As cybersecurity continues to be a worldwide concern, ADSC researchers are working to improve upon and apply a technology that has been proven to be reliable in identifying and analyzing security attacks – a honeypot – to industrial control systems.

According to ADSC Senior Research Scientist Daisuke Mashima, a honeypot is a “decoy” system designed to lure cyber-attackers in hopes of detecting attack attempts early, slowing down or mitigating the impact of attacks, and gathering real-world attack traces for learning attack vectors and designing better cyber security systems.

“Enterprise IT security has used honeypots for a long time, but in industrial control or smart grid systems, the design and use of honeypots is relatively new and immature,” Mashima said.

Industrial control systems, used to automate industrial processes, are the backbone of the power grid, manufacturing facilities, and water treatment plants, among other critical infrastructures. According to Mashima, as more smart grids are being deployed in Singapore, the United States, and other countries, smart grid security is still in its early stages and this method is considered effective in providing additional security.

In November 2018, Mashima received a one-year grant, funded by the Singapore Cybersecurity Consortium under the National Research Foundation, to develop a high-fidelity industrial control systems (ICS) honeypot in the smart grid domain. Mashima has partnered with Custodio Technologies in Singapore to extend research he previously did to develop a state-of-the-art smart grid honeypot system implemented on top of an open-source, software-based smart grid testbed system developed by ADSC.

“The current implementation of the system is minimal, and our honeypots are lacking some realistic aspects from an attacker’s perspective, and that’s why we’re working with Custodio Technologies,” Mashima said. “Their expertise is penetration testing to find vulnerabilities in the system, so they will help us see how our honeypots look to attackers. Based on that, we can improve honeypot systems and effective logging mechanisms to later analyze what the attacker is doing in the system.”

Now, the researchers have a simple prototype that they’ll spend the year improving, with an end goal of making the prototype ready for practical use for immediate data collection and data analysis. According to Mashima, the threat intelligence collected by the smart grid honeypot enables engineers and researchers to design and implement cyber-physical security mechanisms for countering emerging threats against our critical infrastructure.

There are currently open source industrial control systems and enterprise IT honeypots available, but they provide only a cyber view, showing what ports are open to the Internet and what ICS protocols are supported. Mashima’s proposed honeypot will provide a realistic cyber and physical view to the attacker.

“An attacker may have a good amount of knowledge as to whether a system is fake or not, so to fool an attacker, we need to provide a realistic physical side view,” Mashima said. “We want to retain the attacker inside a system for as long as possible to study their attack behavior for a longer time.”

Mashima will also be partnering with National University of Singapore to incorporate NUS’s experience implementing virtual ICS networks, as well as deploying the project outcome on the National Cybersecurity R&D Lab (NCL).

“Through this partnership, our work will expand the coverage and comprehensiveness of the implemented smart grid honeypot system,” Mashima said. “NCL is a great environment for evaluating scalability and sharing the outcome with other researchers.”

ADSC researchers have been celebrating successes in emotion recognition for several years, from ranking among the top three in facial expression recognition competitions at ACM ICMI 2015 and 2016 to incorporating a spin-off, Opsis, that has been commercializing the technology in the domains of human resources, marketing, customer service and others.

Recently, ADSC researchers Songyou Peng, Zhang Le, and Stefan Winkler placed second overall in an emotion recognition challenge held in conjunction with the 2018 IEEE World Congress on Computation Intelligence and the International Joint Conference on Neural Networks.

The One-Minute Gradual-Emotion Behavior Challenge (OMG-Emotion) had teams evaluate one-minute video clips and identify the speakers’ emotions in the videos. The participants were provided with video, audio and text and results were evaluated on their arousal and valance values. The ADSC researchers placed second overall in valence prediction accuracy and first in vision-only arousal prediction.

The researchers’ techniques were unique from other work being done in that they used a pre-trained deep network that’s designed for face verification on large-scale datasets. Additionally, they employed a randomized sparse sampling strategy to look at the frames of the videos to avoid over-fitting and they used Long Short Term Memory networks (LSTM) to aggregate temporal information from each video.

