# Neural Enhancement of Factor Graph-based Symbol Detection

arXiv: https://arxiv.org/abs/2203.03333 · HTML: https://mlsanity.com/paper/2203.03333/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 5 of its 19 references that are in the map.
- First posted on arXiv 7 Mar 2022.

## Builds on (most cited first, 5 of 5)

- [Model-Based Deep Learning](https://mlsanity.com/paper/2012.08405/) (2020)
- [Learning to Decode Linear Codes Using Deep Learning](https://mlsanity.com/paper/1607.04793/) (2016)
- [Neural Enhanced Belief Propagation on Factor Graphs](https://mlsanity.com/paper/2003.01998/) (2020)
- [Learned Factor Graphs for Inference from Stationary Time Sequences](https://mlsanity.com/paper/2006.03258/) (2020)
- [Achievable Information Rates for Fiber Optics: Applications and Computations](https://mlsanity.com/paper/1709.10393/) (2017)

## Built on by (most cited first, 3 of 3)

- [Low-complexity Near-optimum Symbol Detection Based on Neural Enhancement of Factor Graphs](https://mlsanity.com/paper/2203.16417/) (2022)
- [Blind Channel Estimation and Joint Symbol Detection with Data-Driven Factor Graphs](https://mlsanity.com/paper/2401.12627/) (2024)
- [Structural Optimization of Factor Graphs for Symbol Detection via Continuous Clustering and Machine Learning](https://mlsanity.com/paper/2211.11406/) (2022)

## Abstract (arXiv)

We study the application of the factor graph framework for symbol detection on linear inter-symbol interference channels. Cyclic factor graphs have the potential to yield low-complexity symbol detectors, but are suboptimal if the ubiquitous sum-product algorithm is applied. In this paper, we present and evaluate strategies to improve the performance of cyclic factor graph-based symbol detection algorithms by means of neural enhancement. In particular, we apply neural belief propagation as an effective way to counteract the effect of cycles within the factor graph. We further propose the application and optimization of a linear preprocessor of the channel output. By modifying the observation model, the preprocessing can effectively change the underlying factor graph, thereby significantly improving the detection performance as well as reducing the complexity.