# Do graph neural networks learn traditional jet substructure?

arXiv: https://arxiv.org/abs/2211.09912 · HTML: https://mlsanity.com/paper/2211.09912/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Builds on 16 of its 47 references that are in the map.
- First posted on arXiv 17 Nov 2022.

## Builds on (most cited first, 16 of 16)

- [Fast Graph Representation Learning with PyTorch Geometric](https://mlsanity.com/paper/1903.02428/) (2019)
- [Dynamic Graph CNN for Learning on Point Clouds](https://mlsanity.com/paper/1801.07829/) (2018)
- [Sanity Checks for Saliency Maps](https://mlsanity.com/paper/1810.03292/) (2018)
- [Higher-Order Explanations of Graph Neural Networks via Relevant Walks](https://mlsanity.com/paper/2006.03589/) (2020)
- [Fast inference of deep neural networks in FPGAs for particle physics](https://mlsanity.com/paper/1804.06913/) (2018)
- [ParticleNet: Jet Tagging via Particle Clouds](https://mlsanity.com/paper/1902.08570/) (2019)
- [Particle Cloud Generation with Message Passing Generative Adversarial Networks](https://mlsanity.com/paper/2106.11535/) (2021)
- [Learning representations of irregular particle-detector geometry with distance-weighted graph networks](https://mlsanity.com/paper/1902.07987/) (2019)
- [MLPF: Efficient machine-learned particle-flow reconstruction using graph neural networks](https://mlsanity.com/paper/2101.08578/) (2021)
- [Applications and Techniques for Fast Machine Learning in Science](https://mlsanity.com/paper/2110.13041/) (2021)
- [Object condensation: one-stage grid-free multi-object reconstruction in physics detectors, graph and image data](https://mlsanity.com/paper/2002.03605/) (2020)
- [Attribution-based Explanations that Provide Recourse Cannot be Robust](https://mlsanity.com/paper/2205.15834/) (2022)
- [Explainable AI for ML jet taggers using expert variables and layerwise relevance propagation](https://mlsanity.com/paper/2011.13466/) (2020)
- [Machine Learning for Particle Flow Reconstruction at CMS](https://mlsanity.com/paper/2203.00330/) (2022)
- [Explaining machine-learned particle-flow reconstruction](https://mlsanity.com/paper/2111.12840/) (2021)
- [Explainable AI for High Energy Physics](https://mlsanity.com/paper/2206.06632/) (2022)

## Built on by (most cited first, 4 of 4)

- [A Comparison Between Invariant and Equivariant Classical and Quantum Graph Neural Networks](https://mlsanity.com/paper/2311.18672/) (2023)
- [Jet Discrimination with Quantum Complete Graph Neural Network](https://mlsanity.com/paper/2403.04990/) (2024)
- [Automated Molecular Concept Generation and Labeling with Large Language Models](https://mlsanity.com/paper/2406.09612/) (2024)
- [Dissecting Jet-Tagger Through Mechanistic Interpretability](https://mlsanity.com/paper/2605.09881/) (2026)

## Abstract (arXiv)

At the CERN LHC, the task of jet tagging, whose goal is to infer the origin of a jet given a set of final-state particles, is dominated by machine learning methods. Graph neural networks have been used to address this task by treating jets as point clouds with underlying, learnable, edge connections between the particles inside. We explore the decision-making process for one such state-of-the-art network, ParticleNet, by looking for relevant edge connections identified using the layerwise-relevance propagation technique. As the model is trained, we observe changes in the distribution of relevant edges connecting different intermediate clusters of particles, known as subjets. The resulting distribution of subjet connections is different for signal jets originating from top quarks, whose subjets typically correspond to its three decay products, and background jets originating from lighter quarks and gluons. This behavior indicates that the model is using traditional jet substructure observables, such as the number of prongs -- energetic particle clusters -- within a jet, when identifying jets.