Academic Profile : Faculty
Assoc Prof Lam Siew Kei
Associate Professor, College of Computing & Data Science
Assistant Chair (Admissions and Outreach), School of Computer Science and Engineering (SCSE)
Email
External Links
Siew Kei received his BASc, MEng and PhD from School of Computer Engineering, NTU. He was a Visiting Research Fellow in the Imperial College of London, University of Warwick, and RWTH Aachen, Germany. He is currently an Associate Professor in College of Computing & Data Science, NTU. His research investigates custom computing techniques to meet the challenging demands for performance, energy-efficiency, cost, reliability, and security in edge intelligence.
Custom Computing; High Level Synthesis; Heterogeneous and Reconfigurable Computing Systems; Embedded Vision and Autonomous Systems
- FL-SLAM: Federated Lifelong SLAM
- Protecting Critical AI Assets for Edge Intelligence
- ROAD USER TRAJECTORY PREDICTION USING ENHANCED SCENE CONTEXT
- Project SCSE-LSK – Energy-Efficient Architecture and Robust Perception Algorithms for Spatial Edge AI
US 2020/0151305 A1: Computer Program Code Obfuscation Methods And Systems (2022)
Abstract: Methods and systems for obfuscating computer program code are disclosed. In an embodiment, a method of generating obfuscated binary code from input source code for execution on a target processor comprises: generating a set of random obfuscation transform selections; and iteratively optimizing the obfuscation transform selections until a termination criterion is met. The obfuscation transformation selections may comprise indications of custom instructions which are executable on the co-processor in order to reduce side channel leakage.
Abstract: Methods and systems for obfuscating computer program code are disclosed. In an embodiment, a method of generating obfuscated binary code from input source code for execution on a target processor comprises: generating a set of random obfuscation transform selections; and iteratively optimizing the obfuscation transform selections until a termination criterion is met. The obfuscation transformation selections may comprise indications of custom instructions which are executable on the co-processor in order to reduce side channel leakage.