We are excited to announce a notable publication in the Journal of Lightwave Technology:
Pedro Freire, Sasipim Srivallapanondh, Bernhard Spinnler, Antonio Napoli, Nelson Costa, Jaroslaw E. Prilepsky, and Sergei K. Turitsyn,
“Computational Complexity Optimization of Neural Network-Based Equalizers in Digital Signal Processing: A Comprehensive Approach,” J. Lightwave Technol., vol. 42, no. 12, pp. 4177–4201 (2024).
🔗 Access the paper: https://doi.org/10.1109/JLT.2024.3386886
Highlights:
- Introduces a comprehensive methodology to estimate and optimize the complexity of neural network equalizers, encompassing training, inference, and hardware synthesis.
- Provides new complexity metrics that connect neural network architectural choices directly to hardware deployment and performance trade-offs.
- Validates the approach for both feedforward and recurrent models, showcasing substantial complexity reductions while maintaining high equalization performance.
This contribution lays essential groundwork for deploying high-performance, low-complexity neural equalizers in real-time optical systems.