Download Spotlight Publication CIS Publication Spotlight

Publication
Spotlight
Derong Liu, Chin-Teng Lin,
Kay Chen Tan, Graham Kendall,
and Angelo Cangelosi
CIS Publication Spotlight
IEEE Transactions on Neural
Networks and Learning Systems
tion performance, when compared with
other similar MTL approaches.”
Pareto-Path Multitask Multiple Kernel
Learning, by C. Li, M. Georgiopoulos,
and G.C. Anagnostopoulos, IEEE
Transactions on Neural Networks and
Learning Systems, Vol. 26, No. 1, January 2015, pp. 51–61.
A Parametric Classification Rule Based
on the Exponentially Embedded Family,
by B. Tang, H. He, Q. Ding, and S.
Kay, IEEE Transactions on Neural Networks and Learning Systems, Vol. 26,
No. 2, February 2015, pp. 367–377.
Digital Object Identifier: 10.1109/
TNNLS.2014.2309939
“A traditional and intuitively appealing multitask multiple kernel learning
(MT-MKL) method is developed to optimize the sum and the average of objective functions with partially shared kernel
function, which allows information sharing among the tasks. The obtained solution corresponds to a single point on the
Pareto front (PF) of a multiobjective
optimization problem, which considers
the concurrent optimization of all task
objectives involved in the multitask
learning (MTL) problem. A novel support vector machine MT-MKL framework is proposed that considers an
implicitly defined set of conic combinations of task objectives. It is shown that
solving this framework produces solutions along a path on the aforementioned
PF and that it subsumes the optimization
of the average of objective functions as a
special case. Using the algorithms derived,
it is demonstrated through a series of
experimental results that the framework
is capable of achieving a better classifica-
Digital Object Identifier: 10.1109/
TNNLS.2014.2383692
“An approach for estimating model
order and constructing probability density function called exponentially
embedded family (EEF) is extended to
multivariate pattern recognition. Specifically, a parametric classifier rule based on
the EEF is developed, in
which a distribution is constructed for each class based
on a reference distribution.
The proposed method can
address different types of
classification problems in
either a data-driven manner
or a model-driven manner.
The effectiveness is demonstrated with examples of
synthetic data classification
and real-life data classifica- © eyewire
tion in a data-driven manner
and an example of power quality disturbance classification in a model-driven
manner. To evaluate the classification
performance of the approach, the
Monte-Carlo method is used in the
experiments. The promising experimental results indicate many potential applications of the proposed method.”
Digital Object Identifier 10.1109/MCI.2015.2405275
Date of publication: 10 April 2015
14
IEEE Computational intelligence magazine | may 2015
IEEE Transactions on Fuzzy Systems
A Collaborative Fuzzy Clustering Algorithm in Distributed Network Environments, by J. Zhou, C.L.P. Chen, L.
Chen, and H.X. Li, IEEE Transactions
on Fuzzy Systems, Vol. 22, No. 6,
December 2014, pp. 1443–1456.
Digital Object Identifier: 10.1109/
TFUZZ.2013.2294205
“Traditional centralized approaches
using in data clustering possessing privacy and security demands or technical
constraints in a large dynamic distributed
peer-to-peer network are difficult to sort
out. In a P2P network, each peer has
equal functionality. A peer is a facilitator
and a worker at the same
time. Each peer can communicate with others for the
network structure. The local
data in this peer and the
necessary infor mation
exchanged from others have
to be taken into consideration when the P2P distributed clustering algorithm
intends to complete the
locally optimized clusters at
each peer. Therefore, the
authors propose a novel collaborative fuzzy clustering algorithm for
prevailing over a distributed P2P network. This manner searches the optimized clusters at each peer by collaborating with topologically neighboring
peers only step by step, till it reaches the
global consensus of all peers. Furthermore, the proposed algorithm can also
conduct high-dimensional sparse data
clustering and “nonspherical”-shaped
data clustering, which are not considered
by other distributed methods but widely
used in some practical applications.
