IEEE Computational Intelligence Magazine - February 2022 - 15

classification, for which XAI methods are
still not widespread. The problem tackled
by the authors is to make time-series
classes easier to discern by users. To this
end, the authors propose to learn pictorial
class prototypes together with the class of
time-series, so as to explain classes by
using visual clues. They accomplish this
task with a combination of Convolutional
Neural Networks and a Multilayer Perceptron
that is used for encoding timeseries,
generating visual prototypes and
making final classification. The experiments
were conducted by involving endusers,
divided into three groups (STEM,
clinicians and others), which were
employed to evaluate perceived interpretability,
accuracy, response time, and trust.
The third paper, entitled " recoXplainer:
A Library for Development and
Offline Evaluation of Explainable Recommender
Systems " by L. Coba et al.,
goes into a different direction and tries to
fill a gap in explainable recommender systems
(XRS) by proposing a library with
the aim of boosting research reproducibility
and software interoperability in this
emerging field. The library enables the
development of XRSs through a pipeline
(pre-process, train, recommend, explain
and evaluate) and includes several solutions
for building explanations, including
model-based and post-hoc approaches.
The library (which is available as open
source, under the MIT license) has been
applied to develop XRSs for publicly
available benchmark datasets. It shows the
different performance of the implemented
methods, in terms of explainability as
measured by the available metrics.
The fourth paper is entitled " Collective
eXplainable AI: Explaining Cooperative
Strategies and Agent Contribution in
Multiagent Reinforcement Learning with
Shapley Values " by A. Heuillet et al. It
tackles the issue of explainability in Reinforcement
Learning (RL), a field that is
very active in Machine Learning research,
but received little attention in XAI so far.
Nevertheless, the introduction of XAI in
RL could provide important insights in
many scientific fields, including Economy
and Social Sciences, as well as in industrial
applications. The authors focus on
explainability in Cooperative RL, where
agents must cooperate in order to achieve
a common goal. Explainability is introduced
by approximating Shapley values
that evaluate the contribution of each
agent in achieving the common goal.
The Special Issue is closed by the fifth
paper, entitled " Bridging the Gap
Between AI and Explainability in the
GDPR: Towards Trustworthiness-byDesign
in Automated Decision-Making "
by R. Hamon et al. This work is of interdisciplinary
nature as it studies the current
state of the art of XAI solutions and the
legal requirements for explainability, with
special emphasis on the GDPR which sets
out the fundamental rights of individuals
when their personal data are subject to
processing (e.g., in profiling and automated
decision-making). The authors put forward
a number of challenges that limit
current AI approaches in fulfilling the legal
requirements set by GDPR (namely, complexity,
accuracy vs. explainability, and correlation
vs. causation) and describe a number
of proposals that shed light on future re -
search on XAI towards the development
of an AI that is not only technically effective
but also socially and legally acceptable.
The guest editors of this special issue
would like to thank Prof. Chuan-Kang
Ting, the Editor-in-Chief of IEEE Computational
Intelligence Magazine, for his
great support in initiating and developing
this special issue together. Many thanks to
all members of the editorial team for
their kind support during the editing
process of this special issue. Last but not
least, we would also like to thank the
authors for submitting their valuable
research outcomes as well as the reviewers
who critically evaluated the papers.
We sincerely hope and expect that readers
will find this special issue useful.
We want
to hear
from you!
Do you like what you're reading?
Your feedback is important.
Let us know-send the editor-in-chief an e-mail!
FEBRUARY 2022 | IEEE COMPUTATIONAL INTELLIGENCE MAGAZINE 15
IMAGE LICENSED BY GRAPHIC STOCK
IEEE Computational Intelligence Magazine - February 2022

Table of Contents for the Digital Edition of IEEE Computational Intelligence Magazine - February 2022
