Keynotes

Enhancing Security in Artificial Intelligence Systems, Professor Yang Xiang 

Abstract – The integration of machine learning (ML) and artificial intelligence (AI) technologies in different industries has brought about significant changes and countless opportunities. However, this progress has also raised concerns regarding the security of ML and AI systems. To address these concerns, substantial efforts have been made to improve the resilience and reliability of machine learning models. 

In this presentation, we will explore the complex landscape of securing machine learning, discussing the challenges, strategies, and future directions in this rapidly evolving field. We will provide an overview of the current research and emerging trends related to secure and trustworthy ML and AI systems. Specifically, we will highlight the vulnerabilities that can be exploited by malicious actors to compromise ML and AI systems. Furthermore, we will present various approaches that can enhance the security of these systems in real-world scenarios. 

We will also touch upon the ethical considerations associated with deploying ML systems based on large foundation models, emphasizing the importance of accountability and transparency. By drawing on experiences from software vulnerability detection, we will demonstrate how similar approaches can be adapted to the security requirements of machine learning. 

Biography Professor Yang Xiang received his PhD in Computer Science from Deakin University, Australia. He is currently a full professor and the Dean of Digital Research, Swinburne University of Technology, Australia. His research interests include cyber security, which covers network and system security, data analytics, distributed systems, and networking. In the past 20 years, he has been working in the broad area of cyber security, which covers network and system security, AI, data analytics, and networking. He has published more than 300 research papers in many international journals and conferences. He is the Editor-in-Chief of the SpringerBriefs on Cyber Security Systems and Networks. He serves as the Associate Editor of IEEE Transactions on Dependable and Secure Computing, IEEE Internet of Things Journal, and ACM Computing Surveys. He served as the Associate Editor of IEEE Transactions on Computers and IEEE Transactions on Parallel and Distributed Systems. He is the Coordinator, Asia for IEEE Computer Society Technical Committee on Distributed Processing (TCDP). He is a Fellow of the IEEE. 


Human-Centric Activity Recognition, Associate Professor Hailing Zhou 

Abstract – Human activity recognition stands as a pivotal domain, enabling machines to comprehend and respond to human behavior in diverse contexts. With the understanding of human actions, it will facilitate more natural and intuitive interactions between humans and machine/robots. This can lead to improved user experiences across various applications, such as smart environments, collaborative robots, and assistive technologies 

The session will show a journey of human activity recognition development from human gesture recognition to triplet-based human-object interaction recognition, then to symbolic-neural networks, and mostly current to multi-modality approaches such as llava. The journey also witnesses the development of computer vision from a perception level understanding to a cognition level understanding. 

Biography – Dr. Hailing Zhou is an Associate Professor in the Department of Mechanical and Product Design Engineering. She received her PhD from Nanyang Technological University, Singapore. Her research interests focus on Robotic Vision, AI, and the engineering applications. 
She has over three years industry working experience, project delivery and team leadership, including a R&D Principal role in Accenture leading solution deliveries to real-world problems (such as quality control, human behavior analysis, forecasting) in manufacturing and retail domains. Prior to that, she has over seven years of academic experience at Deakin University. She was a senior research fellow on robotic vision, with successful grants and established collaborations with universities and institutes such as DSTG, MIT, USF, and AIMS. She has over 40 publications including those published in high-impact journals such as IEEE Transactions on Image Processing, Pattern Recognition, IEEE Transactions on GeoScience and Remote Sensing and IEEE Transactions on Intelligent Transportation. She also served as an Associated Editor of the IEEE Intelligent Transportation Systems Conference and a program committee member of the IEEE International Conference on Systems, Man, and Cybernetics (SMC). 


Human-Centric Generative and Geospatial AI for Urban Digital Twins, Professor Soheil Sabri 

Abstract The digital transformation of modern cities, driven by the integration of advanced information, communication, and computing technologies, has ushered in a new era of data-driven smart city applications for efficient and sustainable urban management. However, despite their advantages, these applications often depend on large volumes of high-dimensional, multi-domain, and location-based data to monitor and analyze various urban subsystems. This reliance poses challenges in areas where data quality and availability are limited, and where generating urban scenarios and design alternatives is costly. Generative Artificial Intelligence (GenAI), and Geospatial Artificial Intelligence two growing fields within deep learning, have shown great potential in content generation and spatially explicit analysis respectively. This presentation explores how GenAI and GeoAI techniques can be innovatively integrated with Urban Digital Twins to tackle challenges in planning to improve people’s quality of life, focusing on key urban subsystems such as transportation, energy, water, and critical infrastructure. The presentation demonstrates how the intelligence and autonomous capabilities of Generative Adversarial Networks (GAN), Variational Autoencoders (VAE), and the Generative Pre-trained Transformer (GPT), and Geopose enabled Training Data Markup Language (TDML) can be leveraged to enhance research, operations, and the management of critical urban subsystems.

