Leveraging Hidden Structure in Self-Supervised Learning
arXiv:2106.16060
Abstract
This work considers the problem of learning structured representations from raw images using self-supervised learning. We propose a principled framework based on a mutual information objective, which integrates self-supervised and structure learning. Furthermore, we devise a post-hoc procedure to interpret the meaning of the learnt representations. Preliminary experiments on CIFAR-10 show that the proposed framework achieves higher generalization performance in downstream classification tasks and provides more interpretable representations compared to the ones learnt through traditional self-supervised learning.
References in corpus (13)
- Auto-Encoding Variational Bayes
- Categorical Reparameterization with Gumbel-Softmax
- Learning deep representations by mutual information estimation and maximization
- The Concrete Distribution: A Continuous Relaxation of Discrete Random Variables
- Data-Efficient Image Recognition with Contrastive Predictive Coding
- Variational Graph Auto-Encoders
- Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representations
- Learning Representations by Maximizing Mutual Information Across Views
- Estimating divergence functionals and the likelihood ratio by convex risk minimization
- Contrastive Multiview Coding
- GENESIS: Generative Scene Inference and Sampling with Object-Centric Latent Representations
- GENESIS-V2: Inferring Unordered Object Representations without Iterative Refinement
- Self-Supervised Feature Learning by Learning to Spot Artifacts