Deep learning – AI
This year is also likely to see the application of deep learning to
widespread use in autonomous transport. Here the stakes are arguably
higher as lives are on the line and no computer code can be written that
can take account of all possibilities. That means the underlying artifcial
intelligence has to learn to expect the unexpected.
The catalyst for the proliferation of this technology in 2018 may
be the electric car pioneer Tesla. While there are about 100,000 Teslas on
the road, the company has announced that all future models will be
equipped with an onboard “super computer” that can provide full
self-driving capability.
These computers will use an Nvidia platform,
allowing them to process large amounts of data for deep learning.
This deep learning is likely to make it practical to complete the
final step in the three-step system for widespread autonomous vehicle
implementation. The first step is perception, where the sensors identify
objects and classify them. The second is localisation, where the computer
integrates what the sensors are saying with landmarks, maps, and
position information. The science behind these steps has been relatively
well developed over the last few years.
EU Forecast
euf:b18:98/nws-01