Unsupervised Meta-Learning Through Latent-Space Interpolation in Generative Models

Abstract

Several recently proposed unsupervised meta-learning approaches rely on synthetic meta-tasks created using techniques such as random selection, clustering and/or augmentation. In this work, we describe a novel approach that generates meta-tasks using generative models. The proposed family of algorithms generate pairs of in-class and out-of-class samples from the latent space in a principled way, allowing us to create synthetic classes forming the training and validation data of a meta-task. We find that the proposed approach, LAtent Space Interpolation Unsupervised Meta-learning (LASIUM), outperforms or is competitive with current unsupervised learning baselines on few-shot classification tasks on the most widely used benchmark datasets.

Cite

Text

Khodadadeh et al. "Unsupervised Meta-Learning Through Latent-Space Interpolation in Generative Models." International Conference on Learning Representations, 2021.

Markdown

[Khodadadeh et al. "Unsupervised Meta-Learning Through Latent-Space Interpolation in Generative Models." International Conference on Learning Representations, 2021.](https://mlanthology.org/iclr/2021/khodadadeh2021iclr-unsupervised/)

BibTeX

@inproceedings{khodadadeh2021iclr-unsupervised,
  title     = {{Unsupervised Meta-Learning Through Latent-Space Interpolation in Generative Models}},
  author    = {Khodadadeh, Siavash and Zehtabian, Sharare and Vahidian, Saeed and Wang, Weijia and Lin, Bill and Boloni, Ladislau},
  booktitle = {International Conference on Learning Representations},
  year      = {2021},
  url       = {https://mlanthology.org/iclr/2021/khodadadeh2021iclr-unsupervised/}
}