Seminars and Defenses

All are welcome and encouraged to attend the seminars and defenses.

Mar. 26, 1PM, ENG460
Dr. Amir H. Ashouri • Research Seminar
Software Automatic Tuning using Machine Learning
The presentation will focus on the state-of-the-art approaches on optimizing compilers when machine learning is applied. The techniques primarily enhance the quality of the obtained results, and more importantly, make it feasible to tackle the two major compiler optimization problems: optimization selection (choosing which optimizations to apply) and phase-ordering (choosing the order of applying optimizations). Additionally, it will highlight the recent trends in optimizing deep learning applications for more efficient inference tasks.

Dr. Amir H. Ashouri is a Postdoctoral researcher at ECE Department of University of Toronto under the advisement of Prof. Tarek Abdelrahman. His research interests are Deep Learning, Compiler Optimizations, and Automatic Tuning. Prior to joining the U of T, he completed his M. Sc. (2012) and Ph.D. (2016) at Polytechnic University of Milan under the supervision of Prof. Cristina Silvano. His Ph.D. thesis was selected as the winner of IEEE-Italy section best Ph.D. thesis of 2016 and recently, a book entitled "Automatic Tuning of Compilers Using Machine Learning" has been adapted from his Ph.D.
Mar. 21, 10AM, ENG460
Dr. Rebecca Feldman • Research Seminar
Ultra-high field neuroimaging: technical development and translational applications
Ultra-high field magnetic resonance imaging (MRI), such as systems operating at 7T, can produce images of the human body and brain non-invasively, but with outstanding resolution and contrast. The signal to noise ratio (SNR) of MRI increases proportional to the strength of the magnetic field, thus at 7T, higher resolution is possible, given a reasonable scan time. This increased SNR can be parlayed into improvements for a number of imaging sequences. However, the SNR benefit at 7T cannot yet be fully realized due to technical challenges such as inhomogeneity of the applied radio frequency (RF) field, and increased RF power deposition, measured in the specific absorption rate (SAR), at high fields. We will discuss the development of new pulse sequences which decrease the impact of RF field inhomogeneity while also reducing SAR deposition. We will also discuss the application of 7T to human neuroimaging and data analysis techniques which have revealed lesions indicating potential seizure onset zones in epilepsy and other biomarkers of epilepsy in patients with normal clinical MRI exams.

Dr. Rebecca Feldman studied electrical engineering in the Engineering Science program at the University of Toronto. She obtained her PhD from the University of Western Ontario, and is a P.Eng, registered in Alberta. She was selected as a Junior Fellow of the International Society for Magnetic Resonance in Medicine and is currently a senior scientist in the Advanced Neuroimaging Research Program at the Translational and Molecular Imaging Institute. Dr. Feldman's research seeks to amplify the power of magnetic resonance imaging (MRI) to better detect and diagnose illness. Her work is a balance of MR engineering and translational neuroimaging. High field MRI can produce images of the brain with amazing detail and exquisite contrast. However, there are technical challenges associated with imaging at high fields. Dr. Feldman develops new sequences meet these technical challenges and allow us to leverage the power of magnetic resonance imaging to sensitively detect new biomarkers and better treat neurological disorders and diseases.
Mar. 20, 10AM, ENG460
Dr. M. Ali Tavallaei • Research Seminar
Systems and Devices for Image Guided Cardiovascular Interventions
Cardiovascular diseases are a leading global cause of death. Many types of such diseases are diagnosed and treated by image guided cardiovascular interventions. In these procedures, diagnostic or therapeutic devices (e.g. catheters) are introduced into the vasculature and, under the guidance of x-ray fluoroscopic imaging, they are navigated to various target locations for diagnosis or delivery of therapy. The extreme flexibility of such interventional devices, their remote manipulation from outside the patient body, and their guidance with 2D x-ray fluoroscopy, creates various limitations in positioning, tracking and effective navigation of such devices and ultimately leads to high failure rates and long procedure times. In this talk we will focus on technologies and devices developed for addressing these problems in two main clinical applications of interest: cardiac catheterization, for the treatment of cardiac arrythmia (failure rates>25%); and angioplasty, for the treatment of atherosclerosis (failure rates>15%). With the goal of improving the success rate of cardiac catheterization procedures, we will discuss the benefits of using MRI to guide such procedures and study the associated hindering challenges. We will describe the design, development, and evaluation of an MRI-compatible tele-robotic catheter navigation system that addresses the problem of patient access during MRI-guided interventions. We will also describe recent developments in ultra-wideband radar hardware and will discuss its potential applications for estimation and compensation of cardiopulmonary motion in image-guided cardiac catheterization. With the goal of improving the success rate of angioplasty procedures, we will review the challenges of conventional angioplasty techniques and analyze their root causes of failure. We will then describe the design, development, and evaluation of a device and navigation platform (CathPilot), that is intended to address these challenges. The talk will conclude with an overview of potential future research plans and opportunities.

