18
§ Smart homes – residential environments
that are equipped with connected products
for controlling, automating and optimizing
functions such as temperature, lighting,
security, safety or entertainment.
§ Smart manufacturing – a technology-driven
approach using connected machinery to
monitor production processes and data
analytics to improve manufacturing operations.
§ Smart transportation – a scenario in which
mobility-related entities (e.g. personal vehicles,
public transportation, delivery vehicles,
emergency services, parking) are integrated
and automated in order to improve trafc
performance and reduce potential negative
impacts such as congestions, accidents or
pollution.
§ Smart energy – a sustainable and cost-
effective energy system in which renewable
production, consumption, and infrastructures
are integrated and coordinated through energy
services, active users and enabling information
and communication technologies (ICTs).
1.6 Outline of the White Paper
Section 2 describes the need for AI through
identication of several key megatrends posing
major challenges for societies, businesses and
individuals. AI will enable and enhance a wide
range of applications addressing some of these
challenges, such as environmental concerns,
changing demographic trends or economic
disparity, to mention only a few.
Although AI is hardly a new discipline, it was not
until 2010 and later that dramatic technological
enhancements, in particular in the area of machine
learning, paved the way for today’s explosion
of AI. This breakthrough was enabled thanks
to a number of factors explained in Section 3.
Signicant improvements in computational power,
more sophisticated machine learning algorithms
and the availability of large amounts of data to
train AI systems have been the primary enablers of
today’s spectacular AI developments.
A number of additional drivers that have also
contributed to the ourishing eld of AI research
are further outlined in Section 3. These include
information technology (IT) developments such
as cloud and edge computing, the Internet of
Things (IoT), and big data, as well as the increasing
readiness of consumers and society to embrace
new technologies and share data.
Section 4 provides a high-level understanding
of the most common AI systems and machine
learning techniques. Without entering into deep
technical detail, it also reviews the most popular
AI algorithms in use today that constitute the
foundation for tomorrow’s AI developments. Based
on the current state of the art of AI, this technical
overview is complemented by many references
for readers wishing to consolidate their scientic
understanding of how AI actually works from the
inside.
While this White Paper cannot cover all of the
possible AI application scenarios, a representation
describing how today’s main AI systems map to
some of the most popular application domains
is provided in Section 5. This exercise furnishes
a better characterization of the AI needs and
requirements of several industry sectors. The rest
of the section is then devoted to a more detailed
description of the four application domains (smart
homes, smart manufacturing, smart transportation,
and smart energy), for which several AI-related use
cases and scenarios are reviewed, together with
some of the most pressing challenges of current
and emerging AI implementations.
Following this review, Section 6 consolidates the
main AI challenges that can be identied in today’s
implementations or foreseen in emerging AI
developments. Challenges are grouped into several
categories: social and economic challenges;
data-related challenges, including the selection
Introduction
评论0