Finally, this study provides several synthetic and real-world datasets to demonstrate their efficiency and superiority
compared to some existing methods.
Construction of Neurofuzzy Models For
Imbalanced Data Classification, by M.
Gao, X. Hong, and C.J. Harris, IEEE
Transactions on Fuzzy Systems, Vol. 22,
No. 6, Dec. 2014, pp. 1472–1488.
Digital Object Identifier: 10.1109/
TFUZZ.2013.2296091
“The authors propose a new class of
neurofuzzy construction algorithms
with the aim of maximizing generalization capability specifically for imbalanced data classification problems based
on leave-one-out (LOO) cross-validation. The proposed algorithms are in
two stages: First, an initial rule base is
constructed based on estimating the
Gaussian mixture model with analysis
of variance decomposition from input
data; the second stage carries out the
joint weighted least squares parameter
estimation and rule selection using an
orthogonal forward subspace selection
(OFSS) procedure. The authors show
how different LOO based rule selection criteria can be incorporated with
OFSS and advocate either maximizing
the LOO area under curve of the
receiver operating characteristics or
maximizing the LOO F-measure if the
datasets exhibit imbalanced class distribution. Extensive comparative simulations illustrate the effectiveness of the
proposed algorithms.”
IEEE Transactions on
Evolutionary Computation
Reusing Genetic Programming for
Ensemble Selection in Classification of
Unbalanced Data, by U. Bhowan, M.
Johnston, M. Zhang, and X. Yao,
IEEE Transactions on Evolutionary
Computation, Vol. 18, No. 6, December 2014, pp. 893–908.
Digital Object Identifier: 10.1109/
TEVC.2013.2293393
“Classification algorithms can suffer
from performance degradation when the
class distribution is unbalanced. This
paper develops a two-step approach to
evolving ensembles using genetic programming (GP) for unbalanced data. It
combines multiple Pareto-approximated
front members into a single composite
genetic program solution to represent the
(optimized) ensemble. It is shown that
the proposed GP approach evolves
smaller more diverse ensembles compared to an established ensemble selection algorithm, while still performing as
well as, or better than the established
approach. The evolved GP ensembles also
perform well compared to other bagging
and boosting approaches, particularly on
tasks with high levels of class imbalance.”
Learning Value Functions in Interactive
Evolutionary Multiobjective Optimization, by J. Branke, S. Greco, R. Slowinski, and R. Zielniewicz, IEEE
Transactions on Evolutionary Computation, Vol. 19, No. 1, February 2015,
pp. 88–102.
Digital Object Identifier: 10.1109/
TEVC.2014.2303783
“This paper proposes an interactive
multiobjective evolutionary algorithm
(MOEA) that attempts to learn a value
function capturing the users’ true preferences. At regular intervals, the user is asked
to rank a single pair of solutions.This information is used to update the algorithm’s
internal value function model, and the
model is used in subsequent generations to
rank solutions incomparable according to
dominance. This speeds up evolution toward the region of the Pareto front that is
most desirable to the user. It takes into account the most general additive value function as a preference model and empirically
compares different ways to identify the
value function that seems to be the most
representative with respect to the given
preference information, different types of
user preferences, and different ways to use
the learned value function in the MOEA.
Results on a number of different scenarios
suggest that the proposed algorithm works
well over a range of benchmark problems
and types of user preferences.”
IEEE Transactions on
Computational Intelligence
and AI in Games
A Neuroevolution Approach to General
Atari Game Playing, by M. Hausknecht, J. Lehman, R. Miikkilainen
and P. Stone, IEEE Transactions on
Computational Intelligence and AI in
Games, Vol. 6, No. 4, December 2014,
pp. 355–366.
