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

04/01/2021
by   Jong-Chyi Su, et al.
19

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.

READ FULL TEXT

page 2

page 10

11/23/2021

Semi-Supervised Learning with Taxonomic Labels

We propose techniques to incorporate coarse taxonomic labels to train im...
07/13/2021

Learning a Discriminant Latent Space with Neural Discriminant Analysis

Discriminative features play an important role in image and object class...
07/05/2018

Open Logo Detection Challenge

Existing logo detection benchmarks consider artificial deployment scenar...
05/02/2019

Billion-scale semi-supervised learning for image classification

This paper presents a study of semi-supervised learning with large convo...
02/05/2020

Extracting dispersion curves from ambient noise correlations using deep learning

We present a machine-learning approach to classifying the phases of surf...
08/17/2022

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...
06/28/2021

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

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

Code Repositories