Biography –  Dr. Soheil Sabri is an Assistant Professor in Digital Twin at the School of Modeling, Simulation, and Training at the University of Central Florida. He is the director of the Urban Digital Twin Lab, a research facility dedicated to exploring the integration of emerging technologies such as Digital Twins, IoT, AI, and the Metaverse in the realms of Environmentally Sustainable Design (ESD) and Urban Quality of Life (UQoL). Soheil has extensive experience in applied research, often in collaboration with industry and government partners. Over the past two decades, he has provided consulting services to national and international agencies, including work with countries like Australia, Singapore, South Korea, Saudi Arabia, and Columbia facilitated through organizations like the World Bank, Open Geospatial Consortium, and the United Nations Development Program. In addition to his research and consulting efforts, he plays a vital role as the Ambassador and Co-chair of the Academia and Research at the Digital Twin Consortium. In this capacity, he is a staunch advocate for leveraging technological innovation to drive smart and sustainable urban development, ultimately promoting social innovation and progress. Soheil holds an honorary senior research fellow position at the Department of Infrastructure Engineering, The University of Melbourne, Australia.

 


HMI Design for Automated Vehicles, Professor Ronald Schroeter 

Abstract This talk explores cutting-edge developments in Human-Machine Interface (HMI) design within the automotive domain, focusing on enhancing driver engagement and safety in automated vehicles. Leveraging research into Head-Up Displays (HUDs) for non-driving related tasks and human-centric approaches to automation, the talk examines how intelligent interface solutions can improve responsiveness to critical situations, and foster more effective human-automation collaboration in complex environments. 

The presentation will also provide an overview of ongoing projects and emerging trends in the field of automated vehicles, including efforts to promote inclusivity—particularly for older adults and individuals with disabilities—through innovative HMI design. It will highlight work in intelligent agents, developed in collaboration with Seeing Machines through the Empathic Machines project, which aims to create automated systems that respond adaptively to human emotional and cognitive states. 

Biography – Professor Ronald Schroeter is a Principal Research Fellow and Chair in Empathic Machines at QUT. He embraces multidisciplinary research across Human-Computer and Human-Machine Interaction with focus on automotive user interfaces and automated driving), design, human factors and road safety. He has won seven Australian Research Council competitive grants, including a prestigious DECRA fellowship, three ARC Discovery projects, one ARC LIEF, and leads applied research with industry partners such as Seeing Machines Pty Ltd on Driver Monitoring Systems. He also co-led the successful ARC Training Centre bid for Automated Vehicles in Rural and Remote Regions (AVR3) as Co-Director with 28 partners from academia, industry and government.


High-fidelity motion simulators for virtual vehicle prototyping and driver performance analysis, A/Professor Houshyar Asadi 

Abstract Almost 1.3 million people die in road accidents globally each year, and an additional 50 million are injured or disabled globally. In this respect, the use of motion simulators for user training, automotive research & development and prototyping can significantly improve road safety and reduce the number of fatalities. In addition, simulators are the safest and most cost-effective tools not only for training, but also for evaluating new vehicle designs (including driver-based and autonomous vehicles), human performance and perception of comfort and trust in autonomous vehicles. Similar benefits apply to other domains, such as trains, aeroplanes, and ships.

The majority users of driving (and flight) simulators experience motion sickness to some degree, as the existing motion simulators fail to deliver the most accurate driving (or flight) sensation to users. As a result, the key benefits of motion simulators for driver training, virtual vehicle prototyping, and human behaviour analysis cannot be fully realised. Acting as the “brain” of a motion simulator, the Motion Cueing Algorithm (MCA) is responsible for the generation of high-fidelity vehicle motion cues within the simulator’s limited physical workspace through a washout filter, such that realistic driving/flying sensations can be delivered to the user.

This keynote will highlight the key challenges and innovations in simulator technologies including MCAs, demonstrate its applications across industries, and discuss how high-fidelity motion simulators are shaping the future of vehicle design and user performance analysis. Through research findings, we will underscore the importance of this technology in accelerating the development of next-generation vehicles and improving user safety and performance.

