# Probing clustering in neural network representations

arXiv: https://arxiv.org/abs/2311.07864 · HTML: https://mlsanity.com/paper/2311.07864/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 46 of its 55 references that are in the map.
- First posted on arXiv 14 Nov 2023.

## Builds on (most cited first, 30 of 46)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://mlsanity.com/paper/1409.1556/) (2014)
- [Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift](https://mlsanity.com/paper/1502.03167/) (2015)
- [Layer Normalization](https://mlsanity.com/paper/1607.06450/) (2016)
- [Bootstrap your own latent: A new approach to self-supervised Learning](https://mlsanity.com/paper/2006.07733/) (2020)
- [Unsupervised Learning of Visual Features by Contrasting Cluster Assignments](https://mlsanity.com/paper/2006.09882/) (2020)
- [Similarity of Neural Network Representations Revisited](https://mlsanity.com/paper/1905.00414/) (2019)
- [Understanding Black-box Predictions via Influence Functions](https://mlsanity.com/paper/1703.04730/) (2017)
- [Understanding intermediate layers using linear classifier probes](https://mlsanity.com/paper/1610.01644/) (2016)
- [Deep Clustering for Unsupervised Learning of Visual Features](https://mlsanity.com/paper/1807.05520/) (2018)
- [Unsupervised Deep Embedding for Clustering Analysis](https://mlsanity.com/paper/1511.06335/) (2015)
- [Prevalence of Neural Collapse during the terminal phase of deep learning training](https://mlsanity.com/paper/2008.08186/) (2020)
- [Deep Learning on a Data Diet: Finding Important Examples Early in Training](https://mlsanity.com/paper/2107.07075/) (2021)
- [Beyond neural scaling laws: beating power law scaling via data pruning](https://mlsanity.com/paper/2206.14486/) (2022)
- [How Does Batch Normalization Help Optimization?](https://mlsanity.com/paper/1805.11604/) (2018)
- [SCAN: Learning to Classify Images without Labels](https://mlsanity.com/paper/2005.12320/) (2020)
- [GRAD-MATCH: Gradient Matching based Data Subset Selection for Efficient Deep Model Training](https://mlsanity.com/paper/2103.00123/) (2021)
- [Do Wide and Deep Networks Learn the Same Things? Uncovering How Neural Network Representations Vary with Width and Depth](https://mlsanity.com/paper/2010.15327/) (2020)
- [Understanding image representations by measuring their equivariance and equivalence](https://mlsanity.com/paper/1411.5908/) (2014)
- [On Completeness-aware Concept-Based Explanations in Deep Neural Networks](https://mlsanity.com/paper/1910.07969/) (2019)
- [Learning to Discover Novel Visual Categories via Deep Transfer Clustering](https://mlsanity.com/paper/1908.09884/) (2019)
- [Hidden Stratification Causes Clinically Meaningful Failures in Machine Learning for Medical Imaging](https://mlsanity.com/paper/1909.12475/) (2019)
- [No Subclass Left Behind: Fine-Grained Robustness in Coarse-Grained Classification Problems](https://mlsanity.com/paper/2011.12945/) (2020)
- [Revisiting Model Stitching to Compare Neural Representations](https://mlsanity.com/paper/2106.07682/) (2021)
- [Domino: Discovering Systematic Errors with Cross-Modal Embeddings](https://mlsanity.com/paper/2203.14960/) (2022)
- [BREEDS: Benchmarks for Subpopulation Shift](https://mlsanity.com/paper/2008.04859/) (2020)
- [Automatically Discovering and Learning New Visual Categories with Ranking Statistics](https://mlsanity.com/paper/2002.05714/) (2020)
- [Learning to cluster in order to transfer across domains and tasks](https://mlsanity.com/paper/1711.10125/) (2017)
- [Neighborhood Contrastive Learning for Novel Class Discovery](https://mlsanity.com/paper/2106.10731/) (2021)
- [Generalized Shape Metrics on Neural Representations](https://mlsanity.com/paper/2110.14739/) (2021)

## Abstract (arXiv)

Neural network representations contain structure beyond what was present in the training labels. For instance, representations of images that are visually or semantically similar tend to lie closer to each other than to dissimilar images, regardless of their labels. Clustering these representations can thus provide insights into dataset properties as well as the network internals. In this work, we study how the many design choices involved in neural network training affect the clusters formed in the hidden representations. To do so, we establish an evaluation setup based on the BREEDS hierarchy, for the task of subclass clustering after training models with only superclass information. We isolate the training dataset and architecture as important factors affecting clusterability. Datasets with labeled classes consisting of unrelated subclasses yield much better clusterability than those following a natural hierarchy. When using pretrained models to cluster representations on downstream datasets, models pretrained on subclass labels provide better clusterability than models pretrained on superclass labels, but only when there is a high degree of domain overlap between the pretraining and downstream data. Architecturally, we find that normalization strategies affect which layers yield the best clustering performance, and, surprisingly, Vision Transformers attain lower subclass clusterability than ResNets.