Dr. Tavallaei obtained his BSc and MSc degrees in Electrical Engineering from Urmia University and University of Tabriz correspondingly. He obtained his PhD degree in biomedical engineering in 2015. His PhD research was carried out at the Robarts Research Institute at Western University. During his PhD studies he was a trainee of the NSERC Computer-Assisted Medical Intervention program. In parallel to his PhD studies, he co-founded and led Vital Biomedical Technologies, a spinoff from Robarts that commercialized the world's first fully MRI-compatible motion platform for quality assessment of MRI and MRI-guided therapy. After his PhD studies, he completed the Medical Device Innovation Fellowship Program at Western University and University of Minnesota. He is currently a postdoctoral fellow at the Sunnybrook Research Institute (SRI) and his research focuses on minimally invasive therapy and diagnosis of cardiovascular diseases. Dr. Tavallaei is also the co-founder and president of Magellan Biomedical, a spinoff company from SRI, that focuses on addressing major clinical needs in cardiovascular interventions.
Mar. 18, 12PM, ENG106
Dr. Hassan Kojori • Research Seminar
From More Electric to Hybrid/All Electric Aircraft
The More Electric Aircraft (MEA) is based on the concept of utilizing electrical power for driving aircraft subsystems currently powered by hydraulic, pneumatic or mechanical means including utility and flight control actuation, environmental control system, lubrication and fuel pumps, and numerous other utility functions. In this seminar, Dr. Kojori begins with an overview of the More Electric Aircraft and will discuss how various technologies developed over the past three decades have helped reduce the size, weight and life-cycle-cost of the overall system, significantly improve reliability and ease manufacturing and maintenance. Next he will cover emerging advanced technologies for Hybrid and All Electric Aircraft for urban transportation and discuss some of the main opportunities and challenges.

Dr. Hassan Kojori holds a PhD from the University of Toronto and is an IEEE Fellow and licensed Professional Engineer in Ontario. He has over 30 years of experience in the field of power conversion, power distribution, energy storage and related systems optimization and control. Currently, as a Senior Principal Engineer with Honeywell, he is the Conversion Portfolio Leader for Aero Advanced Technologies responsible for research, development and technology demonstration of advanced Electric Power Systems for More Electric Aircraft and tactical vehicles. His original work on numerous technology firsts has resulted in more than 45 patent disclosures (28 granted), several trade secrets and more than 50 technical papers and proprietary industry reports. Dr. Kojori has been actively engaged in collaborative research in the general area of power electronics, Lithium-Ion battery energy storage systems and teaching and supervising graduate students with leading local and international universities for over 20 years. He was adjunct professor in the Department of Electrical and Computer Engineering (ECE) at the University of Toronto and Ryerson University for over 10 years and collaborated as an industry professor in the Institute for Automotive Research and Technology at McMaster. Currently, he is Associate Editor, IEEE Transactions on Transportation Electrification, advisory board member for ECE department at Ryerson University, University of Toronto Institute for Multidisciplinary Design & Innovation, Queen's Centre for Energy and Power Electronics Research (ePOWER), and represents Honeywell at The Downsview Aerospace Innovation and Research Consortium.
Mar. 13, 10AM, ENG460
Dr. Adam Santorelli • Research Seminar
Emerging Electromagnetic Medical Devices: From Concept to Implementation
Medical devices are increasingly investigated in the academic sector with the same rigor present in a commercial setting ensuring a realistic chance of developing a usable technology that will ultimately improve patient outcomes. This talk will give an overview of the development of several emerging technologies that capitalize on advances in wireless technologies to address clinically informed needs. From the high-frequency microwave range to the low-frequency kilohertz range, a contrast in the electrical properties exists between different biological tissues. These differences can be exploited in the development and application of minimally invasive electromagnetic-based medical technologies. Specifically, the design, fabrication, and the clinical testing pathway of prototypes addressing specific clinical needs, including breast cancer screening, brain bleed detection, bladder monitoring, and neutrophil monitoring during chemotherapy will be addressed. This presentation will conclude with a discussion on future, unexplored avenues of research for such electromagnetic-based technologies.

