C-BRIC: Center for Brain Inspired Computing

C-BRIC: Center for Brain Inspired Computing


– Good afternoon. My name
is Ragu Balakrishnan. I am professor at Michael
and Katherine Birck Head of the School of Electrical
and Computer Engineering. And on behalf of ECE,
it’s my distinct pleasure to welcome you all to this
afternoon’s celebration. As the single largest
academic unit at Purdue, ECE has played a significant role in many of the accomplishments
that Purdue has made. What we are all going to
celebrate this afternoon is one singular
accomplishment of leadership in both the educational
and research domains. You’ll hear a lot more about
it from our other speakers. But, now it’s my pleasure and honor to introduce the first
speaker of the afternoon, Mung Chiang, the John A.
Edwards Dean of Engineering. Dr. Chiang is formerly the
Arthur LeGrand Doty Professor at Princeton University. He’s a leader in networking research. A recipient of Guggenheim
Fellowship in 2014. And the 2013 Alan T. Waterman
award, the highest honor given to young US
scientists and engineers. He’s a co-founder of
multiple startup companies, and a pioneer in open, online
courses, reaching more than a quarter million students since 2012. He’s bringing an exciting
mission to Purdue, and we are very happy that he is leading
the College of Engineering. Please welcome Mung Chiang. (applause) – Thank you, Ragu. And thank
you all for coming over to warm up the MSEE atrium on
a somewhat chilly afternoon. Thank you for joining us on
this very special celebration occasion. It is special
for several reasons. It’s special because of
the subject matter of the center for brain-inspired computing. From hardware to software of computers and computing systems. This center will redefine
the frontier of research for intelligent, connected
autonomous systems and artificial intelligence applications. Indeed in engineers’ hand
we have now the opportunity to redefine the boundary
conditions of human existence. This celebration special also
because of the competitive nature of the award. We all know the electrical and
computer engineering school at Purdue has always been a premiere department in the nation. And in many areas, among the very best in the
world. This is the first time that Purdue ECE is the lead of a winning team, out of very competitive process. It also demonstrates, again, that the College of Engineering
at Purdue can, and we must, aim at the pinnacle of
research excellence. Finally, today’s a special
celebration also because of the scale and the
scope of collaboration. Behind the numbers,
like 36 million dollars, for five years, stand a very intensive and diverse ecosystem of partnership, including the private companies
behind the semiconductor research cooperation,
government-funding agency DARPA, Indiana State EDC, and cost-sharing from all
universities including Purdue. It is truly excellence at scale, and excellence through scale. On this exciting occasion, I’m so pleased and proud to bring on
stage my ECE colleague, and the new director of C-BRIC, indeed one more BRIC, one more very special BRIC
higher, Kaushik Roy. Kaushik. (applause) Thanks Mung. Imagine a world where you have your self-driving car waiting
for you to take you home. Once you go home, you
have your personal robotic assistant making coffee for
you, just the way you want it. Exactly the way you want
it. That’s the world we’re going into. If you really
think about it, the C-3-P-O s that George Lucas imagined,
that’s not just for George Lucas anymore. Things have changed. What we’d like to do is, interestingly we’re already
in a AI-centric world. Deep learning has transformed in the field of machine learning a lot. We all hear about deep learning all the time these days in the news. The kind of products we end up using today include web search, include image recognition,
video analysis in some cases, speech recognition, natural
language processing, and so on. But, that’s just the tip of the iceberg. The kind of applications that I’m talking about in the future, includes robotic assistants,
as I said, self-flying drones. To enable such applications,
there’s a need to consider beyond AI, to that effect,
we would be looking at what we call autonomous
intelligence. What autonomous autonomous intelligence is
that a computer is not only going to perceive the
environment around it, it’s gonna to reason about it, then
it’s gonna make a decision, and finally make some
actions. To that effect, our center, the center for
brain-inspired computing is going to develop
new algorithms that can potentially be very, very
close to what the brain does. We’ll be learning from the
brain. There’ll be implemented hardware, which is gonna
be also brain-like. From whatever we have
learned from neuroscience, it turns out that the architecture of the
brain is quite different from what the conventional
architectures that we used today. That’s Von Neumann-computing architecture. Von Neumann-computing
architecture, in general, we have the memory and the
processing unit separate. There’s something called
the memory bottleneck. That causes a whole lot
of problems in achieving the kind of performance
that you’d like to have. It turns out that in the
brain, memory is distributed and the computing, in some
ways, is done in the brain. We’ll be actually rethinking
architecture so that the algorithm that we
develop as part of the center can be implemented in such
architectures to really reduce the order of magnitude,
in an efficiency gap that exists today between the brain and the computing systems. Just
to give you and example of it, we are all familiar with
the fact that back in 2016, about year and a half ago, Google AlphaGo beat the Go Master for one. That was actually quite
achievement. But, the question is, at what cost? It turns out
that if you really look at the energy consumption of
the part consumption of the machine, that was about,
you know, 200-300,000 watts. Where as the brain, it
really does it in about 20 watts or so. It’s a huge
efficiency gap. To that effect, we’ll be really developing architectures, we’ll be developing
algorithms and the software, and we’ll be applying those
algorithms and architectures into application, which
are self-flying drones. You know personal robotic
assistants and so on. To do this, we have a team of
researchers from all across the United States, and
they’re the very best that we could find. They
include universities all the way from the west coast universities, like Portland State University, University of Southern
California, Arizona State. The east coast we have MIT, Princeton, Georgia Tech, Penn State, and in the
midwest we have, of course, our very own Purdue. We are really, really excited to do this research, and
in a collaborative fashion with our partner institutes,
and also with SRC and DARPA. You know, when we were
writing this proposal, this was a large proposal.
In order to really write the proposal, it’s almost required
the help of the village. We did get that help, in