 

BiographyAssociate Professor Houshyar Asadi is currently an Australian Research Council (ARC) DECRA Fellow and leads research in the area of Artificial Intelligence (AI)-based motion simulator technologies. He also leads the Motion Simulation Technologies Lab, where his research focuses on AI-driven Driving and Flight Motion Simulator platforms, Motion Cueing Algorithms (MCA), and Human Performance Assessment in immersive Virtual Environments (VE) such as Immersive Simulations, Interactive Virtual Worlds, Virtual Reality (VR) environments, and Mixed Reality (MR). His research on human performance in VEs includes driver and pilot distraction analysis, behavior prediction, cognitive load assessment, emotion recognition, motion perception, and motion sickness analysis within virtual driving and flight scenarios.

He received his Bachelor of Engineering degree (First Class Honors) in Electrical-Control Systems, his Master’s degree in Industrial Electronic and Control Engineering (High Distinction), and his Ph.D. in the research area of human perception-based washout filtering for motion simulators using AI in 2008, 2012, and 2015, respectively.


Human-Centred Testing and Understanding of Intelligent Software Systems, Dr. Huai Liu 

Abstract We are rapidly entering an age where the advanced artificial intelligence (AI) capabilities are integrated into the contemporary software systems to fulfil critical tasks. Like “traditional” software, intelligent systems are not immune to software faults. In fact, considering their applications into safety-critical domains, some faults in intelligent systems have caused catastrophic disasters. Software testing, the mainstream approach to software quality assurance and control via detecting faults, is confronted with much more severe challenges in the context of AI-based software, because of its unique characteristics, such as the black-box nature and non-deterministic behaviours. In this session, we are going to discuss how to develop effective testing techniques that are easy-to-use even for non-technical stakeholders, such as end users and domain experts, thus providing different perspectives for verifying and validating intelligent systems. As identified by Australian government, besides the more “technical” characteristics like reliability and safety, AI-based practices should also show some “human-centred” values such as diversity, fairness, and contestability. To enhance these human-centred features of intelligent systems, another key topic of this session is the development of good understanding and explanation especially from the perspectives of various stakeholders, most of whom do not have the professional knowledge of AI models and algorithms. 

BiographyDr. Huai Liu is a Senior Lecturer in Department of Computing Technologies at Swinburne University of Technology, Melbourne, Australia. He received the BEng and MEng degree both from Nankai University, China, and the PhD degree from Swinburne University of Technology, Australia. Dr. Liu has worked as a Lecturer at Victoria University and a Research Fellow at RMIT University. He has also worked as an engineer in the IT industry. Dr. Liu is a senior member of the Institute of Electrical and Electronics Engineers (IEEE) and a member of the Association for Computing Machinery (ACM). 
Dr. Liu’s major research interest include:
• Software Engineering, in particular, advanced software testing methodologies and cost-effective software fault tolerance;
• Services and Cloud Computing, in particular, quality assurance of Web services and their compositions; and
• Internet of Things, in particular, spatiotemporal modelling for physical systems as services and data analytics & image processing for optical microscopy.
Dr. Liu is widely published in various international journals and conferences including IEEE Transitions on Software Engineering, ACM Computing Surveys, IEEE Transactions on Computers, and International Conference on Software Engineering. He has also conducted a variety of industry projects. His industry partners range from small and medium-sized enterprises to multinational corporations. He has applied his expertise in IT into various areas, such as business and accounting systems, platform as a service, manufacturing robots, and industrial engineered systems. 


 

Empowering Aerial Robots through Physics-Informed Machine Learning, Professor Mir Feroskhan 

AbstractPhysics-Informed Machine Learning is transforming aerial robotics by embedding physical laws directly into neural networks that model the dynamics of aerial robots. This innovative approach bridges the gap between traditional mathematical models and purely data-driven techniques, offering a more efficient and accurate way to control aerial robots in real-time. In this keynote, I will demonstrate how Physics-Informed Neural Networks (PINNs) significantly reduce data requirements and improve performance by incorporating fundamental dynamic principles directly into the learning process. Unlike data-driven models, which rely on large datasets and provide no inherent understanding of the physical systems, PINNs leverage established physics to capture the complex behavior of multi-rotor aerial robots with greater precision. This enables up to a 65% reduction in data needs while effectively managing uncertainties in dynamic environments, such as wind disturbances and visual servoing errors. 

