News

EECS expands efforts to diversify professoriate by increasing retention of underrepresented undergraduates

The Diversifying LEAdership in the Professoriate (LEAP) Alliance (formerly called the FLIP Alliance), is one of the benefactors of a National Science Foundation (NSF) grant to the Center for Minorities and People with Disabilities in Computing and Information Technology (CMD-IT) to support the Broadening Participation in Computing Alliance (BPC-A).  UC Berkeley is a founding member of the LEAP Alliance, the goal of which is to increase diversity in the field of computing by expanding the number of professors from underrepresented communities at research Universities.  Diversifying the computing professoriate is critical to providing influential role models, shaping departmental programs and policies, and bringing diverse perspectives into research projects and programs.  As part of the first cohort, Berkeley has been partnering with 10 other institutions to focus on increasing the diversity of graduate student populations.  Thanks to their success, the new grant expands the Alliance to 4 cohorts, and Berkeley is now also part of Cohort 4, which is aimed at diversifying undergraduate student populations.  EECS representatives Prof. Armando Fox and Director of Diversity Audrey Sillers have started a mentoring program across institutions, participate in monthly cohort conference calls, attend many professional development events including two All Hands Meetings per year where cohort universities share best practices, and present what they have learned at the annual CMD-IT/ACM Richard Tapia Celebration of Diversity in Computing Conference.

Alt=""

Tsu-Jae King Liu wins 2021 IEEE EDS Education Award

EECS Prof. Tsu-Jae King Liu has been selected to receive the 2021 IEEE Electron Devices Society (EDS) Education Award.  This award is presented annually by EDS to honor "an individual who has made distinguished contributions to education within the field of interest of the Electron Devices Society."  Liu, who is currently the dean of Berkeley Engineering, was cited “For outstanding contributions to education in the field of electron devices and achievements on diversity and inclusion.”  She has been a strong advocate for fostering inclusion and respect for women and members of underrepresented minorities in engineering.  She was the first woman to Chair the EECS department (2014), the second woman to join Intel's board of directors (2016), and the first woman elected dean of the Berkeley College of Engineering (2018).  She won the Chang-Lin Tien Leadership in Education Award in 2020.   Liu is also renowned for her research into novel semiconductor devices, non-volatile memory devices, and M/NEMS technology for ultra-low power circuits.  She is probably best known for the development of polycrystalline silicon-germanium thin film technology for applications in integrated circuits and microsystems; and as the co-inventor of the three-dimensional FinFET transistor  which is the design that is used in all leading microprocessor chips today.

Kathy Yelick named UC Berkeley’s new vice chancellor for research

CS Prof. Katherine Yelick has been named UC Berkeley's next vice chancellor for research.  She will take over the role from EECS Prof. Randy Katz on January 1, 2022.  Yelick is an expert in the field of parallel computing and currently serves as executive associate dean in the Division of Computing, Data Science, and Society (CDSS).  “Kathy Yelick is one of the most talented leaders I have ever worked with — she listens, sees the big picture, and co-creates and implements phenomenal solutions,” said Jennifer Chayes, the CDSS Associate Provost. “I cannot imagine a better vice chancellor for research, and we at CDSS look forward to working with Kathy in her new role.” Yelick spent 11 years in leadership and management roles at Berkeley Lab (LBNL), where she oversaw a variety of initiatives, including the opening of new computing facility Shyh Wang Hall, the founding of the Berkeley Quantum collaboration, the formation of the lab’s machine learning for science initiative, and the launch of the U.S. Department of Energy’s Exascale Computing Project.  “UC Berkeley’s research community is uniquely positioned to tackle some of the world’s most important social and scientific problems, from climate change and public health to equity and social justice,” Yelick said. “I think it’s important to bring together diverse expertise and perspectives, and I look forward to collaborating with my colleagues across academic disciplines, from the humanities and social sciences to the physical and biological sciences, engineering, professional schools and beyond.”

"The Tale of a Success" with Ali Ghodsi

CS Prof. Ali Ghodsi will be the inaugural speaker for "The Tale of a Success" entrepreneurship series, hosted by the Iranian Students of California (ISC) in collaboration with the Berkeley Iranian Students Association in America (ISAA).  Ghodsi is a co-founder and the CEO of enterprise software company Databricks, a start-up which grew out of the AMPLab project that is now valued at $38B. He was one of the original creators of the open source project Apache Spark, and "the ideas from his academic research in resource management and scheduling and data caching have been applied to Apache Mesos and Apache Hadoop."  The lecture series features stories by successful Iranian-American entrepreneurs "who have all built category-defining tech companies."  Ghodsi will give his presentation via Zooom webinar on October 14th.

alt=""

Boubacar Kanté publishes paper introducing additional control knob for optical phase engineering

EECS Associate Prof. Boubacar Kanté is among the authors of a paper published in the journal Science titled "Plasmonic topological metasurface by encircling an exceptional point."  The paper introduces "an additional degree of freedom to address optical phase engineering by exploiting the topological features of non-Hermitian matrices operating near [the] singular points".   The novel phase, which was shown to be topologically protected, enables the construction of novel polarization dependent and chiral phased arrays and holograms. The ease of implementation together with its compatibility with other phase-addressing mechanisms will enable information multiplexing with antenna arrays.

