Cal bullish on B.E.A.R.S.
By Peggy Aycinena | February 18, 2011
If you were able to navigate the wind, rain and traffic by car to reach the U.C. Berkeley campus on Thursday morning, February 17th, you got to navigate yet more wind and rain on foot to struggle across campus to the 2011 Berkeley EECS Annual Research Symposium, otherwise known as B.E.A.R.S.
This year's gathering was held in the Chevron Auditorium at International House next to the football stadium, as opposed to previous years when B.E.A.R.S. was held in the Bechtel Engineering Center on the north side of campus. The good news? This year's SRO crowd included 475 people, whereas in previous years the SRO crowd maxed out at 250. The bad news? It was a longer walk from parking, half the audience arrived late, and all of the audience was soaking wet when they finally got there. Oh well.
This year's program – “Bringing Intelligence to Tomorrow's Client/Cloud Platform” – was organized by EECS Industry Liaison Dr. David Patterson, and featured four 30-minute talks, an hour-long panel discussion, and then lunch and open houses at 26 different EECS-associated labs around campus.
Prof. Ion Stoica went first. Talking about a Berkeley View of Big Data, he told his audience that the world's got a problem with players such as Google and Facebook generating PBs of data daily. Per Stoica, this stuff's expensive, comes from a variety of sources, is un-curated, without schema, suffers from inconsistent semantics, and poses a huge integration challenge. Hence, UCB's new AMP Lab would like to work towards creating a data center that serves as a “real computer,” one that runs smart algorithms and interfaces with even-smarter people to create a “holistically” integrated ecosystem of algorithms, machines, and people.
Prof. Dan Klein was next, and expounded on Natural Language Processing. Of course, he referenced IBM's Watson [who in the AI community isn't referencing Jeopardy this week?], and explained that conflating data with knowledge, ala Watson, is counter-productive and the bain of existence for people working in artificial intelligence. He said, nowadays search is about keywords, knowledge-free structure systems, and superficial patterns. But by 2020, it will be about entities and concepts, talking to embedded systems and mobile robots, and a very deep understanding of data in context.
Prof. Avideh Zakhor was third on the program and spoke about AR, Augmented Reality. Citing examples of heads-up displays in avionics, and scrimmage lines overlaid on NFL broadcasts, Dr. Zakhor explained that GPS capabilities for location will soon be enhanced with orientation information, as well, thanks to work going on in AIBLC – Automatic Image Based Loop Closure.
Finally, Sensor Swarms rounded out the morning's topics, with Prof. Kris Pister's talk outlining pending research in UCB's newly launched SWARM Lab. Headed by EECS Prof. Jan Rabaey, and funded to the tune of millions of dollars by Qualcomm, “the Swarm Lab creates an environment for the creation of the most advanced 'swarm nodes', exploring post-Moore technologies and manufacturing strategies combined with ultra-low power implementation fabrics and architectures for computation, communication, and energy provision. It also contains an experimental playground for the exploration of innovative and disruptive swarm applications. Multi-disciplinary in nature, the lab combines researchers from diverse backgrounds covering the complete spectrum from application to technology and materials.”
The speeches complete, 2 hours into the all-day B.E.A.R.S. event, there was still the morning panel, lunch, and lab tours to go. So there was a 30-minute break, then the 475+ chairs were re-filled and an hour-long discussion lead by Dr. Kurt Keutzer addressed: “Who owns the Future of the Mobile Client?”
As the panelists included folks from Verizon, Qualcomm, and Samsung, of course they had some pointed opinions on the topic. Unfortunately, whoever “owned” the sound system for the Chevron Auditorium needed to have taken better “ownership.” Little could be heard from the back of the room, short of …
- “Local computation will be much quicker if updated via the cloud.”
- “Using the cloud, I'm going to want my book shelf and my entire collection of music to follow me wherever I go.”
- “It's hard to add additional services, while still adequately supporting existing services.”
- “People who make hardware make almost no money. It's the apps today that drive the revenue.”
- “App stores are giant beta providers.”
- “The whole cloud network will be the ultimate playground for innovation.
Titan: Fastest Path to Mixed Signal Silicon
Mar Hershenson provides an overview of Titan and how it allows analog designers to more fully explore the design space to significantly improve performance and dramatically reduce power consumption on both new and existing analog designs.
With lithography limits, double patterning, high-k gates, low-k interconnects and myriad new requirements, the path to silicon success at the 20-nm node can look a little fuzzy. You need an advanced, scalable, integrated SoC design environment that delivers fast and predictable timing and layout closure. Attend this webcast to get a clear vision of how Magma's Silicon One technology can get your 20-nm SoC to silicon fast and with better results.