Classification and comparison via neural networks. (October 2019)

• Record Type: Journal Article
• Authors: Yıldız, İlkay; Tian, Peng; Dy, Jennifer; Erdoğmuş, Deniz; Brown, James; Kalpathy-Cramer, Jayashree; Ostmo, Susan; Peter Campbell, J.; Chiang, Michael F.; Ioannidis, Stratis
• Is Part Of: Neural networks. Volume 118(2019)
• Page Start: 65
• Page End: 80
• Publication Date: 2019-10
• Subjects: Neural network -- Joint learning -- Comparison -- Classification -- Siamese network; Neural computers -- Periodicals; Neural networks (Computer science) -- Periodicals; Neural networks (Neurobiology) -- Periodicals; Nervous System -- Periodicals; Ordinateurs neuronaux -- Périodiques; Réseaux neuronaux (Informatique) -- Périodiques; Réseaux neuronaux (Neurobiologie) -- Périodiques; Neural computers; Neural networks (Computer science); Neural networks (Neurobiology); Periodicals; 006.32
• DOI: 10.1016/j.neunet.2019.06.004
• Languages: English
• ISSNs: 0893-6080

Abstract

Abstract: We consider learning from comparison labels generated as follows: given two samples in a dataset, a labeler produces a label indicating their relative order. Such comparison labels scale quadratically with the dataset size; most importantly, in practice, they often exhibit lower variance compared to class labels. We propose a new neural network architecture based on siamese networks to incorporate both class and comparison labels in the same training pipeline, using Bradley–Terry and Thurstone loss functions. Our architecture leads to a significant improvement in predicting both class and comparison labels, increasing classification AUC by as much as 35% and comparison AUC by as much as 6% on several real-life datasets. We further show that, by incorporating comparisons, training from few samples becomes possible: a deep neural network of 5.9 million parameters trained on 80 images attains a 0.92 AUC when incorporating comparisons.