Securing and Reliable Low-Power Computing at the Edge

Title of the Talk: Securing and Reliable Low-Power Computing at the Edge: Exploring the Fitness Relationship through Approximate Computing
Speaker: Urbi Chatterjeer
Host Faculty: Dr.Shirshendu Das
Date: Friday, 13th February 2026
Time: 3:00 pm
Venue: CS-Seminar Hall 1

Abstract One of the primary challenges in edge computing hardware design is to achieve an optimal fit between the fundamental bottlenecks of resource-constrained platforms and the delivered circuit functionality realized through low power computing techniques. This alignment is essential for energy-harvesting IoT nodes, edge AI inference, and sensor-driven applications, where empirical evidence from VLSI and approximate computing literature demonstrates that well-aligned designs significantly enhance energy efficiency, inference accuracy, and system trustworthiness, while misalignment frequently results in excessive power dissipation, undetected faults, or exploitable vulnerabilities in low power arithmetic units. This talk examines these challenges by first elucidating the static understanding of design trade-offs through contributions by SETTLOR IIT Kanpur, including versatile approximate data formats (ACM TECS 2023), hardware for multimedia acceleration (DATE 2023), soft-error-resilient floating-point representations (ICCAD 2025), and SAT-driven Trojan countermeasures for low power approximate circuits (ACM TECS 2025). Later, the talk focuses on the need of hardware aware optimizations (VLSID 2025) to adapt the fitness relationship across changing edge workloads.

Bio Urbi Chatterjee is an Assistant Professor in the Department of Computer Science and Engineering at the Indian Institute of Technology Kanpur, where she has been a faculty member since March 2021. She earned her Ph.D. from the Indian Institute of Technology Kharagpur, with prior affiliations including the SEAL lab at IIT Kharagpur and an M.Tech. in CSE from the Indian School of Mines IIT Dhanbad. Her research expertise lies in hardware security, with a focus on Physically Unclonable Functions (PUFs), side-channel analysis, secure authentication protocols, embedded systems security, network on chip, low power computing, approximate computing and related areas in secure hardware design. She leads the Secure Embedded and Smart Things Laboratory (SETTLOR) at IIT Kanpur.