team is using printers loaded with a variety of high-tech inks (liquid silver nanoparticles, carbon nanotubes and semiconducting plastics) to make a new generation of medical devices, from wearables to barely noticeable MRI hardware for kids.  They have created light, flexible MRI coils that will improve image quality as well as patient comfort, and  have spun off a company called InkSpace Imaging to speed development.  “What would be best would be electronics that were almost like a second skin,” Arias said. “No adhesive. No straps. Almost like underwear — you forget that you’re wearing it.”