Dr. Adam Santorelli is an IRC Government of Ireland Postdoctoral Fellow in the Translational Medical Device Lab at the National University of Ireland, Galway. He studied at McGill University, Montreal, Canada, where he received his B. Eng, M. Eng, and Ph.D. in Electrical Engineering in 2010, 2012, and 2017, respectively. He has been with the Translational Medical Device Laboratory at National University of Ireland Galway since 2016, initially as a visiting researcher, and full-time as a postdoctoral researcher since 2017.
Mar. 12, 10:30AM, ENG460
Dr. Huiyan Li • Research Seminar
Deciphering the molecular code of diseases: translational proteomics and bioinformatics tools for cancer research
Cancer is a very complex disease. The study of molecular complexity in cancer development and progression requires high-throughput analytical technologies and bioinformatics tools to understand the disease mechanisms. Every cancer is different, and precision approaches have shown promise to improve its diagnosis, prognosis, and treatment. In this talk, I will present the engineering of novel micro-technologies and bioinformatics tools to study cancer proteomics. Using these technologies, I identified promising cancer biomarkers from existing cancer database, discovered potential biomarker panels for personalized diagnosis, and quantified cancer signaling pathway activities for evaluating targeted therapy. In the end, I will briefly discuss the outlook of my research field and future research plans, aiming at deciphering the molecular code of cancer and promoting better clinical management and patient outcome.

Dr. Huiyan Li is currently a CIHR postdoc fellow at Massachusetts General Hospital, Harvard Medical School. She got her PhD in Biomedical Engineering from McGill University, working on developing novel micro-technologies and bioinformatics tools for measuring proteins in cancer samples. After that, she joined the UVic-Genome BC Proteomcis Center as an NSERC postdoc fellow, where she developed novel technologies for the diagnosis of cancer and cardiovascular diseases. She has a Master degree in Electrical and Computer Engineering from the University of Victoria, and a Bachelor degree in bioinformatics from China.
Mar. 4, 12PM, ENG460
Dr. Sharif Razavian • Research Seminar
Why do we move the way we do? Reciprocal benefits of engineering and human motor control
Human movements are captivating. We can dance, lift heavy weights, or touch a delicate petal without damaging it; all with the same body. The nervous system can orchestrate these diverse movements with such ease that we hardly notice the effort. However, the human musculoskeletal system is unimaginably complex: a massively redundant system, with non-linearities, time-delays, and uncertainties. Despite these complexities, the movements are controlled precisely, efficiently, and in real-time. Kinematic and dynamic observations of human movements give us a wealth of information about the features of the motion, but little insight into how they are generated. Fortunately, the engineering tools available in robotics and control have the potential to help unfold the mystery of how the nervous system controls the movements. The benefits of taking a mathematical approach to identify the properties of the human motor control system are multi-faceted. Firstly, we can mimic the biological control algorithms to design more agile and efficient robots that match human capabilities. Additionally, understanding how humans interact with objects and their environment allows us to build intelligent bio-mechatronic systems that can safely interact with the users in a human-like manner. Conversely, the engineering insights into human motor control will expand the physiological knowledge about the human body and enhances clinicians' understanding of movement-related disorders. Consequently, more effective injury-prevention strategies and superior therapies can be administered in the light of these findings. In this presentation, I introduce my past research on the mathematical modelling of the human motor control system, as well as various cases where these models were used for advanced control of bio-mechatronic systems. I will also explain how I foresee my research will grow in the next few years as a result of integrating expertise in engineering and human movement sciences.

