UC San Diego Launches PRLab
October 3, 2016
scientific method traditionally begins with a hypothesis, which is then
tested against data. Powerful new “brain-inspired” computing
capabilities are turning that idea on its head by accelerating a “data
science” experimental method -- a method that detects patterns in data
as a critical first step in generating a hypothesis.
“Pattern recognition is a mode of epistemology, a way of knowing,” says
University of California San Diego’s Larry Smarr, director of the
California Institute for Telecommunications and Information Technology
(Calit2). “It’s taking the same data that’s available to everyone and
trying to let the data talk to you instead of putting your preconceived
notions onto it.”
As both machine learning (ML) techniques and novel computer
architectures continue to rapidly develop, a major challenge is
emerging: how to optimize a variety of ML algorithms on different
architectures and discover which are fastest and most energy efficient
for specific applications across a wide range of disciplines.
Furthermore, there must be flexibility to both process massive static
arrays of data as well as myriad flows of data – and find the
never-before-seen patterns in both.
To explore these trade-offs, Calit2 has created a Pattern Recognition
Laboratory (PRLab), housed in Calit2’s Qualcomm Institute at UC San
Diego. The PRLab is in the early stages of building a “garden of
architectures” capable of performing massive amounts of high-speed
processing without consuming as much power as traditional chips.
UC San Diego Professor Ken Kreutz-Delgado, a long-time member of the
Electrical and Computer Engineering Department, is the PRLab’s first
director. Kreutz-Delgado is taking a broad view of the disciplines to
which pattern-recognition computing can be usefully applied.
“It isn’t just science and engineering problems, but also extends to
arenas in sociology, politics, economics… any discipline where data can
be collected and analyzed with models from the bottom up,” said Kreutz-Delgado.
Besides powerful traditional von Neumann architectures such as
shared-memory multi-core and graphics processing units (GPUs), the PRLab
has acquired non-von Neumann architectures such as high density Field
Programmable Gate Arrays (FPGAs), IBM’s TrueNorth neuromorphic
processor, and KnuEdge’s LambdaFabricTM neural computing systems.
The PRLab is the most recent development born from a decade-long
collaboration between Smarr and Mark Anderson. Anderson is the founder
and publisher of the widely-read Strategic News Service Global Report
and the Future in Review, or “FiRe,”Conference, which explores how
technology drives the global economy and has been described as “the best
technology conference in the world” by The Economist. Smarr and
Anderson, who often refer to Calit2 as “The “FiRe Lab,” jointly
developed the PRLab concept and invited Professor Kreutz-Delgado to FiRe
2015 to announce its formation. FiRe 2016 – which begins today -- will
explore in detail the “Power of Flows” and the necessity of pattern
computing for interpreting them. Mark is a member of the Calit2 Advisory
Board and Larry is a member of the Future in Review Advisory Board.
Both Anderson and Smarr liken pattern recognition to the methodology
they have used over the years to make predictions about the future of
science and technology.
“I’ve learned to be conscious of frames and filters, how they affect our
perceptions, and how to drop them in order to see the patterns of the
world objectively,” says Anderson. “Once you see present patterns
clearly, accurate future predictions become a matter of patterns made
and patterns broken. I believe the PRLab will begin to discover
computational approaches that can do this with Big Data Flows.”
Pattern Recognition Lab will also collaborate with the Pacific Research
Platform (PRP), a regional cyberinfrastructure, funded by the National
Science Foundation (NSF), that securely interconnects many dedicated
research networks at speeds up to 1,000 times that of the commodity
internet. This provides access to extremely large datasets of static and
streaming data whose analysis and interpretation is of keen interest to
scientists, engineers, businesses and policymakers.
“The Pattern Recognition Laboratory, connected to the Pacific Research
Platform, is an early model of how we can do distributed, brain-inspired
computing,” says QI research scientist Tom DeFanti, co-principal
investigator on the PRP. “This will enable people to attempt rapid
turnaround on significant computational problems that can only be
dreamed of today.”