Â鶹´«Ã½

Biosketch

Guan Yue Hong is an associate professor of Computer Science at Â鶹´«Ã½. She received her Ph.D. in software engineering from National University of Singapore. Her research interests include reliable and trustworthy AI, novel machine learning paradigms, and smart cyber-physical-human systems.  She received several grants and awards in support of her scholarly activities. She has published over 70 papers in leading international journals and conferences, e.g. IEEE Trans. Affective Computing, IEEE Trans. Consumer Electronics, and IEEE Trans. IT in Biomedicine. Her published papers have attracted over a thousand scientific citations. 

The first real-world project she worked on was developing an intelligent decision support system for flight control systems while she was pursuing a master’s degree in Artificial Intelligence. She went on to obtain a Ph.D. where she focused her research on how to analyze, model and improve the reliability of complex software systems. After graduation, Dr. Hong began her professional career as a software engineer with Motorola’s Global Software Division, where she worked on the development and testing of cutting-edge software products. She also taught at Motorola University as an instructor in software coding and testing best practices, telecommunication network programming, and systems integration. Dr. Hong received several performance awards and was promoted to senior software engineer. After leaving Motorola, she became a tenured faculty member for ten years at Massey University and Unitec Institute of Technology in New Zealand. During that time, she received three academic promotions in recognition of her outstanding contributions.  Dr. Hong has taught courses at both undergraduate and graduate levels with class sizes ranging from 16 to 350 Â鶹´«Ã½s. She later received a second master’s degree in management and finance from Harvard University.

The overarching theme of her research is based on the vision that the transformative power of AI is beneficial to humans and delivers the most valuable resources for an increasingly safe, reliable, and productive society. Her research aims to address the challenges and Â鶹´«Ã½ we are facing today, specifically in the following:

1 Building Reliable and Trustworthy AI

The more integration of AI systems into our daily lives, the more important quality and reliability evaluations will be due to their increasingly bigger societal implications. Majority of her early research work focused on reliability modeling, estimation and prediction of software and Cyber- Physical Systems. For her PhD, she studied several intriguing issues facing industry, e.g. the optimal software release time based on unbonded NHPP models, reliability growth estimation during the early stage of testing that has been used for commercial product release planning and cost budgeting, as well as using Statistical Process Control techniques for software testing and inspection monitoring. These findings enabled her to see wide applications of reliability theory to every fiber of our daily lives from washing machines, TVs to cars, trains, and airplanes, where increasingly high complexity of software is the major contributing factor of many system failures. 

She subsequently worked as a software engineer at Motorola, where benefited from her Ph.D. research experience. She successfully led the Software Reliability championship for the Asia-Pacific region and 6-Sigma initiatives. Her subsequent research work had more focus on quality and reliability issues facing a cyber- physical system, particularly an interconnected (wireless) network. She studied factors that cause uncertainties in wireless communication among outdoor wearable devices, error localization for video transmission, error propagation over noisy channels, etc. Much of her past research work was driven by real-world issues that she encountered while working in industry.

Her future research work will build upon her past research to answer questions on how to make our increasingly smart AI systems more reliable and trustworthy. Reliable AI aims to detect, predict, and ultimately proactively prevent failures of a smart CPHS that can result from problems of its underlying hardware, software, or human inputs. Another challenging issue she would like to tackle is how to make AI systems trustworthy so that they do not cause unintentional or even intentional harm to humanity. AI is also prone to intentional or unintentional biases that can be problematic and even cause harm to certain groups of people. If the data used for training AI models have any human bias, then the results generated by AI systems will be biased. Trustworthy AI has to minimize potential inherent biases in data in order for its powerful machine learning algorithms to be effective.

2 Novel Machine Learning Paradigms and Contextual Reasoning for Augmented Intelligence

In a robust, reliable, hyper-connected and data intensive Cyber-Physical-Human-System (CPHS), human participants can be the weakest link due to attitude, fatigue, human error, and other human behavioral issues. Augmented intelligence is a conceptualization of artificial intelligence that focuses on AI’s assistant role to enhance human intelligence rather than replace it. 

Her current research activities include improving explainability of machine learning (ML) outputs through generative and ensemble approach, hybrid extensional-intensional learning, and reasoning with semantic objects for augmented intelligence. Her research experience includes applied AI algorithms, including data mining and knowledge discovery and design of machine learning algorithms for classification of multimedia data. In addition, her research addresses some fundamental limitations of current machine learning techniques by optimizing  major dimensions of AI (perception, learning, abstraction, and reasoning) towards the development of a general cyber-physical-human (CPH) contextual-conceptual (CC) model. Taken together, this explainable output is aimed at multiplying human cognitive power thereby augmenting human intelligence.

3 Intelligent Systems

Trustworthy and reliable AI Software is being explored on every continent for large-scale commercial and industrial applications in areas such as manufacturing, agriculture, and education.

3.1 Intelligent Decision Support for Smart Cyber-Physical-Human Systems

Her master thesis work involved the development of a fully funded and award winning intelligent decision support system for flight control systems. Her proposed work is in the area of real-time intelligent decision support in smart manufacturing and smart agriculture applications. One of her research interests is preventive AI. One example is how to apply ML algorithms to conduct preventive maintenance in a smart factory environment. The data generated from these smart devices and ML form the basis of this line of investigation. AI can then help an autonomous CPS make decisions on its own or an augmented CPHS that can inform human decisions. An example was demonstrated in her collaborated research project on intelligent video categorization.

