# Ranking Deep Learning Generalization using Label Variation in Latent Geometry Graphs

arXiv: https://arxiv.org/abs/2011.12737 · HTML: https://mlsanity.com/paper/2011.12737/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 25 Nov 2020.

## Built on by (most cited first, 6 of 6)

- [Measuring Generalization with Optimal Transport](https://mlsanity.com/paper/2106.03314/) (2021)
- [Predicting the generalization gap in neural networks using topological data analysis](https://mlsanity.com/paper/2203.12330/) (2022)
- [On Predicting Generalization using GANs](https://mlsanity.com/paper/2111.14212/) (2021)
- [Weight Expansion: A New Perspective on Dropout and Generalization](https://mlsanity.com/paper/2201.09209/) (2022)
- [Slicing Mutual Information Generalization Bounds for Neural Networks](https://mlsanity.com/paper/2406.04047/) (2024)
- [On the Effects of Artificial Data Modification](https://mlsanity.com/paper/2110.13968/) (2021)

## Abstract (arXiv)

Measuring the generalization performance of a Deep Neural Network (DNN) without relying on a validation set is a difficult task. In this work, we propose exploiting Latent Geometry Graphs (LGGs) to represent the latent spaces of trained DNN architectures. Such graphs are obtained by connecting samples that yield similar latent representations at a given layer of the considered DNN. We then obtain a generalization score by looking at how strongly connected are samples of distinct classes in LGGs. This score allowed us to rank 3rd on the NeurIPS 2020 Predicting Generalization in Deep Learning (PGDL) competition.