Dr. Reza Razavian is a Research Associate in the Department of Bioengineering at Imperial College London, UK. With expertise in control and dynamical systems analysis, his primary research interest is to understand how the human nervous system controls the body movements. He graduated from the University of Waterloo in 2017, with a PhD degree in Systems Design Engineering. His dissertation was on mathematical modelling of the human motor control system, with application to rehabilitation engineering. Dr. Razavian has been engaged in a diverse range of projects, from optimal control of hybrid electric vehicles to modelling of neuronal ensembles. His recent projects include the development of an advanced robotic hand simulator for studying human hand biomechanics, a model-based algorithm to detect abnormalities in pathological gaits (e.g. Parkinson's Disease patients), and an intelligent controller for an upper extremity rehabilitation robot.
Feb. 22, 10AM, ENG460
Dr. Pooya Bagheri • Research Seminar
Data-driven Assessment and Control of Electrical Energy Systems
Electrical power systems are among the largest systems ever built and operated by human beings. Due to its large size, a complete power system can be never tested in a laboratory scale. Hence, these systems have been traditionally assessed by developing circuit models and conducting model-based studies such as load flow and stability analysis. Algorithms for effective and reliable operation have been also designed by relying on circuit models. In recent years, we can witness a significant development in measurement, communication and information technologies available to the power industry. This opportunity has allowed for new perspectives in testing and control of the energy grids. At present, measurement data can be used to calibrate and verify the circuit models used for the operation of a system. In a more direct approach, data can be even used to replace the conventional role of circuit models as the foundation for assessment and control strategies. This talk is devoted to such a data-driven approach for electrical power systems. This approach is particularly beneficial at the power-distribution level where comprehensive and accurate circuit models are hard to attain. The other benefits, as well as challenges with this new approach, are also discussed. A new theoretical framework is presented as a foundation for data-driven methods in power distribution systems. Using the proposed framework, a number of data-driven schemes are developed and presented such as model-free volt-var-control and model-free conservative-voltage-reduction. The promising results of initial feasibility studies on the test systems will be also shared. The talk will be concluded by depicting a roadmap for an innovative research program in this area.

