A.J. Redfern

About

Dr. A.J. Redfern is a leading figure in the design of high-performance VLSI architectures for deep learning inference, with a primary focus on enabling real-time, energy-efficient convolutional neural network (CNN) processing. Their seminal 2017 work, "CNN inference: VLSI architecture for convolution layer for 1.2 TOPS," introduced a groundbreaking hardware accelerator capable of achieving 1.2 trillion operations per second, a critical milestone for deploying CNNs in resource-constrained edge devices such as autonomous vehicles, medical imaging systems, and robotics. This architecture addressed the computational bottleneck of convolution layers—the core of modern CNNs—by optimizing dataflow and memory access patterns, directly influencing subsequent designs in the field. With over 6 citations, this paper has become a foundational reference for researchers and engineers developing custom silicon for AI workloads. Dr. Redfern’s contributions bridge the gap between algorithmic advances in deep learning and practical hardware implementation, making them a key innovator in the push toward ubiquitous, low-latency AI. Their work continues to shape the trajectory of embedded machine learning and hardware-software co-design.

Research Focus

Key Achievements

Top Papers

1. 1
   CNN inference: VLSI architecture for convolution layer for 1.2 TOPS

   6 citations · 2017

Key Collaborators

Contact & Links