EECS professor and UC Berkeley's new Vice Chair for Research, Randy Katz, is interviewed in Cal Alumni's California Magazine about his approach to his new job. The article covers how one might go about creating a nurturing environment for pursuing innovative research, his predictions about future technologies, the integration of Big Data in new research, examples of some exciting projects, and the problem of funding.
A research team led by Assistant Prof. Rikky Muller and Prof. Michel Maharbiz have created StimDust (stimulating neural dust), the smallest volume, most efficient wireless nerve stimulator to date. The innovation adds more sophisticated electronics to neural dust (tiny, wireless sensors first implanted by Maharbiz and Prof. Jose Carmena in 2016) without sacrificing the technology’s size or safety, greatly expanding its range of applications. Powered by ultrasound at an efficiency of 82%, and with a volume of 6.5 cubic millimeters, StimDust can be used to monitor and treat disease in a real-time, patient-specific approach. “StimDust is the smallest deep-tissue stimulator that we are aware of that’s capable of stimulating almost all of the major therapeutic targets in the peripheral nervous system,” said Muller. “This device represents our vision of having tiny devices that can be implanted in minimally invasive ways to modulate or stimulate the peripheral nervous system, which has been shown to be efficacious in treating a number of diseases.” The research will be presented April 10 at the IEEE Custom Integrated Circuits Conference in San Diego.
Prof. Stephen Derenzo is quoted in an article for Australia’s Particle about a new material for a proposed detector of weakly interactive massive particles (WIMPs). Derenzo is the lead author of a study published March 20 in the Journal of Applied Physics about a crystal called gallium arsenide (GaAs) which features added concentrations, or “dopants,” of silicon and boron. This material possesses a scintillation property--it lights up in particle interactions that knock away electrons. According to Derenzo, who is a senior physicist in the Molecular Biophysics and Integrated Bioimaging Division at Berkeley Lab, the new ultrasensitive detector technology could scan for dark matter signals at energies thousands of times lower than those measurable by more conventional WIMP detectors. “It’s a privilege to be working on such an important problem in physics, but the celebration will have to wait until clear signals are seen,” he says. “It’s possible that dark matter particles are even lighter than what we can see with GaAs, and their discovery will have to wait for even more sensitive experiments.”
At the ASPLOS conference in late March, John Kubitowicz and his group from 2000 were celebrated for their paper, "OceanStore: an architecture for global-scale persistent storage." The paper was named Most Influential Paper 2018, and the authors receiving the award included David Bindel, Yan Chen, Steven Czerwinski, Patrick Eaton, Dennis Geels, Ramakrishna Gummadi, Sean Rhea, Hakim Weatherspoon, Chris Wells, and Ben Zhao, as well as Kubi, a long-time Berkeley CS faculty member. The paper was originally published in the Proceedings of the ninth international conference on Architectural support for programming languages and operating systems (ASPLOS IX).
The Foundations of Data Science (Data 8X), which is being offered free online this spring for the first time through the campus’s online education hub, edX, is the fastest-growing course in UC Berkeley’s history. Taught by Prof. David Wagner, Assistant Teaching Prof. John DeNero (recipient of the 2018 Distinguished Teaching Award), and a statistics professor, Data 8X is based on CS C8: Foundations of Data Science and now has more than 1,000 students enrolling every semester. “You’ll learn to program when studying data science — but not for the primary purpose of building apps or games,” says DeNero. “Instead, we use programming to understand the world around us.”
The Association of Computing Machinery (ACM) announced today that the winners of the 2017 ACM Turing Award are UC Berkeley's David A. Patterson and Stanford University's John L. Hennessy for "pioneering a systematic, quantitative approach to the design and evaluation of computer architectures with enduring impact on the microprocessor industry. Hennessy and Patterson created a systematic and quantitative approach to designing faster, lower power, and reduced instruction set computer (RISC) microprocessors." (ACM) Since then, computer architects have been using principles derived from their approach in a wide variety of projects for industry and academia. Today, 99% of the more than 16 billion microprocessors produced annually are RISC processors, to be found in most smartphones, tablets, and the billions of embedded devices that comprise the Internet of Things (IoT).
Assistant Teaching Prof. John DeNero has won the UC Berkeley Distinguished Teaching Award. The award, presented by the Academic Senate, recognizes U.C. Berkeley's brightest teaching stars for their inspiring and transformational teaching. DeNero says his teaching goal is not necessarily to make students happy but to help them learn how to solve problems that they thought they couldn't solve. He has a knack for grabbing attention, exciting students, and in many ways, serving as a pioneer. He teaches his introductory course for computer science majors, CS 61A, to nearly 1,600 students in 47 sections with the help of a course staff of 95 undergraduates. Distinguished Teaching Award winners are frequently called upon by the campus to provide a voice on issues related to teaching. They serve on forums, panels, and committees involving teaching issues, and they are advocates for excellence in teaching at Berkeley.
Prof. Michael Jordan has been named a Plenary Lecturer at the 2018 International Congress of Mathematicians (ICM), which will take place in Rio de Janeiro, Brazil, in August. ICM is considered the world’s premier forum for presenting and discussing new mathematical discoveries. Plenary speakers are invited from around the world to present one-hour lectures which are held without other parallel activities--an honor that has been bestowed on only a small handful of computer scientists over the 121 year history of the ICM.
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.
Robotics researchers from Berkeley's AUTOLab, led by IEOR and EECS professor Ken Goldberg, have built a heaving robotic platform — mimicking the motion of a breathing, heart-beating human patient — to help develop algorithms that robotic surgical assistants can use to guide their cutting. This research is the subject of an article in Wired magazine titled "How Flight Simulation Tech Can Help Turn Robots Into Surgeons." During surgery, when the chest heaves or blood pumps, the surgeon has to compensate for that movement. The researchers took the data from watching the surgeon's movements and developed algorithms that could mimic his strategy for cutting along a line. This new robot, which is a kind of a Stewart platform, mimics that movement. Stewart platforms are normally hefty pneumatic devices that power things like immersive flight simulators. But for this study, the researchers took the concept and shrunk it down to a 6-inch-wide device, opting for servo motors instead of pneumatic power. The machine costs just $250.