# Towards Shape Biased Unsupervised Representation Learning for Domain Generalization

arXiv: https://arxiv.org/abs/1909.08245 · HTML: https://mlsanity.com/paper/1909.08245/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 18 Sep 2019.

## Built on by (most cited first, 10 of 10)

- [Model-Based Domain Generalization](https://mlsanity.com/paper/2102.11436/) (2021)
- [Robust Representation Learning via Perceptual Similarity Metrics](https://mlsanity.com/paper/2106.06620/) (2021)
- [Causal Transportability for Visual Recognition](https://mlsanity.com/paper/2204.12363/) (2022)
- [Combining Diverse Feature Priors](https://mlsanity.com/paper/2110.08220/) (2021)
- [Discriminative Domain-Invariant Adversarial Network for Deep Domain Generalization](https://mlsanity.com/paper/2108.08995/) (2021)
- [The Connection between Out-of-Distribution Generalization and Privacy of ML Models](https://mlsanity.com/paper/2110.03369/) (2021)
- [Explainability-aided Domain Generalization for Image Classification](https://mlsanity.com/paper/2104.01742/) (2021)
- [A Flat Minima Perspective on Understanding Augmentations and Model Robustness](https://mlsanity.com/paper/2505.24592/) (2025)
- [Counterbalancing Teacher: Regularizing Batch Normalized Models for Robustness](https://mlsanity.com/paper/2207.01548/) (2022)
- [Resilience of Vision Transformers for Domain Generalisation in the Presence of Out-of-Distribution Noisy Images](https://mlsanity.com/paper/2504.04225/) (2025)

## Abstract (arXiv)

It is known that, without awareness of the process, our brain appears to focus on the general shape of objects rather than superficial statistics of context. On the other hand, learning autonomously allows discovering invariant regularities which help generalization. In this work, we propose a learning framework to improve the shape bias property of self-supervised methods. Our method learns semantic and shape biased representations by integrating domain diversification and jigsaw puzzles. The first module enables the model to create a dynamic environment across arbitrary domains and provides a domain exploration vs. exploitation trade-off, while the second module allows the model to explore this environment autonomously. This universal framework does not require prior knowledge of the domain of interest. Extensive experiments are conducted on several domain generalization datasets, namely, PACS, Office-Home, VLCS, and Digits. We show that our framework outperforms state-of-the-art domain generalization methods by a large margin.