16 JULY2020|COMPUTER SHOPPER|ISSUE
Thepriceof
thisescape
routeisan
unprecedented
increase in digital
surveillance”
TheTony BlairInstitutefor
Global Changeisn’ttoo keen
on contact-tracing apps
It is highly
likelythat
malicious actors
will use AI to attack
the UK in numerous
ways,and the
intelligence
community will
need to develop
newAI-based
defence measures”
TheRusithink tank foresees
anew machine-based era of
cyberespionage
What better
wayof
plugging this gap
byengaging with
female candidates,
encouraging more
innovation and
generating as
yetuntapped
excitement for
IT in future
generations?”
Tim Sadler,CEO of Tessian,
on thecompany’s reportthat
found onlyone in sixUKIT
professionals are female
Microsoft
Edge+
Outlook=Better
together”
Microsoftis posting ads on its
email service to push peopleto
its newly revamped Edge
web browser
Intel’s neuromorphic computing system
learns like ahuman
ARTIFICIALINTELLIGENCE(AI)techhas
come alongway,but it’s still in its infancy,
and there’s alot ofwork ahead toperfect
the machine-learning algorithms and
techniques that sit behind an AI system.
These hurdles include processing time and
power consumption. Enter Intel and its
neuromorphic computing systems.
The chip maker has taken the covers off
its new Pohoiki Springs system, which aims
to solvethe two aforementioned problems.
Pohoiki Springs is different from
traditional systems in that its memory and
computing systems are intertwined, while
traditional systems separatethem. This eliminates
the shipping of data back and forth between the two
areas, dramatically speeding up processing times.
Forexample,Intel’s research team used just one
neuromorphic research chip to train an AI system to
detect ahazardous odour.This system worked out
how to detect each odour after just asingle training
sample,while the traditional deep-learning method
required 3,000 samples per odour.
How did Intel manage that? Well, the
neuromorphic computing systems learn much as
achild might, by permanently learning sensory-
based objects –those theycan see,smell, hear
or touch –afteronly one interaction, according to
Intel. What’s more,thissystem’s ability to learn
instantaneously allows it to make predictions more
accurately than atraditional deep-learning machine.
The PohoikiSprings system contains about 770
neuromorphic chips in achassis that’s no larger than
astandard server.This gives it the computational
capacity of about 100 million neurons, which is
similar to asmall mammal’s brain.
The by-product of doing everything faster and
more efficientlyisthatitrequires less power to
deliver the same results as adeep-learning machine.
Power consumption has been akey impediment for
large-scale AI development, as researchers at the
University of Massachusetts foundthat developing
just one AI model can have the same carbon
footprint as five cars over their entire lifetime.
FIGHTINGAVIRUSusually takes abit ofmedicine
andsome rest. But the Covid-19 pandemic has shown
the devastation anew virus can cause,which is why
Nvidia is joining the digital fight against coronavirus.
The graphics giant has joined the Covid 19
High-Performance Computing Consortium
to provide its expertise in AI and
large-scale computing optimisations to
use smart tech to find new ways to
take on the virus. Nvidia will also
provide the consortium with access to
its SaturnV supercomputer.
The Covid-19 Consortium is a
private-public effort led by theWhite
House Office of Science and Technology
Policy,the US Department of Energy and IBM. By
bringing together federal government, industry and
academic leaders, consortium members can provide
unfettered access to many of the world’s most
powerful high-performance computing resources to
support Covid-19 research.
This isn’t the first time Nvidia has stepped up to
the platetosupport efforts against the coronavirus.
With more than 27,000 Tesla V100 GPUs inside the
Summit supercomputer,Nvidiahas been diligently
working toward uncovering the virus’s largest spike
protein; this protein is what allows the coronavirus
to invade human cells and infect individuals.
Nvidia is also encouraging PC gamers to donate
unused GPU clock cycles towards the digital fight
against Covid-19.Since kicking off in February,
0,000 volunteers have joined the
Folding@home initiative,which Nvidia,
Intel, AMD and other PC-centric brands
are involved in (seeKay’s Corner,p10).
Muchlike the Covid-19 High
Performance Computing Consortium,
the Folding@homeinitiative looks to
generateinsights dug up from data the
ared GPU power is crunching through,
and use it to better understand diseases and
ways to treat them. And to help battle coronavirus
further,Nvidia has also made access to its GPU-
accelerated genome analysis toolkitParabricks
free for90daystoany researcher currently
working on combatting the virus.
The fight against coronavirus mayseem like
something forepidemiologists and clinicians, but
beyond the front lines, everyone is involved in this
crisis, from people self-isolating and governments
imposing lockdowns, to tech companies putting
their resources and expertise to good use.
Nvidia brings AI tech and high-performance
computing to the coronavirus fight
vid- 19
m
against
400
F
g
sh
d
