Learning Deep Parsimonious Representations
In this paper we aim at facilitating generalization for deep networks while supporting interpretability of the learned representations. Towards this goal, we propose a clustering based regularization that encourages parsimonious representations. Our k-means style objective is easy to optimize and flexible supporting various forms of clustering, including sample and spatial clustering as well as co-clustering. We demonstrate the effectiveness of our approach on the tasks of unsupervised learning, classification, fine grained categorization and zero-shot learning.
PDF Abstract
Datasets
CIFAR-10
CIFAR-100
MNIST
CUB-200-2011
Results from the Paper
| Task | Dataset | Model | Metric Name | Metric Value | Global Rank | Benchmark |
|---|---|---|---|---|---|---|
| Few-Shot Image Classification | CUB-200 - 0-Shot Learning | Sample Clustering | Accuracy | 44.3% | # 3 |
