Yuzhang Shang

Papers

Total Citations

H-Index

About

Yuzhang Shang is a rising researcher at the forefront of efficient deep learning, with a primary focus on model compression and binary neural networks for 3D vision. His most notable contribution is the development of the Fully Binary Point Transformer (FBPT), a groundbreaking model that compresses 32-bit full-precision networks into 1-bit binary values for both weights and activations. This work, published in 2024, addresses the critical challenge of deploying complex point cloud processing on resource-constrained devices like robotics and mobile platforms. While still early in its citation trajectory with 2 citations, FBPT represents a significant step toward making transformer architectures practical for real-time, low-power 3D perception tasks. Shang’s research bridges the gap between state-of-the-art accuracy and extreme computational efficiency, positioning him as an emerging voice in the push for sustainable AI. His work holds particular promise for autonomous systems that require both high performance and minimal energy consumption, marking him as a researcher to watch in the evolving landscape of efficient point cloud processing.