This methodology which was submitted for the OMG-Emotion challenge was developed with support from Digital Emotions, a project by A*STAR’s Science and Engineering Research Council (SERC).

“Emotions can be characterized by the two dimensions of valence and arousal,” Winkler said. “Valence is positive versus negative emotions and arousal measures passive versus active emotions. We measure these on a scale of -1 to 1, which allows us to distinguish small changes in expressions and recognize subtle emotions.”

The dataset provided was a collection of annotated YouTube videos around one minute in length. Each video was selected because of its emotional behaviors, such as facial expressions and language context.

“Previous work in emotion recognition has mostly focused on predicting a limited number of emotion categories, rather than more complex emotion labels, such as arousal and valence, which are much more fine-grained,” Zhang said.

“Our method for the challenge was an application of our already developed research methods,” Songyou said. “We feel this research is exciting from a machine intelligence perspective. Once you have audio or video, you can teach a machine to make automatic predictions on emotion attributes and there are lots of applications for that.”

About Project Digital Emotions

Digital Emotions is a project which aims to develop technological and scientific capabilities in performing multi-modal, multi-lingual emotion analysis of naturally occurring human expressions. Accounting for Asia’s sociocultural context, the project will build a next-generation integrative system that is capable of recognising emotions from visual and non-visual cues, such as in audio and language, for a broad range of potential applications such as in media, consumer, healthcare, finance, transport, hospitality, and professional services. The research team comprises scientists and engineers from A*STAR’s Institute of High Performance Computing (IHPC), Institute for Infocomm Research (I2R), and Illinois at Singapore Pte Ltd’s Advanced Digital Sciences Center (ADSC). Digital Emotions is a project spearheaded by A*STAR’s Science and Engineering Research Council.

You might not think of high-performance computing and high programmability as things that go hand in hand—ADSC researcher Deming Chen and CSL researcher Wen-mei Hwu, or another member of their recently funded Dynamic Data-Aware Reconfiguration, Integration and Generation (DDARING) team.

Led by the Georgia Institute of Technology, DDARING also has team members from the University of Michigan, the University of Southern California, and the University of Illinois at Urbana-Champaign. The different universities are contributing a total of eight researchers with distinct areas of expertise, who will work on different parts of the effort to improve how computers handle information.

“We needed a team working cooperatively to address this big challenge,” said Chen, an electrical and computer engineering (ECE) professor. “That’s why we formed a team among the best researchers across different universities working together to achieve the goal. We want to make sure the whole is bigger than the sum of the parts.”

Currently, many computing platforms have the flexibility to work with many applications and a variety of operating systems, but they aren’t as high-performing as computing platforms that use ASIC chips. An ASIC (“application-specific integrated circuit”) chip is designed to perform a specific task with very high performance and efficiency, but it can perform only that single task; it has no programmability or flexibility. The DDARING team is set to pursue the challenging goal of approaching the high processing capability of an ASIC chip, without losing the programming flexibility of standard computing platforms.

“To overcome this critical challenge, our approach is to build reconfigurable software through smart compiler techniques targeting dynamically reconfigurable hardware to achieve near-ASIC performance and efficiency,” said Chen.

Chen and Hwu’s piece of the DDARING project is developing a program in which dynamic changes of different algorithmic phases and data patterns take place. Hwu is creating highly optimized code for the application. He is making sure that the program—a compiler—selects the right code versions for the application and executes the functions properly on the dynamically reconfigured hardware. Chen is using machine learning to train the compiler on what to do while the algorithm phases and data are changing. That will involve offline training that uses machine learning techniques, and online inferencing that makes code optimization and hardware configuration decisions. The resulting decisions will then be passed to the hardware, and a check will ensure that the hardware is configured as it should be.

“The main two parts of our tasks are figuring out how to reconfigure the hardware, and then compiling and generating the code to take advantage of the reconfigured hardware,” said Hwu, who is the acting head of the ECE department.

Hwu likens the process to trying to reconfigure a highway to reduce congestion. Normally, of course, highways have the same number of lanes on each side, under the assumption that equal amounts of traffic will go in the two directions. But if a large traffic imbalance between the two sides were to occur, in theory you could reduce congestion by allowing traffic on the overloaded side to use more than half of the road’s lanes.