Pooya Bagheri obtained the B.Sc. degree in electrical engineering from Sharif University of Technology, Tehran, Iran, in 2010, and the M.Sc. and Ph.D. degrees in Energy Systems from the University of Alberta, Edmonton, Canada, in 2013 and 2019, respectively. He was an electrical power engineer at Stantec Consulting Ltd., Edmonton, Canada from 2013 to 2015. As a researcher at the University of Alberta, he has also contributed to several projects for power utility companies such as ATCO, EPCOR, and FortisAlberta. During his Ph.D. studies, Pooya has been awarded the NSERC Alexander Graham Bell Canada Graduate Scholarship, the Alberta Innovates Technology Futures (AITF) Scholarship, and the President's Doctoral Prize of Distinction. His research interests are power system data-analytics and power quality.
Feb. 19, 10AM, ENG471
Amr Adel Fathy Mohamed • PHD final thesis defense
Line-Wise Power Balance Equations And Applications
Optimal power flow (OPF) refers to a class of optimization formulations that optimize power systems considering a chosen objective and a set of constraints. The objective might be to minimize generation costs, or to minimize transmission losses, etc. The set of constraints might include equality constraints such as bus-wise power balance equations, and, inequality constraints such as limits on magnitude of bus voltage phasors, limits on reactive power output of generators, etc. Existing accurate OPF formulations used to settle electricity markets include a set of bus-wise power balance equations that is comprised exclusively of 3rd or 4th order terms (i.e. terms with three or four products) and all of those terms have a sinusoidal component. Accordingly, such OPF formulations remain nonlinear and nonconvex mathematical optimization problems. Even though commercial implementations of OPF algorithms are robust and efficient, due to their nonlinear and nonconvex nature, they still cannot guarantee a global optimum. The US Federal Energy Regulatory Commission estimates that the best commercial OPF solvers are off by 10%, amounting to an economic loss of US $400 billion per year worldwide. In addition, the continuous necessity to ensure voltage stability of power systems while they are being optimized, necessitates incorporation of voltage stability constraints into OPF formulations. For these motivating technical and economic reasons, OPF remains a major research focus and forms the topic of this thesis. To address the nonlinear and nonconvex nature of accurate OPF formulations, this thesis aims to: (1) develop new sets of power balance equations with lower order terms and lesser numbers of sinusoidal terms yielding faster power flow algorithms with lower order solution space that are better suited for OPF applications, (2) build new OPF formulations using this new set of power balance equations, and (3) incorporate voltage stability constraints into the newly developed OPF formulations. The genesis of the new set of power balance equations stems from: (1) the fact that power of a constant impedance load is proportional to the square of voltage magnitude, and, (2) that power flow in transmission lines and transformers can be expressed in terms of square of magnitude of voltage phasors. Accordingly, a set of line-wise power balance equations is developed, both in polar and rectangular forms. The solution algorithms of these two formulations are developed using Newton-Raphson technique and are tested on several IEEE and real power system data up to 9241 busses. Tests show that the proposed line-wise power flow (LWPF) algorithms are accurate, provide monotonic convergence, and scale well for large systems. They are up to twice and thrice faster for polar and rectangular forms respectively, when compared with conventional bus-wise power flow (BWPF) methods that use bus-wise power balance equations. Analysis of BWPF and LWPF methods shows that the LWPF methods require much lower numbers of calculations, resulting in higher speed.
Feb. 15, 10AM, ENG460
Dr. Amir Abiri Jahromi • Research Seminar
Towards Sustainable and Smart Electricity Grids
Electric power systems are at the verge of significant technology transformation moving toward a smarter operating environment in which system components are enabled to engage in optimal power system operation and planning. This is happening while the uncertainties in power systems are rapidly increasing due to the massive integration of renewable energy resources, electrification of road transportation, and continuously aging power system legacy assets. Concurrently, the cyber vulnerability of power systems is on the rise by the increased reliance on distributed control and information technology. These rapid shifts in the electric power industry should be addressed properly. Otherwise, power system economics and reliability will be challenged by disruptive and costly electricity interruptions. In this talk, we address the challenges and opportunities facing the future smart electricity grids. First, we demonstrate the increasing need for the deployment of flexibility resources, and advanced control and monitoring systems to facilitate the integration of renewable energy resources and adoption of electric vehicles. Next, we show that the extensive reliance of these advanced control and monitoring systems on information and communication technologies opens up new cybersecurity vulnerabilities. Co-simulation platforms are considered as a cost- effective solution for studying the vulnerabilities of smart grids to cyberphysical attacks. The importance of data analytics and artificial intelligence for addressing the cyber vulnerability of smart grids is further underlined.

Currently a postdoctoral researcher at the University of Toronto, Amir Abiri Jahromi received his Ph.D. degree in Electrical Engineering from McGill University in 2016. He was a postdoctoral researcher at the University of Denver in 2017. His research interests are in the fields of cyber-security, protection, reliability, and optimization of smart electricity grids.
Feb. 14, 2PM, ENG460
Dr. Mohammadreza Fakhari Moghaddam Arani • Research Seminar
Effective Utilization of Distributed and Renewable Energy Resources to Stabilize and Enhance Smart Power Grids Performance
Clean and renewable energy resources such as wind power and photovoltaics, along with advanced environmentally-friendly loads like electric vehicles, will be significant components of future grids. These resources are rapidly increasing, but they have an inherent tendency to degrade system performance due to their intermittent generation, complicated system dynamics, and distributed nature with an increasing penetration level. This talk focuses mainly on studying and analyzing voltage and frequency dynamics in grids with these resources contributing to frequency and/or voltage stabilization.

