Pengbo Yu

Pengbo Yu is a researcher at the Embedded Systems Laboratory (ESL) within the School of Engineering at École polytechnique fédérale de Lausanne (EPFL). His work focuses on developing energy-efficient hardware architectures for edge artificial intelligence applications, with particular emphasis on variable-precision computing techniques.

His research interests span computer architecture, edge AI hardware acceleration, neural network quantization, memory systems, and low-power computing. Dr. Yu's work bridges the gap between algorithmic robustness in quantized neural networks and specialized hardware implementations that can dynamically adjust precision to maximize efficiency in resource-constrained environments.

His recent publications demonstrate a strong focus on the SoftSIMD paradigm, Dynamic Bitwidth-Frequency Scaling techniques, and near-memory computing solutions that achieve significant energy savings (up to 67%) compared to state-of-the-art approaches. His work shows consistent output in top venues including IEEE Transactions on VLSI Systems and ACM Transactions on Embedded Computing Systems.

Dr. Yu's research is supported by multiple funding agencies including EC H2020, H2020, and SNSF. He maintains active collaborations with researchers across Europe, including institutions like IMEC and National Technical University of Athens.

His laboratory work centers around the Embedded Systems Laboratory at EPFL, where he develops and validates novel hardware architectures for energy-efficient AI acceleration. Current projects include silicon validation of his proposed architectures on open-source RISC-V platforms and integration of variable-precision computing techniques into memory systems.