Digital Object Identifier: 10.1109/
TCIAIG.2013.2294713
“This paper addresses the challenge of
learning to play many different video
games with little domain-specific knowledge. Specifically, it introduces a neuroevolution approach to general Atari 2600 game
playing. Four neuroevolution algorithms
were paired with three different state representations and evaluated on a set of 61 Atari
games.The neuroevolution agents represent
different points along the spectrum of algorithmic sophistication - including weight
evolution on topologically fixed neural networks (conventional neuroevolution), covariance matrix adaptation evolution strategy
(CMA-ES), neuroevolution of augmenting
topologies (NEAT), and indirect network
encoding (HyperNEAT). State representations include an object representation of
the game screen, the raw pixels of the game
screen, and seeded noise (a comparative
baseline). Results indicate that directencoding methods work best on compact
state representations while indirect-encoding methods (i.e., HyperNEAT) allow scaling to higher dimensional representations
(i.e., the raw game screen). Previous
approaches based on temporal-difference
(TD) learning had trouble dealing with the
large state spaces and sparse reward gradients often found in Atari games. Neuroevolution ameliorates these problems and
evolved policies achieve state-of-the-art
results, even surpassing human high scores
on three games. These results suggest that
neuroevolution is a promising approach to
general video game playing (GVGP).”
(continued on page 52)
may 2015 | IEEE Computational intelligence magazine
15
[22] R. Romo, A. Hernández, A. Zainos, C. Brody, and
L. Lemus, “Sensing without touching: Psychophysical
performance based on cortical microstimulation,” Neuron, vol. 26, no. 1, pp. 273–278, 2000.
[23] W. Schultz, “Getting formal with dopamine and reward,” Neuron, vol. 36, no. 2, pp. 241–263, 2002.
[24] Y. Wang, X. Su, R. Huai, and M. Wang, “A telemetry navigation system for animal-robots,” Robot, vol. 28,
no. 2, pp. 183–186, 2006.
[25] L. Bourdev and J. Brandt, “Robust object detection
via soft cascade,” in Proc. IEEE Computer Society Conf. Computer Vision Pattern Recognition, 2005, vol. 2, pp. 236–243.
[26] C. Harris and M. Stephens, “A combined corner and
edge detector,” in Proc. Alvey Vision Conf., 1988, vol. 15, p. 50.
[27] B. D. Lucas and T. Kanade, “An iterative image registration technique with an application to stereo vision,”
in Proc. 7th Int. Joint Conf. Artificial Intelligence, 1981, vol.
81, pp. 674–679.
[28] G. Paxinos and C. Watson, The Rat Brain in Stereotaxic Coordinates: Hard Cover Edition. New York: Academic,
2006.
[29] F. Lucivero and G. Tamburrini, “Ethical monitoring
of brain-machine interfaces,” AI Soc., vol. 22, no. 3, pp.
449–460, 2008.
Publication Spotlight (continued from page 15)
IEEE Transactions on Autonomous
Mental Development
© GRAPHIC STOCK
A Hierarchical System for a Distributed
Representation of the Peripersonal Space of
a Humanoid Robot, by A. Antonelli, A.
Gibaldi, F. Beuth, A.J. Duran, A. Canessa, M. Chessa, F. Solari, A.P. del Pobil,
F. Hamker, E. Chinellato, and S.P.
Sabatini, IEEE Transactions on Autonomous Mental Development, Vol. 6, No. 4,
December 2014, pp. 259–273.
52
Digital Object Identifier: 10.1109/
TAMD.2014.2332875
“This work demonstrates in a
humanoid torso the feasibility of an
integrated working representation of
the robot’s per ipersonal space. To
achieve this goal, the paper presents a
cognitive architecture that connects
several models inspired by neural circuits of the visual, frontal and posterior
parietal cortices of the brain. The outcome of the integration process is a
We want
to hear
from you!
IEEE Computational intelligence magazine | MAY 2015
system that allows the robot to create
its internal model and its representation
of the surrounding space by interacting
with the environment directly, through
a mutual adaptation of perception and
action. The robot is eventually capable
of executing a set of tasks, such as recognizing, gazing and reaching target
objects, which can work separately or
cooperate for supporting more structured and effective behaviors.”
Do you like what you're reading?
Your feedback is important.
Let us know—send the editor-in-chief an e-mail!