BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//University of California\, Berkeley//UCB Events Calendar//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
BEGIN:VTIMEZONE
TZID:America/Los_Angeles
BEGIN:STANDARD
TZOFFSETFROM:-0700
TZOFFSETTO:-0800
DTSTART:19701029T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:19700402T020000
TZOFFSETFROM:-0800
TZOFFSETTO:-0700
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20190301T231830Z
DTSTART;TZID=America/Los_Angeles:20190305T160000
DTEND;TZID=America/Los_Angeles:20190305T170000
TRANSP:OPAQUE
SUMMARY:Efficient Computational Methods for Complex Societal Systems
UID:124374-ucb-events-calendar@berkeley.edu
ORGANIZER;CN="UC Berkeley Calendar Network":
LOCATION:Wozniak Lounge (430) Soda Hall
DESCRIPTION:Somayeh Sojoudi\, Assistant Professor in Residence\, University of California\, Berkeley\n\nComputation plays a crucial role in the design\, analysis and operation of intelligent societal systems appearing in smart cities\, such as modernized power grids. This area requires significant advances in optimization\, control and learning techniques to be able to solve nonlinear real-world problems with a high accuracy at the scale and speed needed by emerging complex systems. We motivate the talk by discussing how advances in computation can revolutionize energy systems and then study two problems. First\, we design efficient global optimization techniques for power systems. We show how the passivity of the power infrastructure can be leveraged to design efficient computational methods for the real-time operation of power systems and demonstrate them on real-world grids with thousands of nodes. Then\, we study the problem of learning a model from data\, where the model could be statistical or physical. For statistical learning\, we show how to design low-complexity algorithms to find graphical models with over 400\,000 nodes in a few minutes. For physical learning\, we first study how much data is needed to learn an interconnected dynamical system and then design optimal distributed data-driven controllers. We also outline our contributions to the mathematics of learning through the notions of global functions and restricted isometry properties. The talk is concluded by mentioning our participation in the ARPA-E $4M cash prize competition on Grid Optimization and how different techniques from optimization theory\, numerical algorithms\, graph theory\, control theory\, and machine learning could be used for this purpose. \n\nBiography: Somayeh Sojoudi is an Assistant Professor in Residence in the Departments of Electrical Engineering & Computer Sciences and Mechanical Engineering at the University of California\, Berkeley. She is also on the faculty of the Tsinghua-Berkeley Shenzhen Institute (TBSI). She received her PhD degree in Control & Dynamical Systems from California Institute of Technology in 2013. She has been working on several interdisciplinary problems in optimization theory\, control theory\, machine learning\, and power systems. Somayeh Sojoudi is an Associate Editor for the journals of the IEEE Transactions on Smart Grid\, IEEE Access\, and Systems & Control Letters. She is also a member of the conference editorial board of the IEEE Control Systems Society. She has received the 2015 INFORMS Optimization Society Prize for Young Researchers and the 2016 INFORMS Energy Best Publication Award. She has been a finalist (as an advisor) for the Best Student Paper Award of the 2018 American Control Conference and a finalist (as a co-author) for the Best Student Paper Award of the 53rd IEEE Conference on Decision and Control. Her paper on graphical models has received the INFORMS 2018 Data Mining Best Paper Award. Her research has been supported by National Science Foundation\, Air Force\, Navy\, and Department of Energy.
URL:http://events.berkeley.edu/index.php/calendar/sn/pubaff.html?event_ID=124374&view=preview
SEQUENCE:0
CLASS:PUBLIC
CREATED:20190301T231830Z
LAST-MODIFIED:20190301T235417Z
X-MICROSOFT-CDO-BUSYSTATUS:BUSY
X-MICROSOFT-CDO-INSTTYPE:0
X-MICROSOFT-CDO-IMPORTANCE:1
X-MICROSOFT-CDO-OWNERAPPTID:-1
END:VEVENT
END:VCALENDAR