Keynote Speaker I
Dr. JIE ZHANG
Newcastle University, UK
Speech Title: Modelling and Optimisation of Batch Processes Using Machine Learning Techniques
Abstract: Batch processes are suitable for the agile manufacturing of high value added products, such as specialty polymers, pharmaceuticals, and bio-products. In contrast to continuous processes, batch processes have strong nonlinear behaviour and always operated in transient states. A further difficulty in batch process control is that product quality variables usually cannot be measured on-line and can only be obtained through laboratory analysis after a batch has finished. This talk presents several robust neural network based data driven modelling, inferential estimation, and reliable optimal control methods for batch processes. Bootstrap aggregated neural networks have enhanced model generalisation capability and can also provide model prediction confidence bounds. One of the most important issues of empirical model based batch process optimal control is that the calculated optimal control profile can degrade very significantly when applied to the actual process due to model plant mismatches. In order to address this issue, the optimisation objective function can be augmented by an additional term (or an additional objective in multi-objective optimisation) to penalise wide model prediction confidence bound at the end-point of a batch. By such a means, the calculated optimal control profile is much reliable in the sense that, when being applied to the actual process, the degradation in control performance is limited.
Bio: Dr Jie Zhang received his PhD in Control Engineering from City University, London, in 1991. He has been with the School of Engineering, Newcastle University, UK, since 1991 and is currently a Reader in Process Systems Engineering and Degree Programme Director for MSc in Applied Process Control. His research interests are in the general areas of process system engineering including process modelling, batch process control, process monitoring, and computational intelligence. He has published over 290 papers in international journals, books, and conference proceedings. He is on the Editorial Boards of a number of journals including Neurocomputing published by Elsevier, PLOS ONE and International Journal of Automation and Control.
Keynote Speaker II
Dr. Jeong Jin Hong
Manager, Data Scientist, Applied Materials, USA
Speech Title: Data-driven Applications for Semiconductor/Display Manufacturing
Abstract: The semiconductor and display manufacturing industries have many different processing steps involved and a variety of sensors techniques are applied to the process tools to gather potential useful information as many as possible. It allows us to imagine that there would be a huge amount of data available. The data-driven approaches including machine learning methodologies are actively being studied and being developed to obtain better product quality monitoring or control methods tackling limitations of the conventional methods such as the univariate sensor monitoring. Now it is applied for many different problems in the manufacturing environment from automatic defect detection to quality prediction. Some examples will be discussed. One of main applications widely applied to the manufacturing site is automated defect detection and classification using deep learning methods. There are several inspection steps during the production process and most of them are carried out manually by human inspection engineers. Many researches demonstrate superior detection performance of DL models. Another application using machine learning techniques is fault detection & diagnosis of the process tool. Process tools normally have number of sensors which are highly correlated and the level of difficulties to monitor every sensor’s behaviors with conventional methods. The machine learning techniques can extract certain patterns that are useful for fault detection and engineers now are able to apply it to their workflows. Machine learning approaches can be beneficial in speed as it can detect faults earlier and faster. However, there is always a limitation we need to overcome and this limitation is how well we can interpret the model’s results. To overcome this, we have to understand the physical background of the process and try to integrate model result with the physics to obtain meaningful solutions.
Bio: Jeong Jin Hong, Ph.D. has been a data science manager at AMAT since 2019 with data-driven model algorithm development using various types of data collected from the tool. He is currently leading a group of data scientists developing machine learning-based solutions for internal & external R&D. Prior to joining AMAT, Dr. Hong was a principal engineer position for Samsung Display. During his 8 years at Samsung, Dr. Hong spent time leading a research group developing machine learning-based solutions for the OLED manufacturing line such as prediction modeling, inspection image classification, APC modeling, and fault detection modeling. His prediction model-based solution development project has been awarded the company’s annual best performance award by CEO. Some of his other projects also got technical awards from CTO and CEO at Samsung Display. Dr. Hong received the B.Eng. degree in Chemical Engineering from Hongik University, Seoul, South Korea in 2006. And he received the M.Sc. degree with Distinction and the Ph.D. degree in Chemical Engineering from Newcastle University, Newcastle upon Tyne, the United Kingdom in 2007 and 2011 respectively. He was a recipient of several scholarships including the industrial studentship between Newcastle University and Syngenta, the global Swiss company in agricultural industry. And one of his papers was also awarded as the best paper from the EU-Korea Conference in Engineering and Science. His main research interests include machine learning algorithms for the semiconductor industry, plasma-assisted process monitoring, advanced process control, interpretable deep learning modeling and physics integrated machine learning modeling.
Keynote Speaker III
Dr. Xin-She Yang(Reader in Modelling and Optimization)
Design Engineering & Maths, Middlesex University London, UK
Speech Title: Nature-Inspired Algorithms: Challenges and Opportunities
Abstract: Nature-inspired algorithms have become effective tools for solving optimization problems concerning engineering designs, data mining and computational intelligence. Swarm intelligence based algorithms, such as particle swarm optimization, cuckoo search and firefly algorithm, have demonstrated their flexibility and effectiveness. However, many challenging issues still exist. This talk highlights some of the recent developments, applications, key issues and future research directions.
Bio: Xin-She Yang obtained his DPhil in Applied Mathematics from the University of Oxford. He then worked at Cambridge University and National Physical Laboratory (UK) as a Senior Research Scientist. Now he is Reader in Modelling and Optimization at Middlesex University, and the Chair of the IEEE CIS Task Force on Business Intelligence and Knowledge Management. He has been on the prestigious list of Highly Cited Researchers (Web of Science) for consecutive four years (2016, 2018, 2018 and 2019).
Keynote Speaker IV
Hans Ulrik Staehr
CEO, MarketScape, Denmark
Bio: Hans Ulrik Staehr is the CEO of the market intelligence software company MarketScape. He has more than 17 years of experience working with the largest companies in Europe. Three years ago Hans Ulrik founded Munit.io, a company that is now a market leader in software for dark web investigations for law enforcement agencies. Hans Ulrik is security cleared and works with some of the worlds leading agencies hunting criminals on the dark web.
Hans Ulrik Staehr is security cleared at the highest level in Nato by the Danish and Swiss authorities (CTS/Cosmic Top Secret), and he has extensive experience with deep- / dark web intelligence projects both with law enforcement agencies and among corporations in finance, energy/utilities, retail, pharma & telcos. During the last years some examples of his speaks:
- Risk Connect Conference in Warsaw. November 2019.
- Swiss Cyber Security Days. February 2019.
- 16th Austrian IT & Consultants Day. November 2018.
- DeepINTEL & DeepSEC, Vienna. November 2018.
- Digital Economic Forum, Zurich. DEF2018.
- SIPUG Day. Swiss Information Providers user group 2017.
- Europol November 2019. Event with leading oil companies regarding card fraud.
- ISS World from Telestrategies. Dubai 2017 and Prague 2018
| Dr. JIE ZHANG Newcastle University, UK Multivariate Statistical Process Performance Monitoring |
| Prof. Peng Yang Hebei University of Technology, China Motion intention recognition and advanced control strategies for wearable exoskeletons |
 | Prof. Massimo Marchiori (UNIPD), Technical Director (EISMD) University of Padua, Italy; European Center for Science, Media and Democracy, Belgium
Cyber Physical Systems and Data Science: the public and private challenge |
