APNNS/IEEE-CIS Education Forum Series

Pau-Choo (Julia) Chung of National Cheng Kung University

Gait and balance patterns are two major measurements for the evaluation of the elderly’s physical conditions and are therefore often used in the early detection of certain diseases. For example, patients with Alzheimer’s disease (AD) were reported of revealing gait disorders and balance problems. In this talk we will present an inertial-sensor based wearable device, which are designed for objective quantitative measurement of gait patterns and balance capabilities. Accompanied with the device are the algorithms and CI models, which integrally provide the quantitative evaluations. We will also discuss several essential indicators from gait and balance patterns for AD
diagnosis. The gait analyzing algorithm, which is composed of stride detection followed by gait cycle decomposition, and the balance analysis algorithm, which is measured by the sway speed in anterior-posterior (AP) and medial-lateral (ML) directions of the
projection path of body’s center of mass (COM), will also be introduced.

Biography:

Pau-Choo (Julia) Chung (S’89-M’91-SM’02-F’08) received the Ph.D. degree in electrical engineering from Texas Tech University, USA, in 1991. She then joined the Department of Electrical Engineering, National Cheng Kung University (NCKU), Taiwan, in 1991 and has become a full professor in 1996. She served as the Director of Institute of Computer and Communication Engineering (2008-2011), the Vice Dean of College of Electrical Engineering and Computer Science (2011), the Director of the Center for Research of E-life Digital Technology (2005-2008), and the Director of Electrical Laboratory (2005-2008), NCKU. She was elected Distinguished Professor of NCKU in 2005. She also
served as Chair of Department of Electrical Engineering, NCKU.

Dr. Chung’s research interests include image/video analysis and pattern recognition, biosignal analysis, computer vision and computational intelligence. She applies most of her research results to healthcare and medical applications. Dr. Chung served as the program committee member in many international conferences. She was a Member on IEEE International Steering Committee, IEEE Asian Pacific Conference on Circuits and Systems (2006-2008), the Special Session Co-Chair of ISCAS 2009 and 2010, the Special Session Co-Chair of ICECS 2010, and the TPC of APCCAS 2010. Dr. Chung was the Chair of IEEE Computational Intelligence Society (CIS) (2004-2005), Tainan Chapter. She was the Chair of the IEEE Life Science Systems and Applications Technical Committee (2008-2009) and a member of the BioCAS Technical Committee
and the Multimedia Systems & Applications Technical Committee of the CAS Society. She also serves as the Associate Editor of IEEE Transactions on Neural Networks and served as the Editor of Journal of Information Science and Engineering, the Guest Editor of Journal of High Speed Network, the Guest Editor of IEEE Transactions on Circuits and Systems-I, and the Secretary General of Biomedical Engineering Society of the Republic of China. She is one of the co-founders of Medical Image Standard Association (MISA) in Taiwan and served on the Board of Directors of MISA.

Pau-Choo Chung was a member in BoG of CAS Society (2007-2009, 2010-2012). She served as an IEEE CAS Society Distinguished Lecturer (2005-2007). She is an ADCOM member of IEEE CIS and current Vice President for Education. She is a Member of Phi
Tau Phi honor society and is an IEEE Fellow since 2008.

Saman Halgamuge of The University of Melbourne

In the past decade, Deep Neural Networks (DNNs) based on supervised learning have revolutionized various fields including Computer Vision, Natural Language Processing, Bioinformatics and Robotics. Behind this revolution is the increasing demand for computational power, with reportedly the amount of computing used in the training of largest manually designed DNN models doubling every 3.5 months since 2012, much faster than the two-year doubling period of Moore’s law in electronic hardware advancement. While continually meeting such a demand is unsustainable and unlikely, a need arises for significant innovations in discovering or designing DNN architectures and training procedures that are significantly more efficient and demand much less computing power, i.e., low cost. Such innovations could also benefit the wider use of DNNs by researchers without expertise in DNN design in many new areas including in energy, environmental and social sciences and arts and humanities.

To reduce the development cost of DNNs, a recent idea proposed is to automate the DNN design, which leads to an emerging field called automatic machine learning (Auto-ML). However, this idea was previously applied by the author on shallow Neural Networks using Self generation/growing [1-3]. Existing Auto-ML methods have attempted to optimize every step of the data analysis pipeline including data
preparation, feature engineering, model generation, training, and evaluation. Among them, Neural Architecture Search (NAS) methods explicitly find DNN architectures for a given supervised learning task. This is achieved by encoding the candidate architecture as a solution in some search space and treating the architecture design as an optimization problem. Growing Neural Network Architectures instead of “searching for the best” has been our strategy to this problem.

Our research shows that Self Growing Neural Networks can be used both in Unsupervised Learning and Supervised Learning. In the former, they can be applied to continuous data streams. In the latter, our techniques also lead to interpretable architectures. We also refer to recent work jointly published with Dr Damith Senanayake, Dr Wei Wang and others [3-6].

Acknowledgment: This research is funded by ARC Discovery project grant DP210101135.

References
1. SK Halgamuge, M Glesner, “Neural Networks in Designing Fuzzy Systems for Real World Applications”, Fuzzy Sets and Systems 65 (1), 1-12, 1994
2. SK Halgamuge, W Pochmueller, M. Glesner, “An Alternative Approach for Generation of Membership Functions and Fuzzy Rules Based on Radial and Cubic Basis Function Networks”, International Journal of Approximate Reasoning 12 (3), 4, 1995
3. SK Halgamuge, “Self Evolving Neural Networks for rule based data processing”, IEEE Transactions on Signal Processing, 45 (11), 2766-2773, 1997
4. D Senanayake, W Wang, SH Naik, S Halgamuge, “Self Organizing Nebulous Growths for Robust and Incremental Data Visualization”,
IEEE Transactions on Neural Networks and Learning Systems, 2020
5. W Wang, Y Sun, S Halgamuge, “Improving MMD-GAN Training with Repulsive Loss Function”, International Conference on Learning Representations (ICLR 2019), 2019
6. PN Hameed, K Verspoor, S Kusljic, S Halgamuge, “Positive-unlabeled learning for inferring drug interactions based on heterogeneous attributes”, BMC bioinformatics 18 (1), 1-15, 2017.

Biography:

Prof Saman Halgamuge is a Fellow of IEEE, a Professor in the School of Electrical, Mechanical and Infrastructure Engineering and a Distinguished Speaker/Lecturer on Computational Intelligence (2019-2021). He served as Director/Head, Research School of Engineering of the Australian National University (ANU) (2016-18), a member of Australian Research Council (ARC) College of Experts for Engineering, Information and Computing Sciences (2016-18), the founding Director of the PhD training centre Melbourne India Postgraduate Program (MIPP) of University of Melbourne and Associate Dean (2013-15) and Assistant Dean (2008-13) in International Engagement in the Melbourne School of Engineering. Saman is also an honorary Professor of ANU.