3.2 Intelligent Web Applications: Intelligent Tutoring and Web Services

Her later research work involved proposing a web content recommender system based on human behavior modeling. Understanding human behavior is fundamental to augmented intelligence and applications like intelligent tutoring. Ontology is a useful knowledge representation for semantic analysis. It facilitates knowledge discovery through reasoning about concepts, which can be applied to a range of prediction problems, e.g. fault diagnosis, resources allocation, what-if analysis, budgeting, and cost control. In this area, she is interested in intelligent web applications that address business and educational needs.

Selected Publications: 

1.Augmented Intelligence supported by Web of Wisdom, G.Y. Hong, Kybernetes (under review)

2.Boosted supervised intensional learning supported by unsupervised learning, ACM Fong and G. Y. Hong, accepted by International Journal of Machine Learning and Computing, to appear in 2020.

3.Multi-prong framework toward quality-assured AI decision making, G. Y. Hong and A.C.M. Fong, in Proc. 4th International Conference on Contemporary Computing and Informatics (iC3I 2019), Singapore, December 2019.

4.Ontology-powered hybrid extensional-intensional learning, A.C.M. Fong and G.Y. Hong, in Proc. International Conference on Information Technology and Computer Communications (ITCC 2019), ACM, pp. 18-23, Singapore, August 2019.

5.Augmented intelligence with ontology of semantic objects, A.C.M. Fong and G.Y. Hong, in Proc. 4th International Conference on Contemporary Computing and Informatics (iC3I 2019), Singapore, December 2019.

6.Self-organizing communicating semantic objects for augmented intelligence, A.C.M. Fong and G.Y. Hong, in Proc. 3rd IEEE International Conference on Communication and Information Systems (ICCIS 2018), pp.189-192, Singapore, Dec 2018.

7.Short-range tracking using smart clothing sensors, ACM Fong, B. Fong, and G. Y. Hong, 3rd IEEE International Conference on Communication and Information Systems (ICCIS 2018), pp.169-172, Singapore, December 2018.

8.A Prognostics Framework for Health Degradation and Air Pollution Concentrations, B. Fong and G. Y. Hong, Journal of Advances in Information Technology. 3 (1): 64-68, 2012.

9.Generation of personalized ontology based on consumer emotion and behavior analysis, A.C.M. Fong, B. Zhou, S.C. Hui, J. Tang, and G.Y. Hong, IEEE Trans. Affective Computing, Vol. 3/2, pp.152-164, 2012.

10.Web content recommender system based on consumer behavior modeling, Fong ACM, Zhou B, Hui SC, Hong G.Y. and Do T.A., IEEE Transactions on Consumer Electronics, Vol. 57/2, pp. 962-969, 2011.

11.An intelligent video categorization engine, G. Y. Hong, Fong B and Fong ACM, Kybernetes. Vol. 34/6, pp. 784-802, 2005. 

12.Factors Causing Uncertainties in Outdoor Wearable Wireless Communications, B. Fong, P. B. Rapajic, G. Y. Hong and ACM. Fong, IEEE Pervasive Computing, Vol. 2/2, pp. 16-19, 2003. 

13.Adaptive QoS control of multimedia transmission over band-limited networks, G.Y. Hong, ACM Fong and B. Fong, IEEE Transactions on Consumer Electronics, Vol. 48/3, pp. 644 –649, 2002.

14.Challenges for Providing Mobile Multimedia Services in a Campus Environment, B. Fong, P. B. Rapajic and G. Y. Hong, IEEE Conf. Proc. ITRE, 2003.

15.Constrained error propagation for efficient image transmission over noisy channels, B. Fong, G.Y. Hong and ACM Fong, IEEE Transactions on Consumer Electronics, Vol. 48/1, pp. 49-55, 2002.

16.A modulation scheme for broadband wireless access in high capacity networks, B. Fong, G. Y. Hong and ACM Fong, IEEE Transactions on Consumer Electronics, 48 (3), pp.457-462, 2002.

17.Error localization for robust video transmission, G.Y. Hong, B. Fong and ACM Fong, IEEE Transactions on Consumer Electronics, Vol. 48/3, pp. 463-469, 2002.

18.On the performance of telemedicine system using 17 GHz orthogonally polarized microwave links under the influence of heavy rainfall, Fong B, Fong ACM and Hong GY, IEEE Trans. IT in Biomedicine, Vol. 9/3, pp. 424-429, September 2005. 

19.Six Sigma in Software Quality, G. Y. Hong and T.N. Goh, The TQM Magazine, Vol. 15/6, pp.364-373, October, 2003.

20.Modeling the Optimum Software Release Time, G.Y. Hong, Proceedings of the 5th Annual Simulation and Modeling Symposium, Ft. Lauderdale, Florida, USA, pp. 1137-1140, 1999. 

21.A Practical Method for the Estimation of Software Reliability Growth in the Early Stage of Software Testing, M. Xie, G.Y. Hong and C. Wohlin, Proceedings of the IEEE International Symposium on Software Reliability Engineering, ISSRE’97, pp. 116-123, 1997.