David Gregg

Trinity College Dublin, Trinity College London

Papers

Total Citations

H-Index

About

David Gregg is a researcher whose work focuses on the optimization of deep neural networks (DNNs) for resource-constrained computing environments, with particular emphasis on enabling efficient inference on mobile and embedded devices. His research addresses one of the central challenges in modern AI deployment: making state-of-the-art convolutional neural networks (CNNs) practical for platforms with severely limited memory and energy budgets. Gregg's most notable contribution is TASO (Time and Space Optimization), published in 2020, which tackles the co-optimization of memory usage and execution time for DNN inference — a problem critical to applications ranging from industrial robotics to mobile biometric identification. This work has garnered 12 citations, reflecting its relevance to the embedded AI community. Complementing this, his 2019 poster introduced an ahead-of-time primitive selection framework leveraging Integer Linear Programming, demonstrating his interest in formal optimization techniques applied to neural network deployment. Together, these works position Gregg as a contributor to the growing field of efficient deep learning, bridging the gap between cutting-edge neural network performance and the practical constraints of real-world embedded systems. His research is particularly valuable for engineers and scientists working on deploying AI in latency-sensitive, resource-limited environments.