Latest Accenture Report came out towards the end of Nov.
Unleashing the full potential of AI
The semiconductor industry is key to unlocking the full potential of #AI. Semiconductor manufacturers have the opportunity to play a leading role in driving the next wave of computing. See how:
www.accenture.com
Not a bad read for a general overview on their assessment of AI and Semi Industry.
Quite liked this image they used which appears borrowed from a 2015 paper.
The U.S. Department of Energy's Office of Scientific and Technical Information
www.osti.gov
A table I also found interesting was the costs.
On their web I found a small research section where they did their own POC voice control in 2021:
NEUROMORPHIC COMPUTING: ENERGY-EFFICIENT SMART CARS WITH ADVANCED VOICE CONTROL
Was based on Intel but hey, I don't mind if they blaze away generating industry interest for us to swoop in as the best available for immediate uptake
Every organization needs to shape its computational variety strategy to meet growing demands from consumers. Learn how to stay ahead of increasing competition.
www.accenture.com
Speaker 1: Tim Shea, a research scientist at
Accenture Labs, has been working with an
automotive client on prototyping the use of
neuromorphic computing to help people interact
with smart vehicles.
Speaker 2: Consumer demand for AI driven
experiences is increasing rapidly, especially in
the automotive industry. Customers expect
responsive voice, gesture, and contextual
intelligence from their vehicles. But current AI
hardware is too power hungry, which can impact
vehicle performance and limit the possible
applications. Smart vehicles need more efficient
edge AI devices to meet the demand.
Using edge AI devices to compliment cloud
based AI could also increase responsiveness
and improve reliability when connectivity is poor.
So we've built a proof of concept system with
one of our major automotive partners to
demonstrate that neuromorphic computing can
make cars smarter without draining the
batteries. We're using Intel's Kapoho Bay to
recognize voice commands that an owner would
give to their vehicle. The Kapoho Bay is a
portable and extremely efficient neuromorphic
research device for AI at the edge.
We're comparing that proof of concept system
against a standard approach using a GPU. To
build the system, we trained spiking neural
networks to differentiate between command
phrases. Then we ran the trains networks on
the Kapoho Bay. We connected the Kapoho
Bay to a microphone and a controller similar to
the electronic control units that operate various
functions of a smart vehicle.
We're targeting commands that reflect features
that can be accessed from outside a smart
vehicle, such as park here, or unlock passenger
door. These functions also need to be energy
efficient, so the vehicle can remain responsive
even when parked for long stretches of time.
As a first step, we trained the system to
recognize simple commands, such as lights on
and lights off, open door, close door, or start
engine. Using a combination of open source
voice recordings and a smaller sample of
specific commands, we can approximate the
kinds of voice processing needed for smart
vehicles. We tested this approach by comparing
our train spiking neural networks running on
Intel's neuromorphic research cloud against a
convolutional neural network, running on a
GPU.
Both systems achieved acceptable accuracy
recognizing our voice commands, but we found
that the neuromorphic system was up to a
thousand times more efficient than the standard
AI system with a GPU. This is extremely
impressive and it's consistent with the results
from other labs, as Intel will show further in their
session on benchmarking the Intel OAE.
The neuromorphic system also responded up to
200 milliseconds faster than the GPU. This
dramatic improvement in energy efficiency for
our task comes from the fact that computation
in Loihi is extremely sparse. While the GPU
performs billions of computations per second,
every second, the neuromorphic chip only
processes changes in the audio signal and
neuron cores inside low Loihi communicate
efficiently with spikes.
This project demonstrates that neuromorphic
systems can prove more efficient and more
responsive than conventional solutions for AI in
smart vehicles. This research is helping our
partners in the automotive industry understand
how Intel's neuromorphic systems might impact
their next-generation products. And it helps us
develop a roadmap for future neuromorphic
applications.
Copyright © 2021 Accenture
All rights reserved.
Accenture, its logo, and High
Performance Delivered are
trademarks of Accenture