Efficient Gaussian Process Model on Class-Imbalanced Datasets for Generalized Zero-Shot Learning
This addresses the class imbalance issue in ZSL for researchers and practitioners, offering an efficient solution with incremental improvements over existing methods.
The paper tackles the problem of class imbalance in Zero-Shot Learning (ZSL) datasets by proposing a model that combines a neural network for latent feature embedding with a Gaussian Process regression to predict unseen class prototypes, achieving state-of-the-art performance on imbalanced benchmarks like AWA2, AWA1, and APY with an average training time of 5 minutes.
Zero-Shot Learning (ZSL) models aim to classify object classes that are not seen during the training process. However, the problem of class imbalance is rarely discussed, despite its presence in several ZSL datasets. In this paper, we propose a Neural Network model that learns a latent feature embedding and a Gaussian Process (GP) regression model that predicts latent feature prototypes of unseen classes. A calibrated classifier is then constructed for ZSL and Generalized ZSL tasks. Our Neural Network model is trained efficiently with a simple training strategy that mitigates the impact of class-imbalanced training data. The model has an average training time of 5 minutes and can achieve state-of-the-art (SOTA) performance on imbalanced ZSL benchmark datasets like AWA2, AWA1 and APY, while having relatively good performance on the SUN and CUB datasets.