A Realistic Evaluation of Semi-Supervised Learning for Fine-Grained Classification

by   Jong-Chyi Su, et al.

We evaluate the effectiveness of semi-supervised learning (SSL) on a realistic benchmark where data exhibits considerable class imbalance and contains images from novel classes. Our benchmark consists of two fine-grained classification datasets obtained by sampling classes from the Aves and Fungi taxonomy. We find that recently proposed SSL methods provide significant benefits, and can effectively use out-of-class data to improve performance when deep networks are trained from scratch. Yet their performance pales in comparison to a transfer learning baseline, an alternative approach for learning from a few examples. Furthermore, in the transfer setting, while existing SSL methods provide improvements, the presence of out-of-class is often detrimental. In this setting, standard fine-tuning followed by distillation-based self-training is the most robust. Our work suggests that semi-supervised learning with experts on realistic datasets may require different strategies than those currently prevalent in the literature.


page 2

page 10


Semi-Supervised Learning with Taxonomic Labels

We propose techniques to incorporate coarse taxonomic labels to train im...

Learning a Discriminant Latent Space with Neural Discriminant Analysis

Discriminative features play an important role in image and object class...

Open Logo Detection Challenge

Existing logo detection benchmarks consider artificial deployment scenar...

Billion-scale semi-supervised learning for image classification

This paper presents a study of semi-supervised learning with large convo...

Extracting dispersion curves from ambient noise correlations using deep learning

We present a machine-learning approach to classifying the phases of surf...

How does the degree of novelty impacts semi-supervised representation learning for novel class retrieval?

Supervised representation learning with deep networks tends to overfit t...

Rail-5k: a Real-World Dataset for Rail Surface Defects Detection

This paper presents the Rail-5k dataset for benchmarking the performance...

Code Repositories