Tarek Abdelzaher
· Professor of Computer ScienceUniversity of Illinois Urbana-Champaign · Computer Science
Active 1997–2024
About
Tarek Abdelzaher received his Ph.D. in Computer Science from the University of Michigan in 1999. He is currently a Professor and Willett Faculty Scholar at the Department of Computer Science at the University of Illinois at Urbana-Champaign. His research interests broadly encompass understanding and influencing the performance and temporal properties of networked embedded, social, and software systems, especially in the context of increasing complexity, distribution, and interaction with physical environments. Abdelzaher has authored or coauthored more than 400 refereed publications in areas including real-time computing, cyber-physical systems (CPS), Internet of Things (IoT), distributed systems, intelligent networked sensing, machine learning, and control. He has served as Editor-in-Chief of the Journal of Real-Time Systems for 15 years and as an associate editor for several prominent journals in his field. Additionally, he has chaired numerous conferences such as RTAS, RTSS, IPSN, Sensys, DCoSS, ICDCS, Infocom, and ICAC. Recognized for his contributions, Abdelzaher has received the IEEE Outstanding Technical Achievement and Leadership Award in Real-time Systems, the Xerox Award for Faculty Research, and several best paper awards. He is a fellow of both IEEE and ACM.
Research topics
- Artificial Intelligence
- Computer Science
- Machine Learning
- Distributed computing
- Computer network
- Computer Security
- Data Mining
- Data science
- Software engineering
- Medicine
- Telecommunications
- Seismology
- Programming language
- Computer vision
- Theoretical computer science
- Mathematics
- Mathematical optimization
- Geology
- Embedded system
- Engineering
Selected publications
Real-Time Task Scheduling for Machine Perception in In Intelligent Cyber-Physical Systems
IEEE Transactions on Computers · 2021 · 29 citations
Senior authorCorresponding- Computer Science
- Artificial Intelligence
- Computer Science
This paper explores <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">criticality-based real-time scheduling</i> of neural-network-based machine inference pipelines in cyber-physical systems (CPS) to mitigate the effect of algorithmic priority inversion. We specifically focus on the perception subsystem, an important subsystem feeding other components (e.g., planning and control). In general, priority inversion occurs in real-time systems when computations that are of lower priority are performed together with or ahead of those that are of higher priority. In current machine perception software, significant priority inversion occurs because <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">resource allocation</i> to the underlying neural network models does not differentiate between critical and less critical data within a scene. To remedy this problem, in recent work, we proposed an architecture to partition the input data into regions of different criticality, then formulated a utility-based optimization problem to batch and schedule their processing in a manner that maximizes confidence in perception results, subject to criticality-based time constraints. This journal extension matures the work in several directions: (i) We extend confidence maximization to a generalized utility optimization formulation that accounts for criticality in the utility function itself, offering finer-grained control over resource allocation within the perception pipeline; (ii) we further instantiate and compare two different criticality metrics (distance-based and relative velocity-based) to understand their relative advantages; and (iii) we explore the limitations of the approach, specifically how inaccuracies in criticality-based attention cueing affect performance. All experiments are conducted on the NVIDIA Jetson AGX Xavier platform with a real-world driving dataset.
Machine Learning/Artificial Intelligence for Sensor Data Fusion–Opportunities and Challenges
IEEE Aerospace and Electronic Systems Magazine · 2021 · 181 citations
Senior authorCorresponding- Artificial Intelligence
- Computer Science
- Machine Learning
During Fusion 2019 Conference (https://www.fusion2019.org/program.html), leading experts presented ideas on the historical, contemporary, and future coordination of artificial intelligence/machine learning (AI/ML) with sensor data fusion (SDF). While AI/ML and SDF concepts have had a rich history since the early 1900s—emerging from philosophy and psychology—it was not until the dawn of computers that both AI/ML and SDF researchers initiated discussions on how mathematical techniques could be implemented for real-time analysis. ML, and in particular deep learning, has demonstrated tremendous success in computer vision, natural language understanding, and data analytics. As a result, ML has been proposed as the solution to many problems that inherently include multi-modal data. For example, success in autonomous vehicles has validated the promise of ML with SDF, but additional research is needed to explain, understand, and coordinate heterogeneous data analytics for situation awareness. The panel identified opportunities for merging AI/ML and SDF such as computational efficiency, improved decision making, expanding knowledge, and providing security; while highlighting challenges for multi-domain operations, human-machine teaming, and ethical deployment strategies.
New Frontiers in IoT: Networking, Systems, Reliability, and Security Challenges
IEEE Internet of Things Journal · 2020 · 64 citations
- Computer Science
- Computer Science
- Computer Security
The field of IoT has blossomed and is positively influencing many application domains. In this article, we bring out the unique challenges this field poses to research in computer systems and networking. The unique challenges arise from the unique characteristics of IoT systems such as the diversity of application domains where they are used and the increasingly demanding protocols they are being called upon to run (such as video and LIDAR processing) on constrained resources (on-node and network). We show how these open challenges can benefit from foundations laid in other areas, such as fifth-generation network cellular protocols, machine learning model reduction, and device-edge-cloud offloading. We then discuss the unique challenges for reliability, security, and privacy posed by IoT systems due to their salient characteristics which include heterogeneity of devices and protocols, dependence on the physical environment, and the close coupling with humans. We again show how open research challenges benefit from the reliability, security, and privacy advancements in other areas. We conclude by providing a vision for a desirable end state for IoT systems.