turns out, from Purdue. We’d also like to thank the College of Engineering for really providing us all the help that we needed. In particular, actually, I
would like to thank Mung for all the support and enthusiasm
that he has shown. It’s been amazing. On
the other side of things, from the cost-sharing point of
view, you know Melba Crawford played a very important role.
So did Ragu Balakrishnan, our department head. [Christine
Chuel]. I’d also like to thank [Ling Dauman] for all
the help that he provided during our proposal writing.
We always ended up making the last minute changes. I found, that I never
met Kristine Sherman, and she’s here, and did the contract, budget for us. I’d also like to thank Jim Krogmeier and Robert Frosch for providing us the space. IEDC, that actually certainly played a very very important loan
in matching. They provided us the money to actually set
up an intelligence systems lab, which we’ll be actually
working on very soon. I’d also like to finally
thank [Christian Bobak], and his team, and also
[Christine Malaventa] for all the help during this process. [Suresh Garemelo], we had a
lot of discussions with him. I’m probably forgetting some people, I’m sure, but
during the contracting time, the contracting really did
take a long, long time. I’d like to also thank
Laura Gray and Ken Sandel. With that, we are really,
really excited and we look forward to doing some really exciting work in the near future. It’s my distinct honor now to introduce Purdue’s
President, Mitch Daniels. (applause) Probably the number one
joy of my job is the chance to come and celebrate the achievements. The spectacular achievements,
which are pretty routine at Purdue University these
days. But, there’s nothing routine about today. I want
to just start by saying to Dr. Roy, Ragu, to all your
colleagues, how deeply grateful we are. You have
brought an enormous honor, equally enormous
responsibility. But a tremendous honor and recognition to our University I’ll mention in just a second. I’m very personally interested, excited to watch this proposal move
forward, kept my fingers crossed. We’re all intrigued about
autonomous vehicles Kaushik. I’ve said for some time,
I want them to perfect the autonomous car one day after they take my driver’s license away. So,
speed up, ’cause that might not be too far off, I don’t know. I also must confess that
when I try to play Go, it uses many more than 20 watts. (laughing) We’ve all been reading,
every single day, about artificial intelligence,
and all the ways it’s likely to change our world. All
the possibilities that it will surface. Much of
this work is already going on at Purdue. But, now
in a very very visible, and very important way.
Some of the most important and noteworthy research
is gonna happen here. I hope everyone already
got the sense of what an achievement this is.
Just a few months ago, Purdue was chosen by the NSF
for an energy research center. It was the first time in 40 years. We celebrated that, as well we should. This is the first time ever, most people here I suppose know what the consortium is all about. The national semi-conductor consortium, and has a sense about what the defense advance research projects agency, DARPA, is about. I can tell
you that I have, in past lives, have worked with both of them. Every grant our people, our faculty, compete for here is a very fair, and tough competition I’m sure. But there are no competitions tougher than the ones that
Kaushik and his colleagues have just succeeded in. These are very unsentimental people. They are interested in the
hardest, practical realites they expect to see. On behalf
of the great corporations who pool together to create SRC. Certainly, I can tell you on
behalf of the people at DARPA, whose responsibility is the
national security of us all, in a dangerous world. They are only, they will
only award their projects to the very, very finest and
most promising researchers and institutions. We are
deeply honored that you have brought this achievement to us. I can’t underscore it enough. Finally, I want to thank
our friends from the state. I’m constrained to mention,
in my previous employment the first action of the first
day was to create something called the Indiana Economic
Development Corporation. A public/private corporation,
later enacted in statute. On that day, we said that
our objective was to make Indiana, to lift Indiana
to be one of the most attractive places on the
planet for people to invest and to create new learning
and create new opportunities. They’ve done a great job
in that organization over the ensuing decade. Their
contribution, their support their recognition of the
importance of research like this and Purdue as a world leader
in it, was instrumental, I think, in the success, and
we want to say thanks to them. Finally, I’ll just say that, the whole university, I hope, will pay close attention to this work, and will not only join in celebrating it, But AI, in ways that Ragu and Kaushik just
presaged a little bit, is gonna change our world
in ways we cannot imagine. Most I’m sure will be for
the infinite good of mankind. But, there may also be some very serious downsides to guard against and consider. The thought that the front edge of so much of this will be
happening right here, I think, gives us a second responsibility,
which is for those of us who will never be capable
of contributing directly to the work of C-BRIC to
pay close attention to it, and to think about all the
ways in which we make sure that this is an unmitigated
good for all of humanity. That any problems or dangers
that it brings with it are managed effectively. What
a wonderful set of missions for our institution. All
because of the folks who made C-BRIC possible. I’ve rattled
on only because I’m so very excited about this and have
been so hopeful that this day would come, and now
for the next several years, let’s do the very best work
of which Purdue is capable. Show those tough-minded
people who have selected us that they made exactly the
right choice. Thank you. (applause) – See, I’m back here
again. When I was talking, and thanking, I said that I
must be forgetting some people. Indeed, I did. It’s the
most important people. That turns out to be Anand Raghunathan. Without him, the center
would not have been possible. Absolutely not possible. We spent endless nights
discussing the center. Our families were really upset about it. Thank you very much.
Without Anand It would not have been possible.
Finally, I also wanted to thank my students. We spent,
again, many sleepless nights discussing and getting the writeups done. Thank you very much. I knew I’d forgotten some of
the most important people. Thank you again. (applause) – Well thank you Kaushik, and
thank you President Mitch, Dean Mung. Thank you all for
joining us this afternoon, and helping us celebrate
the launch of C-BRIC. A great jewel. One more
jewel in Purdue’s crown.

Leave a Reply

Your email address will not be published. Required fields are marked *