# Generalizing similarity in noisy setups: the DIBS phenomenon

arXiv: https://arxiv.org/abs/2201.12803 · HTML: https://mlsanity.com/paper/2201.12803/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 39 of its 54 references that are in the map.
- First posted on arXiv 30 Jan 2022.

## Builds on (most cited first, 30 of 39)

- [A Simple Framework for Contrastive Learning of Visual Representations](https://mlsanity.com/paper/2002.05709/) (2020)
- [Representation Learning with Contrastive Predictive Coding](https://mlsanity.com/paper/1807.03748/) (2018)
- [Momentum Contrast for Unsupervised Visual Representation Learning](https://mlsanity.com/paper/1911.05722/) (2019)
- [Bootstrap your own latent: A new approach to self-supervised Learning](https://mlsanity.com/paper/2006.07733/) (2020)
- [Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms](https://mlsanity.com/paper/1708.07747/) (2017)
- [Going Deeper with Convolutions](https://mlsanity.com/paper/1409.4842/) (2014)
- [Supervised Contrastive Learning](https://mlsanity.com/paper/2004.11362/) (2020)
- [Contrastive Multiview Coding](https://mlsanity.com/paper/1906.05849/) (2019)
- [Reconciling modern machine learning practice and the bias-variance trade-off](https://mlsanity.com/paper/1812.11118/) (2018)
- [Learning Representations by Maximizing Mutual Information Across Views](https://mlsanity.com/paper/1906.00910/) (2019)
- [Self-supervised Learning: Generative or Contrastive](https://mlsanity.com/paper/2006.08218/) (2020)
- [Deep Double Descent: Where Bigger Models and More Data Hurt](https://mlsanity.com/paper/1912.02292/) (2019)
- [Learning from Noisy Labels with Deep Neural Networks: A Survey](https://mlsanity.com/paper/2007.08199/) (2020)
- [A Survey on Contrastive Self-supervised Learning](https://mlsanity.com/paper/2011.00362/) (2020)
- [The generalization error of random features regression: Precise asymptotics and double descent curve](https://mlsanity.com/paper/1908.05355/) (2019)
- [Hard Negative Mixing for Contrastive Learning](https://mlsanity.com/paper/2010.01028/) (2020)
- [Understanding the Behaviour of Contrastive Loss](https://mlsanity.com/paper/2012.09740/) (2020)
- [Debiased Contrastive Learning](https://mlsanity.com/paper/2007.00224/) (2020)
- [Keystroke Dynamics as Part of Lifelogging](https://mlsanity.com/paper/2010.10144/) (2020)
- [Unsupervised Label Noise Modeling and Loss Correction](https://mlsanity.com/paper/1904.11238/) (2019)
- [Supervised Contrastive Learning for Pre-trained Language Model Fine-tuning](https://mlsanity.com/paper/2011.01403/) (2020)
- [EMNIST: an extension of MNIST to handwritten letters](https://mlsanity.com/paper/1702.05373/) (2017)
- [Learning to Learn from Noisy Labeled Data](https://mlsanity.com/paper/1812.05214/) (2018)
- [Demystifying Contrastive Self-Supervised Learning: Invariances, Augmentations and Dataset Biases](https://mlsanity.com/paper/2007.13916/) (2020)
- [Deep Self-Learning From Noisy Labels](https://mlsanity.com/paper/1908.02160/) (2019)
- [Boosting Contrastive Self-Supervised Learning with False Negative Cancellation](https://mlsanity.com/paper/2011.11765/) (2020)
- [Can contrastive learning avoid shortcut solutions?](https://mlsanity.com/paper/2106.11230/) (2021)
- [Optimal Regularization Can Mitigate Double Descent](https://mlsanity.com/paper/2003.01897/) (2020)
- [Double Trouble in Double Descent : Bias and Variance(s) in the Lazy Regime](https://mlsanity.com/paper/2003.01054/) (2020)
- [Robust Audio-Visual Instance Discrimination](https://mlsanity.com/paper/2103.15916/) (2021)

## Abstract (arXiv)

This work uncovers an interplay among data density, noise, and the generalization ability in similarity learning. We consider Siamese Neural Networks (SNNs), which are the basic form of contrastive learning, and explore two types of noise that can impact SNNs, Pair Label Noise (PLN) and Single Label Noise (SLN). Our investigation reveals that SNNs exhibit double descent behaviour regardless of the training setup and that it is further exacerbated by noise. We demonstrate that the density of data pairs is crucial for generalization. When SNNs are trained on sparse datasets with the same amount of PLN or SLN, they exhibit comparable generalization properties. However, when using dense datasets, PLN cases generalize worse than SLN ones in the overparametrized region, leading to a phenomenon we call Density-Induced Break of Similarity (DIBS). In this regime, PLN similarity violation becomes macroscopical, corrupting the dataset to the point where complete interpolation cannot be achieved, regardless of the number of model parameters. Our analysis also delves into the correspondence between online optimization and offline generalization in similarity learning. The results show that this equivalence fails in the presence of label noise in all the scenarios considered.