Connecting power-electronic-interfaced generation, particularly in some offshore wind plants, to long transmission lines with very high-impedances can be challenging. Benefiting from a comprehensive analytical model, a detailed analysis of the voltage source converter (VSC) dynamics is presented and showed how the assumptions which are usually made for designing VSC regulators in conventional grids are not valid in grids with high impedances. Two solutions which enabled the VSC to operate at its maximum theoretical active power by making minimal modifications in the widely accepted control method are proposed and compared. This talk also discusses analyzing and improving the performance of the proposed methods under faults. Analyses are used not only to show the factors which threaten the successful low-voltage ride-through of the converter, but also to find a solution to improve the converter performance with a minimum change in the control structure and parameters, while satisfying required standards. This talk also focuses on frequency regulation in smart grids. The studies on asymmetrical distributed single-phase sources and loads such as photovoltaics and electric vehicles show that a central controller may result in delays and that a distributed control may lead to asymmetry. Both of these controllers may destabilize the system. This talk discusses the disadvantages of independent utilization of wind generators and electric vehicles for frequency regulation and come up with a coordinated control method in which these two sources compensate for each other weakness. In summary, this talk shows that neglecting the impact of emerging power system components on system stability will lead to large and complicated problems in the near future. The research work then evolved to develop new mechanisms to mitigate the adverse impacts of the new components on system stability, thereby contributing to the creation of a sustainable future.

Mohammadreza Fakhari Moghaddam Arani received the M.Sc. from University of Waterloo, Waterloo, Canada, in 2012, in electrical engineering and the Ph.D. degree in Energy Systems in the Department of Electrical and Computer Engineering, University of Alberta, Edmonton, Canada, in 2017. In 2012, he worked as a research associate in University of Waterloo. He is currently a post-doctoral fellow with University of Toronto, Toronto, ON, Canada. He has been awarded the Alberta Innovates Technology Future (AITF) Scholarship and Natural Science and Engineering Research Council of Canada (NSERC) postdoctoral fellowship. His paper on implementing frequency support controllers in wind power generation has been recognized by the IEEE Transactions on Energy Conversion Editorial Board as a high-quality work in 2017. His research interests include cyber-physical security of smart grids, renewable and distributed generation, plug-in hybrid electric vehicles, microgrids dynamics and control, power system stability, and power electronics.
Feb. 8, 10AM, ENG460
Dr. Talal Halabi • Research Seminar
Toward Intelligent Cybersecurity
Cyberattacks are rapidly evolving in complexity and power. Artificial Intelligence has already begun to shape these attacks by promoting their automation and exploding their scale. On the other hand, our critical infrastructures including transportation, energy production, and health are becoming significantly reliant on coupling the Cyber and physical worlds. This renders our economy and safety highly dependent on the security of emerging smart cities and involved Cyber-Physical Systems. Such paradigms are constantly altering our perception of Cybersecurity and revolutionizing our visions on defense and protection. All of this makes the reliance on yesterday’s incidents to detect and mitigate tomorrow’s attacks a lost cause. Today’s security mechanisms require a certain level of smartness to be realistic, which is reflected in their ability to adhere to specific system requirements such as decentralization and large-scale, to make strategic and contextual decisions, to adapt to unexpected situations, and to be proactive. This can be achieved by giving these mechanisms the ability to learn. This talk will highlight the critical role of intelligence integration into the design of security solutions and will give new insights on smartening next generation defenses in IT infrastructures and Cyber-Physical Systems through advanced research ideas and examples related to Cloud Computing and Intelligent Transportation Systems. More specifically, the talk will emphasize on several emerging concepts such as cooperative security and Machine Learning-based security for attack prevention, mitigation, and recovery. The talk will be concluded by sharing a tentative vision of establishing a new research lab at Ryerson’s Electrical and Computer Engineering Department, leading innovative research at the intersection of Cybersecurity and Artificial Intelligence.

