IEEE Computational Intelligence Magazine - November 2021 - 16

serious game community. Nevertheless,
a framework for developing a generic
tool has been proposed to lift its barriers.
In this article, a generic stealth
assessment (SA) software tool was
developed based on this framework to
examine the robustness of the stealth
assessment methodology under various
conditions. In specific, the conditions
relate to processing data sets of different
distribution types and sizes (960 data
sets containing a total of 72.336.000
data points are used for this reason),
utilizing two different ML algorithms
(Gaussian Naïve Bayes Network and
C4.5), and using statistical models relating
to two different competency constructs.
Results show that stealth
assessment is a robust methodology,
whilst the generic SA tool is a highly
accurate tool capable of handling efficiently
a wide range of conditions. "
IEEE Transactions on Cognitive
and Developmental Systems
Intrinsically Motivated Hierarchical Policy
Learning in Multiobjective Markov
Decision Processes, by S. Abdelfattah, K.
Merrick, and J. Hu, IEEE Transactions
on Cognitive and Developmental Systems,
Vol. 13, No. 2, June 2021, pp.
262-273.
Digital Object Identifier: 10.1109/
TCDS.2019.2948025
" The multiobjective Markov decision
processes (MOMDPs) are sequential
decision-making problems that
involve multiple conflicting reward
functions that cannot be optimized
simultaneously without a compromise.
This type of problem cannot be solved
by a single optimal policy as in the conventional
case. Alternatively, multiobjective
reinforcement learning (RL)
methods evolve a coverage set of optimal
policies that can satisfy all possible
preferences in solving the problem.
However, many of these methods cannot
generalize their coverage sets to
work in the nonstationary environments.
In these environments, the
parameters of the state transition and
reward distribution vary over time. This
limitation results in significant performance
degradation for the evolved
policy sets. In order to overcome this
limitation, there is a need to learn a
generic skillset that can bootstrap the
evolution of the policy coverage set for
each shift in the environment dynamics,
therefore, it can facilitate a continuous
learning process. In this article, intrinsically
motivated RL (IMRL) has been
successfully deployed to evolve generic
skillsets for learning hierarchical policy
to solve the MOMDPs. We propose a
novel dual-phase IMRL method to
address this limitation. In the first phase,
a generic set of skills is learned, while in
the second phase, this set is used to
bootstrap policy coverage sets for each
shift in the environment dynamics. We
show experimentally that the proposed
method significantly outperforms the
state-of-the-art multiobjective reinforcement
methods on a dynamic
robotics environment. "
IEEE Transactions on Emerging
Topics in Computational
Intelligence
Training Deep Photonic Convolutional
Neural Networks with Sinusoidal Activations,
by N. Passalis, G. M. Alexandris,
A. Tsakyridis, N. Pleros, and A. Tefas,
IEEE Transactions on Emerging Topics in
Computational Intelligence, Vol. 5, No.
3, June 2021, pp. 384-393.
Digital Object Identifier: 10.1109/
TETCI.2019.2923001
" Deep learning (DL) has achieved
state-of-the-art performance in many
challenging problems. However, DL
requires powerful hardware for both
training and deployment, increasing the
cost and energy requirements and rendering
large-scale applications especially
difficult. Recognizing these difficulties,
several neuromorphic hardware solutions
have been proposed, including
photonic hardware that can process
information close to the speed of light
and can benefit from the enormous
bandwidth available on photonic systems.
However, the effect of using these
photonic-based neuromorphic architec16
IEEE COMPUTATIONAL INTELLIGENCE MAGAZINE | NOVEMBER 2021
tures, which impose additional constraints
that are not usually considered
when training DL models, is not yet
fully understood and studied. The main
contribution of this paper is an extensive
study on the feasibility of training deep
neural networks that can be deployed
on photonic hardware that employ sinusoidal
activation elements, along with
the development of methods that allow
for successfully training these networks,
while taking into account the physical
limitations of the employed hardware.
Different DL architectures and four
datasets of varying complexity were
used for extensively evaluating the proposed
method. "
IEEE Transactions on
Artificial Intelligence
Privacy and Artificial Intelligence, by J.
Curzon, T. Kosa, R. Akalu, and K. ElKhatib,
IEEE Transactions on Artificial
Intelligence, Vol. 2, No. 2, Apr 2021.
Digital Object Identifier: 10.1109/
TAI.2021.3088084
" While an appreciation of the privacy
risks associated with artificial
intelligence is important, a thorough
understanding of the assortment of different
technologies that comprise artificial
intelligence better prepares those
implementing such systems in assessing
privacy impacts. This can be achieved
through the independent consideration
of each constituent of an artificially
intelligent system and its interactions.
Under individual consideration, privacyenhancing
tools can be applied in a
targeted manner to reduce the risk associated
with specific components of an
artificially intelligent system. A generalized
North American approach to assess
privacy risks in such systems is proposed
that will retain applicability as
the field of research evolves and can be
adapted to account for various sociopolitical
influences. With such an approach,
privacy risks in artificial intelligent systems
can be well understood, measured,
and reduced. "
IEEE Computational Intelligence Magazine - November 2021

Table of Contents for the Digital Edition of IEEE Computational Intelligence Magazine - November 2021