“Although it is extremely difficult to achieve that kind of reconfiguration architecture for a highway, it is actually totally possible when we are in the computer architecture field,” Hwu said.

The project is funded by the Defense Advanced Research Projects Agency (DARPA) and is expected to take two and a half years, with a possibility for further extension. It is part of the Software Defined Hardware (SDH) program under the new DARPA Electronics Resurgence Initiative (ERI).

Inspirit IoT, a startup company founded by researchers from the Advanced Digital Sciences Center (ADSC)and the University of Illinois, recently secured late-seed funding from Senscape Technologies, another startup company founded by an ADSC alumnus. The seed funding will help Inspirit IoT grow and establish a research partnership between the two companies, enabling the team to use its combined expertise to advance smart sound and smart vision applications.

The Internet of Things (IoT) is a network of connected devices that interact via the Internet. The IoT promises to bring sensor intelligence to a vast array of applications. Inspirit IoT was co-founded by Professor Deming Chen of ADSC, the Coordinated Science Lab, and the Dept. of Electrical and Computer Engineering; the company focuses on sensor intelligence and on providing hardware accelerators, coupled with an advanced high-level synthesis and machine learning optimization toolkit, to enable smart IoT applications. The technology developed by Inspirit IoT includes the DNN Optimizer, the Xcelo Compiler, and Smart Sound solutions.

The Deep Neural Network (DNN) Optimizer can take a trained neural network, go through analysis and optimization, and produce a hardware-friendly representation of the same DNN. The DNN is compressed so it can be mapped to hardware efficiently, with high performance and low power, without loss of accuracy. The Xcelo Compiler takes the compressed DNN and generates Verilog code, which allows even software engineers to design hardware. The results produced by this high-level synthesis tool are higher quality than those of market competitors, according to Chen. Further, Inspirit IoT has created Smart Sound solutions that include a smart soundboard that can localize and enhance sounds. The company is currently developing machine-learning models that can detect specific sounds related to security and use that information as the basis for identifying dangerous events, increasing monitoring of the environment, and informing authorities about dangerous situations.

This area of research has won Chen, with ECE graduate students Xinheng Liu and Dae Hee Kim, a best paper award at the 2018 IEEE/ACM International Workshop on System-Level Interconnect Prediction. Their paper outlines a backward pipeline scheduling idea that effectively enables layer overlapping across multiple layers of a DNN and reduces latency significantly compared to conventional pipelining solutions, according to Chen.

In addition, a team of Illinois, IBM, Inspirit IoT and Boeing researchers placed third in the FPGA category of the System Design Contest at the 2018 Design Automation Conference. Sixty-one teams participated in the contest, whose goal was to detect miniscule objects in a video shot from a UAV with the highest accuracy and lowest energy consumption. Collaborators on the effort included Chen, Illinois’ AMD Jerry Sanders Chair of ECE Wen-mei Hwu, IBM’s Jinjun Xiong, Inspirit IoT’s Kyle Rupnow, and Boeing’s Huafeng Yu, along with Illinois postdoc Cong Hao and graduate students Yuhong Li, Sitao Huang, Xiaofan Zhang, and Tianqi Gao.

“While we’re proud of our third-place finish, we’re not completely satisfied,” Chen said. “We realized we still have some room to further improve in terms of the DNN design, optimization, and implementation. We are carrying out some new ideas now.”

Inspirit IoT was co-founded by Chen along with two former ADSC research scientists, Rupnow and Swathi Gurumani. Rupnow serves as Inspirit IoT’s chief technology officer, and Gurumani is its vice president of engineering. The company grew out of the work that Chen, Rupnow and Gurumani began in 2011 at ADSC, which is a University of Illinois research center in Singapore. Their work aimed to create energy-efficient hardware accelerators in an automated way, and to increase hardware design productivity and quality. Their technology was commercialized in 2016, and the company now has 10 employees and operations in the United States, Singapore and India.

Senscape Technologies CEO Hongbo Xiao worked at ADSC in 2013, with Chen, Rupnow and Gurumani’s overlapping time with Xiao there. Senscape Technologies is a leader in embedded artificial intelligence (AI) sensing platforms and has the goal of giving machines and applications perception and intelligence. The new funding will help Inspirit IoT grow over the next 18 months and foster collaborations between it and Senscape.

