Work

Knowledge Distillation Applied to Optical Channel Equalization: Solving the Parallelization Problem of Recurrent Connection

Year: 2022

Type: preprint

Abstract: To circumvent the non-parallelizability of recurrent neural network-based equalizers, we propose knowledge distillation to recast the RNN into a parallelizable feedforward structure. The latter shows 38\% latency decrease, while impacting the Q-factor by only 0.5dB.

Source: arXiv (Cornell University)

DOI: 10.48550/arxiv.2212.04569

Authors: Srivallapanondh, Sasipim, Freire, Pedro J., Spinnler, Bernhard, Costa, Nelson, Napoli, Antonio, Turitsyn, Sergei K., Prilepsky, Jaroslaw E.

Institutions: Aston University, Infinera (United States)

Language: English

Topic: Optical Network Technologies

Subfield: Electrical and Electronic Engineering

Field: Engineering

Domain: Physical Sciences

Open Access status: green

Funder: Engineering and Physical Sciences Research Council

Awards: G453597767, Transforming networks - building an intelligent optical infrastructure (TRANSNET)

OpenAlex ID: W4311247804

Cites:

Cited by: 1

Related to: 10