Signal Processing - November 2017 - 127

from user interaction with the device. The learning models for
coadaptation will be based on unsupervised domain adaptation,
which would involve gleaning patterns from unlabeled user interaction data. There has been no work so far in the domain adaptation literature for person-centered device adaptation, and these
person-centered adaptive models would make technology more
accessible, especially to individuals with disabilities.

Conclusions
The current generation of artificial intelligence systems can
outperform humans in a narrow set of tasks like playing chess
or GO. Even though deep neural networks have contributed to
the unprecedented progress of artificial intelligence research in
the last few years, artificial general intelligence has, so far, been
elusive. To advance artificial intelligence, computational systems will need the ability to transfer learning and progressively
augment knowledge. Transfer learning paradigms like domain
adaptation will be key to heralding the next generation of artificial intelligence systems. The current generation of domain
adaptation models is dominated by deep-learning systems. Prior
to the advent of deep learning, domain adaptation approaches
had to develop adaptive computational models based on fixed
representations of data. Deep-learning systems have found great
success in domain adaptation because of their ability to extract
domain aligned features specific to the adaptation task. This
has led to a surge in domain adaptation research in recent years,
and this article has provided a survey of literature in the area
of domain adaptation based on deep learning. In this survey,
we have outlined the concept of knowledge transfer across
computational models and categorized the different paradigms
of transfer and compared them with domain adaptation. This
article is meant to provide a clear understanding of the scope of
research in domain adaptation and also highlight the promising
directions for future research.

Acknowledgments
We are grateful to the reviewers who provided highly valuable
comments that have helped to improve the quality of the article.
This material is based on work supported by the National Science
Foundation under grant 1116360.

Authors
Hemanth Venkateswara (hemanthv@asu.edu) graduated
summa cum laude with master's degrees in physics and computer
science from the Sri Sathya Sai University, India, in 2005 and
2007, respectively. He is a postdoctoral researcher at the School
of Computing Informatics and Decision Systems Engineering at
Arizona State University. His areas of research includes machine
learning and computer vision, with applications in domain adaptation using deep learning. Prior to earning his Ph.D. degree, he
worked as a senior software engineer at Alcatel-Lucent Technologies, India. He is a Member of the IEEE and the Association
for Computing Machinery.
Shayok Chakraborty (shayok.chakraborty@asu.edu)
received his Ph.D. degree in computer science from Arizona
State University (ASU) in 2013. He was an assistant research

professor of computer science at ASU and is now an assistant
professor of computer science at Florida State University,
Tallahassee. He has worked as a postdoctoral researcher at Intel
Labs and also in the Electrical and Computer Engineering
Department at Carnegie Mellon University, Pittsburgh,
Pennsylvania. His research interests include computer vision and
machine learning. He has published his research in premier conferences and journals such as the IEEE Conference on Computer
Vision and Pattern Recognition, the Association for Comput--
ing Machinery (ACM) SIGKDD, ACM Multimedia, IEEE
Transactions on Pattern Analysis and Machine Intelligence,
IEEE Transactions on Neural Networks and Learning Systems,
and Pattern Recognition. He has also extensively served in the
program committee and as a reviewer of these conferences and
journals. He is a Member of the IEEE and ACM.
Sethuraman Panchanathan (panch@asu.edu) receieved his
bachelor's degree in electronics and communication engineering
from the Indian Institute of Science in 1984, his master's degree in
electrical engineering from Indian Institute of Technology,
Madras, in 1986, and his Ph.D. degree in electrical and computer
engineering from the University of Ottawa, Canada, in 1989. He
leads the Knowledge Enterprise Development, Arizona State
University (ASU), Tempe, which advances research, innovation,
strategic partnerships, entrepreneurship, global, and economic
development at ASU. In 2014, he was appointed by U.S.
President Barack Obama to the U.S. National Science Board. He
has also been appointed by U.S. Secretary of Commerce Penny
Pritzker to the National Advisory Council on Innovation and
Entrepreneurship. He is a fellow of the National Academy of
Inventors, the Canadian Academy of Engineering, and the Society
of Optical Engineering. He currently serves as the chair-elect in
the Council on Research within the Association of Public and
Land-Grant Universities. He has authored more than 425 papers in
refereed journals and conferences. His research interests include
human-centered multimedia computing, haptic user interfaces,
person-centered tools, and ubiquitous computing technologies for
enhancing the quality of life for individuals with disabilities,
machine learning for multimedia applications, medical image processing, and media processor designs.

References

[1] S. J. Pan and Q. Yang, "A survey on transfer learning," IEEE Trans.
Knowledge Data Eng., vol. 22, no. 10, pp. 1345-1359, 2010.
[2] L. Chen, W. Li, and D. Xu, "Recognizing RGB images by learning from
RGB-D data," in Proc. IEEE Conf. Computer Vision and Pattern Recognition,
2014, pp. 1418-1425.
[3] J. Hoffman, S. Gupta, J. Leong, S. Guadarrama, and T. Darrell, "Cross-modal
adaptation for RGB-D detection," in Proc. IEEE Robotics and Automation Int.
Conf., 2016, pp. 5032-5039.
[4] P. Koniusz, Y. Tas, and F. Porikli, "Domain adaptation by mixture of alignments of second-or higher-order scatter tensors," in Proc. IEEE Conf. Computer
Vision and Pattern Recognition, 2017.
[5] B. F. Klare, S. S. Bucak, A. K. Jain, and T. Akgul, "Towards automated caricature recognition," in Proc. IEEE Biometrics 5th IAPR Int. Conf., 2012, pp. 139-146.
[6] E. J. Crowley and A. Zisserman, "In search of art," in Proc. Workshop on
Computer Vision Art Analysis, European Conf. Computer Vision (ECCV), 2014,
pp. 54-70.
[7] S. Saxena and J. Verbeek, "Heterogeneous face recognition with CNNs," in
Proc. European Conf. Computer Vision, 2016.
[8] L. Castrejon, Y. Aytar, C. Vondrick, H. Pirsiavash, and A. Torralba,
"Learning aligned cross-modal representations from weakly aligned data," in

IEEE SIGNAL PROCESSING MAGAZINE

November 2017

127

Table of Contents for the Digital Edition of Signal Processing - November 2017