“We established mutual trust and mutual interest and realized that the two companies’ technologies and strengths complement each other,” Chen said. “Now we have Inspirit’s software design tools and Senscape’s hardware platforms, and as partners, we’ll find new products and meet the future market needs. Actually, we already identified a couple of new product concepts with very good potential.”

Chen added that such industry and academia partnerships, especially on challenging issues, are extremely valuable to both sides.

“We realize that when we come up with an elegant solution, we can significantly improve the state of the art and change the status quo and improve people’s lives,” Chen said. “This would require strong teamwork involving both brain power and practical problem-solving skills. Like in the competition, when companies help by providing not only funding, but also expertise and key insights, it enables graduate students and smaller companies like Inspirit IoT to revolutionize how we can achieve the great potential of IoT technology.”

The University of Illinois has received funding for a new research program in Singapore – the CREATE Programme for a Trustworthy and Secure Cyber-Plexus (TSCP). The new initiative will work to make information systems both trustworthy, meaning it behaves as expected even during disruptions, and secure, or hardened against malicious attacks.

TSCP is funded for five years by Singapore’s National Research Foundation and will focus on four main thrusts, primarily as they relate to the power grid:

  • system architecture and the organization of trustworthy and secure cybersecurity components and their interactions;
  • development of new hardware, software, protocols and applications;
  • standards, verification and validation of components;
  • and methodologies and tools to continuously monitor, detect and react to problems in real-time.

The TSCP CREATE Programme will be led by David Nicol, the Franklin W. Woeltge Professor of Electrical and Computer Engineering and director of the Information Trust Institute. Zbigniew Kalbarczyk, a research professor of electrical and computer engineering in the Coordinated Science Laboratory, will share the leadership role.

“We are looking to provide a much deeper check of what the power grid system is allowed to do and if its behavior is correct, safe, and secure, as well as develop new technology to solve security challenges,” Nicol said. “We’ll also be working with monitoring the information being gathered by a system to develop ways to identify the important from the unimportant.”

The CREATE Programme partners the University of Illinois with the Singapore University of Technology and Design. TSCP also will collaborate with the National University of Singapore (NUS), Nanyang Technological University, the Institute for Infocomm Research, and other research agencies at Singapore’s Agency for Science, Technology and Research (A*STAR).

The TSCP CREATE Programme will build upon the research and relationships that were established during the past nine years in which the Advanced Digital Sciences Center has operated in Singapore. ADSC has been a top research center in information technology, data and video analytics, and smart grid cybersecurity arenas. The center was founded by the University of Illinois at Urbana-Champaign in 2009, and its researchers have received nearly 30 best paper awards, more than a dozen contest victories, 21 technology licenses, and 12 spinoff entrepreneurial ventures. ADSC’s future research on cybersecurity will focus on the CREATE Programme’s four main thrusts. Externally funded projects will strive to enhance and go beyond these boundaries.

“Through the leadership of ADSC Director Doug Jones and former directors Marianne Winslett and Ben Wah, ADSC became one of the premier research and innovation programs in Singapore,” said Andreas Cangellaris, dean of the University of Illinois College of Engineering. “Doug was especially instrumental in creating a future home for ADSC’s research in the TSCP CREATE Programme and we look forward to continuing Illinois’ tradition of creating societal impact in Singapore.”

CREATE (Campus for Research Excellence and Technological Enterprise) is an international research campus and innovation hub funded by NRF. CREATE programmes focus on research in human, energy, environmental and urban systems, resiliency, smart cities, sustainability, and other areas.

“As a member of the Illinois at Singapore PTE board, former Associate Director of ADSC, and an ADSC faculty member since its inception, I’m excited about the prospect of continuing ADSC’s cybersecurity work and the chance for Illinois faculty to partner with additional researchers at SUTD and at other organizations in Singapore,” said William H. Sanders, head of Illinois’ Dept. of Electrical and Computer Engineering. “The CREATE TSCP Programme will be a great new home for ADSC.”