Building on my research in visual servoing for multi-rotor control, I will show how PINNs can handle up to 70% of system uncertainties, enabling real-time control tasks to be performed up to 10 times faster than conventional models. Visual servoing, where camera-based feedback controls the movement of the robot, is particularly sensitive to uncertainties in both the robot’s dynamics and the camera system. By embedding physics into the neural network’s cost function, PINNs offer robust and efficient solutions, making them ideal for autonomous flight in unpredictable conditions. Beyond visual servoing, PINNs have vast potential in more complex aerial tasks, such as modular parcel delivery and electric vertical takeoff and landing (eVTOL) operations. For example, in decentralized multi-agent tasks like collaborative parcel delivery, PINNs efficiently model multi-agent control strategies. Additionally, their ability to model unsteady aerodynamics during critical phases of eVTOL flight, such as hover-to-cruise transitions, will be discussed. This talk will offer insights into the latest advancements, current research, challenges and future directions, positioning Physics-Informed Machine Learning as a key driver in the next generation of intelligent, adaptable, and efficient aerial robots, capable of executing complex tasks in ever-changing environments. 

Biography: Professor Mir Feroskhan received his Bachelor’s degree with First Class Honors in Aerospace Engineering from Nanyang Technological University (NTU), Singapore, and earned his Ph.D. in Aerospace Engineering from the Florida Institute of Technology, USA. He continued postdoctoral research at NTU and later at École Polytechnique Fédérale de Lausanne (EPFL), Switzerland, working in the Laboratory of Intelligent Systems under the supervision of Professor Dario Floriano. 

Currently, as an Assistant Professor in the School of Mechanical and Aerospace Engineering at NTU, Prof. Feroskhan leads the Intelligent Cybernetics Group. His research focuses on integrating artificial intelligence with robotics, particularly in the modeling and control of aerial robots, including avian-inspired drones and multi-agent systems. 

Prof. Feroskhan has received notable grants and awards, including A*STAR’s Young Individual Research Grant in 2022 for his pioneering AI-driven UAV solutions, EDB’s OSTIn Grant in 2023 for the porposed development of an onboard, mobile vision-based space situational awareness system, and the Dr. S. K. Leung Excellence Award in 2023 for outstanding teaching. He also serves as the Project Lead for UAS Traffic Management (UTM) at NTU’s Air Traffic Management Research Institute (ATMRI), where he plays a vital role on the management committee. In addition, he is an active member of the IEEE community, where he serves on the Robotics and Intelligent Sensing Technical Committee within the IEEE Systems, Man, and Cybernetics (SMC) Society. Prof. Feroskhan also serves on the editorial board of Scientific Reports (Nature). 


Cooperative and Highly Automated Driving (CHAD), feedback from a 5 years study on Automated Driving, Professor Sebastien Glaser 

Abstract: The Cooperative and Highly Automated Driving (CHAD) Safety Study is a long initiative to understand the safety impact of automated vehicle (AV) systems in Australia. This project, led by the Centre for Accident Research & Road Safety – Queensland (CARRS-Q) in collaboration with iMOVE Australia and the Queensland Department of Transport and Main Roads, involves the development and testing of a Level 4 Electric Cooperative and Automated Vehicle (CAV) named ZOE2. This presentation will focus on two specific work packages: the interaction with the driver (WP1) and the demonstration activities and challenges encountered (WP4). WP1 focuses on the driving task transition in automated vehicles, exploring public participants’ experiences in a controlled setting and assessing their acceptance of and trust in AV technology. WP4 targets public awareness and perception, aiming to increase knowledge and acceptance of AVs and develop a large activity on demonstration on public roads in Australia. This presentation will focus on the technical details for implementing the ADS specific to the encountered environment. 

Biography: Sebastien Glaser is a Professor at CARRS-Q, QUT, (since 2018) and Centre Director for the ARC Training Centre on Automated Vehicles in Rural and Remote Regions, having previously contributed to and led research projects & teams in Europe. He has significant experience working with the automotive industry. His research focuses on the safe interaction between automated vehicles and other road users. He is managing the iMOVE CRC Cooperative and Highly Automated Driving project and several extensions, with the Australian Research Council, and the Queensland Department of Transport and Main Roads, looking at the deployment of automated vehicles and the safety of automated technologies for their uses in on-road applications. . 


Intelligent Data Analytics Systems, Professor Chee Peng Lim 

Abstract: In this talk, computational intelligence (CI)-based models for undertaking data analytics and decision support problems will be presented. The underlying computational frameworks exploit the capabilities of individual and hybrid data-based learning models, which include artificial neural networks, fuzzy systems, and evolutionary algorithms. Ensemble architectures are also leveraged to further enhance robustness and efficacy of the resulting frameworks. In addition, the importance of incremental learning without suffering from the catastrophic forgetting problem in perpetual data-based learning environments will be highlighted. Application of the resulting CI-based systems to a number of decision support problems in industrial and healthcare domains will be demonstrated.