EPIC Lab receives $2M NSF grant to build tools for criminal justice big datasets

CS Prof. Joseph Hellerstein, and Assistant Profs. Aditya Parmeswaran and Sarah Chasins, are among the principal investigators of a new lab that has just received a $2M grant from the National Science Foundation to make big datasets used by the criminal justice system more accessible to non-technical researchers.  The Effective Programming, Interaction, and Computation with Data (EPIC) Lab will create tools that utilize machine learning, program synthesis, and human-centered design, to improve the ability of public defenders, investigators and paralegals to research police misconduct, judicial decision-making, and related issues, for their cases.  The tools, which will initially be used in San Francisco, Alameda and Sacramento, are designed to address systemic power and resource disparities in California by helping under-resourced practitioners better defend their clients.

Gopala Anumanchipalli named Rose Hills Innovator

EECS Assistant Prof. Gopala Anumanchipalli has been selected for the Rose Hills Innovator Program which supports distinguished early-career UC Berkley faculty who are "interested in developing highly innovative research programs" in STEM fields.  The program will provide discretionary research support of up to $85,000 per year for "projects with an exceptionally high scientific promise that may generate significant follow-on funding."   Anumanchipalli's project, titled "Multimodal Intelligent Interfaces for Assistive Communication," proposes to "improve the current state of assistive communication technologies by integrating multiple neural and behavioral sensing modalities, and tightly integrating the graphical interfaces, and personalizing them to the user’s context."  His team will use "state-of-the-art neural engineering and artificial intelligence to develop novel communication interfaces" including Electrocorticography, non-invsive in-ear Electroencephalography sensors and functional near infrared spectroscopy.  They will also use on-device speech recognition and dialog management to incorporate the acoustic context of the user.

Sanjit Seshia wins Computer-Aided Verification Award

EECS Prof. Sanjit Seshia was a recipient of the CAV Award at the 2021 International Conference on Computer-Aided Verification (CAV) earlier this month.  This award is presented annually "for fundamental contributions to the field of Computer-Aided Verification," and comes with a cash prize of $10K that is shared equally among recipients.  This year's award specifically recognizes pioneering contributions to the foundations of the theory and practice of satisfiability modulo theories (SMT).”  Seshia's Ph.D. thesis work on the UCLID verifier and decision procedure helped lay the groundwork for this field.  SMT solvers are critical to verification of software and hardware model checking, symbolic execution, program verification, compiler verification, verifying cyber-physical systems, and program synthesis. Other applications include planning, biological modeling, database integrity, network security, scheduling, and automatic exploit generation.  CAV is the premier international conference on computer-aided verification and  provides a forum for a broad range of advanced research in areas ranging from model checking and automated theorem proving to testing, synthesis and related fields.

NSF awards $20M for researchers to launch National AI Institute for Advances in Optimization

A team of researchers from UC Berkeley, Georgia Tech, and USC, have been awarded $20M by the National Science Foundation (NSF) to launch an institute which will deploy AI to tackle massive optimization challenges.  The researchers hope the new National Artificial Intelligence (AI) Institute for Advances in Optimization will deliver a paradigm shift in automated decision-making by fusing AI and optimization to address grand challenges in highly constrained settings, such as logistics and supply chains, energy and sustainability, and circuit design and control.  EECS/IEOR Prof. Pieter Abbeel will lead the Reinforcement Learning Team, and EECS/IEOR Prof. Laurent El Ghaoui will be on both the End to End Optimization and the New Learning Methods Teams.  EECS Profs. Borivoje Nikolic and Vladimir Stojanovic will also be participating.  The group intends to integrate ethics and values into their complex systems design, from inception through operation, to ensure that all scientific advances will ultimately serve the interests of society.  The institute also plans to partner with historically Black colleges and universities (HBCUs) in Georgia, and Hispanic-serving community colleges in California, to build longitudinal education and workforce development programs.  Partners include Clark Atlanta University, Spelman College, and the University of Texas at Arlington.

New AI system allows legged robots to navigate unfamiliar terrain in real time

A new AI system, Rapid Motor Adaptation (RMA), enhances the ability of legged robots, without prior experience or calibration, to adapt to, and traverse, unfamiliar terrain in real time.  A test robot figured out how to walk on sand, mud, and tall grass, as well as piles of dirt, pebbles, and cement, in fractions of a second.  The project is part of an industry-academic collaboration with the Facebook AI Research (FAIR) group and the Berkeley AI Research (BAIR) lab that includes CS Prof. Jitendra Malik as Principal Investigator, his grad student Ashish Kumar as lead author, and alumnus Deepak Pathak (Ph.D. 2019, advisors: Trevor Darrell and Alexei Efros), now an assistant professor at Carnegie Mellon, among others.  RMA combines a base policy algorithm that uses reinforcement learning to teach the robot how to control its body, with an adaptation module that teaches the robot how to react based on how its body moves when it interacts with a new environment.  “Computer simulations are unlikely to capture everything,” said Kumar. “Our RMA-enabled robot shows strong adaptation performance to previously unseen environments and learns this adaptation entirely by interacting with its surroundings and learning from experience. That is new.”  RMA's base policy and adaptation module run asynchronously and at different frequencies so that it can operate reliably on a small onboard computer.