Abstract:

A Recommender System, (or Recommendation System, RS), is a subclass of Information Filtering systems. It is used to predict the ranking or preference of items being offered to users. The system has been used for commercial applications such as making recommendations of merchandises, entertainment items, online dating services and in academic activities, such as search for research articles, collaborators and subject experts. One of the challenges in the development of an efficient RS is Information Sparseness, in particular with new users or users with few interests.

Collaborative-based personalized model has achieved much success in addressing the issue and it replaces the traditional use of personal profiles by associating user preferences from users who have similar preferences. Collaborative interests are developed through an analysis on the interaction of users to their preferred items as the interaction influences the learning capability of the model. Therefore, the methodology to establish an effective interaction is essential for a good collaborative-based personalized model.

Convolutional Neural Network (CNN) is one of the actively used model-based techniques. It can be used for the generation of a feature embedding map for user-item interaction either with or without integration of knowledge graph. The success of CNN-based embedding relies heavily on the network design and interaction learning mechanism.

This talk aims to provide the research question, backgrounds and introduces the state-of-the-art techniques being in use. It will explore the roles of CNN as an embedding technique in collaborative-based personalization, and in particular for recommender systems and facet selection. Example results will be shown to illustrate the work so far.

Biography :

Emeritus Professor Lance Chun Che Fung was born in Hong Kong and trained/worked as a Marine Radio and Electronic Officer from 1972 to 1978. He graduated with a B.Sc. Degree with First Class Honours in Maritime Studies (1981) and a Master of Engineering Degree in System Test Technology (1982) from the University of Wales.His PhD Degree was awarded by the University of Western Australia in 1994 with a thesis on the Applications of Artificial Intelligent Techniques to Electrical Power System Engineering under the supervision of the late Professor Kit Po Wong. He taught at the Department of Electronics and Communication Engineering, Singapore Polytechnic (1982-1988), and at the School of Electrical and Computer Engineering, Curtin University of Technology (1989-2003). He joined Murdoch University in 2003 and was appointed as Emeritus Professor in 2015. In 2017, he was awarded an Honorary PhD Degree in Information Technology by Walailak University Thailand, in recognition of his contributions towards the development and advancement of their research and postgraduate programs. He have held positions as Academic Program Chair, Associate Dean of Research, Postgraduate Research Director and Director of the Center for Enterprise Collaborative in Innovative Systems. He have supervised over 30 postgraduate and doctoral students and published over 330 academic articles in international journals and conference proceedings in the areas of Neural Networks, Intelligent Systems, Computational Intelligence, Cybernetics, Electrical Power Systems, Image Processing, Data Mining, Machine Learning, Knowledge Management, Education and Web Technology.

Abstract:

Human-activity recognition (HAR) has enabled many applications including home-based behavior analysis, healthcare, and smart homes. It is an important technology for humanity, as it allows one to analyze and provides assistance for one’s daily life. It mainly comprises two types, namely, video based and sensor based. The video-based HAR uses cameras to capture videos to recognize the behaviors of people. The sensor-based HAR uses sensors, such as accelerometers and gyroscopes. It has been studied and widely used, with privacy being well protected. Many machine-learning methods have been used in the sensor-based HAR. In this talk, the efforts of the speaker regarding the sensor-based HAR using machine-learning methods have been introduced thus far.

Biography :

Makoto Koshino is an Associate Professor at the Department of Electronics and Information Engineering, National Institute of Technology, Ishikawa College, Japan. He received his master’s degree in Engineering from Kanazawa University in 2002 and joined Fujitsu Limited in 2002. He has been working at the National Institute of Technology, Ishikawa College since 2003. He received his Ph.D. in Engineering from Kanazawa University in 2004. He has been an Associate Professor at the National Institute of Technology, Ishikawa College since 2010. His awards include a grand prize in the Kanazawa Smartphone Application Contest in 2012, president's award of the National Institute of Technology in 2013, good presentation award at the Education Forum of the National Institute of Technology in 2014, and creative award at the Human Symbiosis System Design Contest in 2015. His research area covers AI and IoT.

Abstract:

For more than four decades, Moore’s law has been driving the semiconductor industry where the number of transistors per chip roughly doubles every 18-24 months at a constant cost. Transistors have been relentlessly evolving from the first Ge transistor invented at Bell Labs in 1947 to planar Si metal-oxide-semiconductor field-effect transistor (MOSFET), then to strained SiGe source/drain (S/D) in the 90- and 65-nm technology nodes, and high-k/ metal gate stack introduced at the 45- and 32-nm nodes, then to the current 3D transistors (Fin field-effect transistors (FinFETs)) trained at the 22-nm node. Thus, the miniaturization of devices so far has been possible due to changes in the dielectric, S/D, and contact materials/ processes, and innovations in lithography processes, in addition to changes in the device architecture. The gate length of current transistors has been scaled down to 14nm and below, with over 10⁹ transistors in state-of-the-art microprocessors. Yet, further scaling down the CMOS technology is leading to a more massive interconnect delay and higher power density. The complexity of physical design is also increasing with a higher density of devices. So, what is next?

To continuously face the miniaturization challenges, adapting computational modeling and simulation approach towards issues aroused in nanoelectronics devices seems to be potentially applied. The computational modeling of ballistic quantum transport in nanoscale MOSFET is presented in this talk. The modeling approach would elaborate on the fundamental theory of ballistic saturation velocity for a meaningful interpretation of nanoscale MOSFET with the data validations to the fabricated device. In addition, the computational simulation using technology computer-aided design (TCAD) tools for vertical strained impact ionization MOSFET (VESIMOS) is also discussed. The computational simulation using TCAD has been widely accepted to be accurately simulated and insight into various nanoelectronics device drawbacks prior to carrying out the device fabrication process. Thus, the cost, time, and major defects would be surpassed extensively. Besides, applying artificial intelligence (AI) techniques to make the fast and accurate simulation of ultra-scaled nanodevice is also explored in this talk.

Biography :

Ismail Saad is an Associate Professor at the Electrical & Electronics Engineering Program, Faculty of Engineering, Universiti Malaysia Sabah (UMS). He received his Bachelor's Engineering Degree from Universiti Putra Malaysia (UPM) in 1999 and a Master's Degree from the University of Southampton in 2001. In 2009, he received his Ph.D. from Universiti Teknologi Malaysia (UTM). During his Ph.D. Convocation Ceremony, he was awarded the Chancellor and Best student awards. He has been an Associate Professor at the Faculty of Engineering, UMS, since 2013. His research area covers Micro and Nanoelectronics device and materials, Digital Integrated Circuit system design, and Nanofiber membrane for energy applications. He is a graduate member of the Board of Engineers Malaysia (BEM) and Institute of Engineers Malaysia (IEM) since 2010. He was also an active member of IEEE since 2007 and was appointed as the First Chair of the IEEE Sabah Subsection in 2017. He is passionate about research and has won awards in Malaysia Technology Expo (MTE), Bio-Innovation, and UMS Research and Innovation (PEREKA) research competition from 2011 to 2016. He has secured over a million research grants with numerous dedicated collaborative research partners. He has published more than 100 Scopus index papers as primary and corresponding authors. In administration experience, he proves active involvement with holding a position as head of the Artificial Intelligence Research unit, Program head of Electrical & Electronics Engineering, Deputy Dean of Academic & Internationalization, and currently as a Dean in Faculty of Engineering, Universiti Malaysia Sabah (UMS).