As CREATE TSCP’s director, Nicol will be responsible for leveraging Illinois’ strength in cyber security and cyber resiliency in Singapore. Nicol is an ACM Distinguished Lecture and an IEEE and ACM fellow. He received the 2007 ACM SIGSIM Distinguished Contributions Award. His research focuses on modeling and simulation of large-scale systems, network security, and high-performance computing. He holds an MS and PhD in computer science from the University of Virginia and was a professor of computer science at the College of William and Mary and at Dartmouth before coming to Illinois in 2003.

“We think this is a really exciting opportunity to focus on an area we’re keenly interested in and is of critical importance to Singapore,” Nicol said. “We’re looking forward to leveraging the research that’s done in Illinois to enhance the research done in Singapore and using what we glean in Singapore to aid our research in Illinois. We see Singapore as a very promising testbed where we can test systems and scale ideas in ways we can’t in the United States. Illinois is excited to look forward into the next five years with NRF as our sponsor, SUTD as our collaborator, and Singapore as our continued place of research.”

ADSC recently received a grant to develop an integrated safety-security engineering approach and best-practice guideline for unmanned aerial systems (UAS) in Singapore. The 12-month project, which will begin on December 1, received S$99,500 from Singapore’s National Research Foundation’s Singapore Cybersecurity Consortium.

UAS usage is increasing at a rapid pace, as more companies are using drones for last-mile delivery, infrastructure inspection, land surveys, surveillance, and search and rescue. Drone usage in Singapore poses unique challenges as Singapore is a dense city with a large airport taking up a large amount of land space and often has unpredictable tropical weather. Monitoring dynamic factors, such as flight plans, weather, civil events and building model updates, are essential to ensuring UAS operation safety. Furthermore, with the increasing risk of cyberattacks, it is important to design an UAS traffic management system that ensures both safety and security.

“This topic is something we have a lot of interest in and we have done research in similar areas such as power grids and trains, so unmanned aerial vehicles are a natural extension of our results in a new domain,” said ADSC Senior Research Scientist Binbin Chen.

Currently, various users in Singapore are beginning to explore the creation of a central management system for UAS that will schedule and manage UAS traffic and monitor the status of the drones while in flight. ADSC researchers hope to make contributions to the establishment of best practices, as the system is being developed.

The ADSC researchers will be creating guidelines for these systems and addressing the safety and security issues regarding UASs. They will develop an integrated safety-security approach and study the effectiveness of their approach by applying it to the UAS traffic management engineering process, in the future.

“When you look at safety, you don’t worry about the attacker, but you worry about the human errors that might occur, for example,” Chen said. “When looking at security, you are concerned with attackers that might want to do harm via your system. Many measures and practices have been built to ensure safety, but because security is a much more recent concern. It leaves many systems safe, but not secure.”

The researchers are partnering with Nova Systems, who are able to help them put the researchers’ ideas into actual designs and apply their methods in real life. Nova is a technical consulting firm that provides a range of services, including systems engineering support, cyber security, and smart technology development. They are currently involved with the research and development of systems to support multiple drone operations in Australia and Singapore.

“In manpower and land scarce Singapore, UAS will potentially be a key part of the Smart Nation initiative to transform how people work, play, and live,” Nova Systems Managing Director Ryan Lee said.

Chen added that working with Nova will allow the researchers to participate in the design, development, and testing of the traffic management system.

“We’ll bring our safety and security engineering methods and apply them in the process, observe how well they work and learn how we can improve our methods,” he said.

In addition to Nova, Chen is collaborating with Electrical and Computer Engineering Professor Zbigniew Kalbarczyk at the University of Illinois at Urbana-Champaign, ADSC researchers Daisuke Mashima and Yue Wu, who developed security and safety guidelines for Singapore metro systems, and researchers with Critical Systems Labs Inc., Canada, and NSHC Pte. Ptd.

There are a lot of interesting engineering challenges because there are many types of information that needs to be analyzed, such as dynamic geofencing, UAS locations, or the weather, that can greatly affect an UAS,” Chen said. “Once you collect all that information from the sources, you have to determine if the information is correct or if an attacker might be affecting the data.”