Biography: Professor Chee Peng Lim received his Ph.D. degree from the University of Sheffield, UK in 1997. His research focuses on CI-based models and their applications. He has published over 600 research papers, edited over 10 books, and received more than 10 best paper awards in international conferences. In collaboration with co-researchers, he has developed multiple award-winning CI-based software tools. These include Gold Medal at Invention and New Product Exposition, Pittsburgh, USA, Gold Medal and Special Award at British Innovation Show, UK, Gold Medal at Geneva International Exhibition of Inventions, Switzerland, and Silver Prize at Open Source Software World Challenge, South Korea.


Bringing Order to Chaos: Innovation Management for the Modern Age, Dr Khoh Soo Beng

Abstract: In a world where rapid technological advancements and shifting market demands continually challenge organisations, managing innovation is no longer optional—it’s essential. This talk explores how structured innovation management can transform an organisation’s approach to creativity and change, turning abstract ideas into tangible, value-driven outcomes.

Attendees will gain insights into the principles behind innovation management systems, such as ISO 56001, and learn how these frameworks create order by integrating cross-functional collaboration, risk management, and stakeholder engagement. The talk will examine how companies use innovation management to move beyond ad-hoc experimentation, building repeatable processes that harness creativity, increase agility, and drive sustained competitive advantage.

Organisations can unlock new growth opportunities and stay ahead in a fast-evolving landscape by aligning innovative efforts with strategic goals. Whether you’re an engineer, project manager, or business leader, this session will equip you with actionable strategies for fostering a structured, dynamic innovation culture that thrives on creativity and clarity

 

Biography: Dr Khoh Soo Beng is an innovative leader with a passion for technology and a drive to solve challenges. With 28 years of experience across Fortune 500 companies, academia, and start-ups, Dr Khoh has expertise in R&D, manufacturing, IT, project and change management, innovation management and new product development.

Starting as an embedded systems engineer at Rover Advanced Technology Centre, Dr Khoh developed pioneering automotive technology, including the CANbus for vehicles and DeviceNet for factory automation at Allen Bradley. These early projects emphasised the importance of interoperability and real-time operations.

At Motorola Penang, Dr Khoh was a catalyst for change, transforming manufacturing into a digital, paperless process. His integration of MES, ERP, and Digital Six Sigma expert systems significantly boosted efficiency and productivity. His initiatives established data-driven decision-making as the core of operational excellence and set a new standard for digital transformation in the industry.

As a Motorola Penang Design Centre leader, Dr Khoh introduced Design for Six Sigma (DFSS) to train junior engineers in high-quality product development, earning him recognitions such as DFSS consultant black belt, Quality Star, Inventor mentor and Innovation Champion of the Year.

Driven by a deep commitment to societal impact, Dr Khoh has dedicated his career to leveraging technology to improve society. His work with CREST to promote connected health, co-founding Digital Health Malaysia, and supporting numerous start-ups all underscore his unwavering commitment to making a difference. His co-founding of PMO Innovations, a boutique consultancy focused on project management, innovation, and sustainability, further demonstrates his dedication to societal progress.

Dr Khoh now serves as Malaysia’s chair for the ISO56002 Innovation Management System, merging technology with strategic business solutions for clients. His technical interests include IoT, automotive CAN bus, and Industry 4.0, focusing on MedTech, digital health, and sustainability. He invented the 5-Step G.R.E.E.N. methodology to train engineers on implementing carbon reduction and sustainability-oriented innovations.

Holding a PhD in Electrical Engineering from the University of Warwick, UK, Dr Khoh is a member of professional bodies such as IET, IEM, and IEEE, with over 50 published papers. He is an associate editor of the International Journal of Innovation Science (IJIS). His expertise continues to influence innovation and business success.


THE CONTINUING TECHNOLOGY EVOLUTION WIRELESS TOWARDS 6G: A TECHNICAL OVERVIEW OF 6G RESEARCH ITEMS, Edwin Li

Biography: Edwin Li is a sales engineer at Rohde & Schwarz responsible for industry, component and research (ICR) for Victoria, Tasmania and, South Australia.

He has a degree in applied science (microprocessor applications) and an engineering degree in electronic telecommunication (photonics).

His first role at Hewlett Packard Instrument (1998) kick off his interests in test and measurement (T&M). Edwin has held varied roles including application engineer, marketing, business development and project management, with geographical coverage from Asia Pacific to a global product overview for the test and measurement industry.

I would like to take this opportunity to share with you my experience, my passion in engineering,and the wonderful journey of 6G mm Wave technical research and development.