Martin Herbordt

Papers

1

Total Citations

35

H-Index

1

About

Martin Herbordt is a leading researcher in high-performance and reconfigurable computing, with a primary focus on accelerating deep neural network (DNN) inference for real-time, latency-critical applications. His major contributions center on developing novel hardware architectures, particularly using FPGAs, to overcome the performance bottlenecks of DNNs in domains like autonomous driving and robotic control. His highly cited work, "LP-BNN: Ultra-low-Latency BNN Inference with Layer Parallelism" (2019, 35 citations), exemplifies his approach by introducing a binarized neural network (BNN) design that achieves ultra-low latency through aggressive layer parallelism, directly addressing the challenge of deploying DNNs where speed is paramount. Beyond this, Herbordt has made significant strides in computer architecture for scientific computing, including molecular dynamics simulations and sparse matrix operations. His research consistently bridges the gap between algorithmic innovation and practical hardware implementation, earning him recognition for pushing the boundaries of what is achievable with reconfigurable logic. For students and researchers, his work offers a compelling blueprint for designing efficient, real-time AI systems.

Research Focus

Key Achievements

1
H-Index
1
Papers
35
Total Citations
35
Avg Citations/Paper
🏆 Most Cited Paper
LP-BNN: Ultra-low-Latency BNN Inference with Layer Parallelism
35 citations · 2019
📈 Most Prolific Year: 2019 (1 Papers)
🤝 Key Collaborators: 6

Top Papers

  1. 1

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 12 days ago