ASIC Implementation of Time-Domain Digital Backpropagation with Deep-Learned Chromatic Dispersion Filters

Fougstedt, Christoffer; Häger, Christian; Svensson, Lars; Pfister, Henry D.; Larsson-Edefors, Per

doi:10.1109/ecoc.2018.8535430

Cited by 17 publications

(23 citation statements)

References 19 publications

(38 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The obtained FIR filters are as short as 5 or 3 (symmetric) taps per step, leading to very simple and efficient hardware implementation. This is confirmed by recent ASIC synthesis results which show that the power consumption of LDBP becomes comparable to linear equalization [24]. LDBP can also be extended to subband processing to enable low-complexity DBP for multi-channel or other wideband transmission scenarios [25].…”

Section: Optimization Resultssupporting

confidence: 64%

What Can Machine Learning Teach Us about Communications?

Lian

Häger

Pfister

2018

2018 IEEE Information Theory Workshop (ITW)

Self Cite

View full text Add to dashboard Cite

Rapid improvements in machine learning over the past decade are beginning to have far-reaching effects. For communications, engineers with limited domain expertise can now use off-the-shelf learning packages to design high-performance systems based on simulations. Prior to the current revolution in machine learning, the majority of communication engineers were quite aware that system parameters (such as filter coefficients) could be learned using stochastic gradient descent. It was not at all clear, however, that more complicated parts of the system architecture could be learned as well.In this paper, we discuss the application of machine-learning techniques to two communications problems and focus on what can be learned from the resulting systems. We were pleasantly surprised that the observed gains in one example have a simple explanation that only became clear in hindsight. In essence, deep learning discovered a simple and effective strategy that had not been considered earlier.

show abstract

Section: Optimization Resultssupporting

confidence: 64%

What Can Machine Learning Teach Us about Communications?

Lian

Häger

Pfister

2018

2018 IEEE Information Theory Workshop (ITW)

Self Cite

View full text Add to dashboard Cite

show abstract

“…The filters are then progressively pruned to a given target length by forcing the outermost taps to zero at certain iterations during SGD [26]. The zero forcing is done using a masking operation in TensorFlow.…”

Section: B Pre-training and Filter Pruningmentioning

confidence: 99%

“…In terms of complexity, it has been shown that the power consumption and chip area for time-domain DBP [48] and LDBP [26] are dominated by the linear steps, whereas the nonlinear steps have efficient hardware implementations using a Taylor expansion. Therefore, we focus on the linear steps for simplicity.…”

Section: Testingmentioning

confidence: 99%

“…Fixed-point requirements and other DSP hardware implementation aspects for DBP have been investigated in [22], [26], [48], [49], [62]. A potential benefit of multi-step architectures is that they empirically tend to have many "good" parameter configurations that lie relatively close to each other.…”

Section: Coefficient Quantization and Asic Implementationmentioning

confidence: 99%

“…All FIR filters can then be jointly optimized, similar to optimizing the weight matrices in a deep neural network (NN) [23], [25]. Complexity is reduced via pruning, i.e., progressively shortening the filters during the optimization procedure [26]. This can be seen as a form of model compression, which is commonly used in machine learning to reduce the size of NNs [27], [28].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Revisiting Efficient Multi-Step Nonlinearity Compensation With Machine Learning: An Experimental Demonstration

Oliari

Goossens

Häger

et al. 2020

J. Lightwave Technol.

Self Cite

View full text Add to dashboard Cite

Efficient nonlinearity compensation in fiber-optic communication systems is considered a key element to go beyond the "capacity crunch". One guiding principle for previous work on the design of practical nonlinearity compensation schemes is that fewer steps lead to better systems. In this paper, we challenge this assumption and show how to carefully design multi-step approaches that provide better performance-complexity tradeoffs than their few-step counterparts. We consider the recently proposed learned digital backpropagation (LDBP) approach, where the linear steps in the split-step method are re-interpreted as general linear functions, similar to the weight matrices in a deep neural network. Our main contribution lies in an experimental demonstration of this approach for a 25 Gbaud singlechannel optical transmission system. It is shown how LDBP can be integrated into a coherent receiver DSP chain and successfully trained in the presence of various hardware impairments. Our results show that LDBP with limited complexity can achieve better performance than standard DBP by using very short, but jointly optimized, finite-impulse response filters in each step. This paper also provides an overview of recently proposed extensions of LDBP and we comment on potentially interesting avenues for future work.

show abstract