Talal Halabi is a postdoctoral fellow at the School of Computing of Queen's University, Canada. He received his B.Eng. in Electrical and Electronic Engineering from the Lebanese University in 2012, his M.Sc. degree in Computer Networks from the Lebanese University and Saint-Joseph University in Lebanon in 2013, and his Ph.D. in Computer and Software Engineering from the Polytechnic School of Montreal in 2018. His doctoral thesis focused on addressing the challenges to security evaluation and operational integration in Cloud Computing. It presented numerous contributions of high quality and relevance and was nominated for the best thesis award in his department. Dr. Halabi’s current research mainly involves the design of smart security solutions for emerging Cyber-Physical Systems including Intelligent Transportation Systems and the Internet of Things. His research interests lie in the broad areas of Cybersecurity, network security, Cloud security, information systems security, Cyber-Physical security, trust management, cooperative and distributed security, Game Theory for security, learning for security, and data-driven security, and his research methodologies involve operations research, algorithm design, machine learning, and game and decision theories. Dr. Halabi has published in journals and conferences of very high impact, and has served as a program committee member for several conferences of high quality.
Feb. 6, 10AM, ENG460
Dr. Reza Samavi • Research Seminar
Cybersecurity: AI for Security or Security for AI?
In recent years the areas of Artificial Intelligence (AI) and Cybersecurity, have both received tremendous attention from the research community. Interestingly these two areas mutually benefit from each other. On the one hand, recent advances in AI have enabled new security capabilities. For example, machine learning, and specifically deep learning methods, have helped researchers develop more accurate and effective intrusion, malware or spam detection systems. On the other hand, AI solutions are becoming major components of the modern economy and as such AI algorithms are targets for attackers who intend to compromise their integrity and confidentiality. AI needs to be reliable, and security can enable a reliable AI system. In this talk I will discuss why conducting research at the intersection of cybersecurity and AI is important and timely. I will describe a number of projects that we are currently working on in this area and the challenges that we are facing. I will argue that if we mainly focus on AI as an enabler for security, and not sufficiently invest on security as an enabler for AI, we may lose the premise of reliable AI systems (including using AI for cybersecurity). I will explore research opportunities in the application areas of machine learning to computer security problems, and the vulnerabilities in software systems that are exposed by machine learning methods. I will conclude my talk with a roadmap for a research program that enables advancing the state-of-the-art in security applications with AI, and secure AI applications.
Feb. 4, 12PM, ENG106
Prof. Chang Wen Chen • Research Seminar
Internet of Video Things (IoVT): Next Generation IoT with Visual Sensors
The worldwide flourishing of the Internet of Things (IoT) in the past decade has enabled numerous new applications through the internetworking of a wide variety of devices and sensors. More recently, visual sensors has seen their considerable booming because they usually capable of providing richer and more versatile information. Internetworking of large scale visual sensors has been named Internet of Video Things (IoVT). IoVT has its own unique characteristics in sensing, transmission, storage, and analysis, which are essentially different from conventional IoT. These new characteristics of IoVT are expected to impose significant challenges to existing technical infrastructures. In this talk, an overview of recent advances in various fronts of IoVT will be introduced and a broad range of technological and system challenges will be presented.

Chang Wen Chen is currently Dean of School of Science and Engineering at the Chinese University of Hong Kong, Shenzhen. He is also an Empire Innovation Professor of Computer Science and Engineering at the University at Buffalo, State University of New York since 2008. He was Allen Henry Endow Chair Professor at the Florida Institute of Technology from July 2003 to December 2007.  He was on the faculty of Electrical and Computer Engineering at the University of Rochester from 1992 to 1996 and on the faculty of Electrical and Computer Engineering at the University of Missouri-Columbia from 1996 to 2003.