On Removing Algorithmic Priority Inversion from Mission-critical Machine Inference Pipelines
2020 · 52 citations
Senior authorCorresponding- Computer Science
- Computer Science
- Artificial Intelligence
The paper discusses algorithmic priority inversion in mission-critical machine inference pipelines used in modern neural-network-based cyber-physical applications, and develops a scheduling solution to mitigate its effect. In general, priority inversion occurs in real-time systems when computations that are of lower priority are performed together with or ahead of those that are of higher priority. <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup> In current machine intelligence software, significant priority inversion occurs on the path from perception to decision-making, where the execution of underlying neural network algorithms does not differentiate between critical and less critical data. We describe a scheduling framework to resolve this problem, and demonstrate that it improves the system’s ability to react to critical inputs, while at the same time reducing platform cost.
2020 · 135 citations
Senior authorCorresponding- Computer Science
- Computer Science
- Artificial Intelligence
With recent advances, neural networks have become a crucial building block in intelligent IoT systems and sensing applications. However, the excessive computational demand remains a serious impediment to their deployments on low-end IoT devices. With the emergence of edge computing, offloading grows into a promising technique to circumvent end-device limitations. However, transferring data between local and edge devices takes up a large proportion of time in existing offloading frameworks, creating a bottleneck for low-latency intelligent services. In this work, we propose a general framework, called deep compressive offloading. By integrating compressive sensing theory and deep learning, our framework can encode data for offloading into tiny sizes with negligible overhead on local devices and decode the data on the edge server, while offering theoretical guarantees on perfect reconstruction and lossless inference. By trading edge computing resources for data transmission time, our design can significantly reduce offloading latency with almost no accuracy loss. We build a deep compressive offloading system to serve state-of-the-art computer vision and speech recognition services. With comprehensive evaluations, our system can consistently reduce end-to-end latency by 2X to 4X with 1% accuracy loss, compared to state-of-the-art neural network offloading systems. In conditions of limited network bandwidth or intensive background traffic, our system can further speed up the neural network inference by up to 35X 1.
Revisiting Over-smoothing in Deep GCNs
arXiv (Cornell University) · 2020 · 55 citations
Senior authorCorresponding- Artificial Intelligence
- Computer Science
- Medicine
Oversmoothing has been assumed to be the major cause of performance drop in deep graph convolutional networks (GCNs). In this paper, we propose a new view that deep GCNs can actually learn to anti-oversmooth during training. This work interprets a standard GCN architecture as layerwise integration of a Multi-layer Perceptron (MLP) and graph regularization. We analyze and conclude that before training, the final representation of a deep GCN does over-smooth, however, it learns anti-oversmoothing during training. Based on the conclusion, the paper further designs a cheap but effective trick to improve GCN training. We verify our conclusions and evaluate the trick on three citation networks and further provide insights on neighborhood aggregation in GCNs.
Recent grants
CSR-EHS: Towards Ubiquitous Wearable Embedded Computing
NSF · $150k · 2006–2009
FIA: Collaborative Research: Named Data Networking (NDN)
NSF · $600k · 2010–2016
Collaborative Research: NeTS-NOSS: The Sensor Network Development and Deployment Studio
NSF · $324k · 2006–2010
NSF · $100k · 2007–2011
NSF · $900k · 2010–2014
Frequent coauthors
- 192 shared
Bhaskar Krishnamachari
- 129 shared
Jie Gao
- 129 shared
Viktor K. Prasanna
- 128 shared
Luca Mottola
Politecnico di Milano
- 128 shared
Sotiris Nikoletseas
- 128 shared
Magnús M. Halldórsson
Reykjavík University
- 128 shared
Shengzhong Liu
Dalian National Laboratory for Clean Energy
- 99 shared
Dongxin Liu
Chinese Center For Disease Control and Prevention
Labs
Siebel School of Computing and Data SciencePI
Education
- 1998
Ph.D., Computer Science
University of Illinois at Urbana-Champaign
- 1994
M.S., Computer Science
University of Illinois at Urbana-Champaign
- 1991
B.S., Computer Science
University of Science and Technology of China
Awards & honors
- IEEE Outstanding Technical Achievement and Leadership Award…
- Xerox Award for Faculty Research (2011)
- Fellow of IEEE
- Fellow of ACM
- Best Paper Award at IEEE RTAS (2010)
Similar researchers at University of Illinois Urbana-Champaign
- Resume-aware match score
- Save to shortlist
- AI-drafted outreach
See your match with Tarek Abdelzaher
PhdFit ranks faculty by your research interests, methods, and publications — grounded in their actual work, not templates.
- Free to start
- No credit card
- 30-second signup