Task-Prior Conditional Variational Auto-Encoder for Few-Shot Image Classification
This addresses a more realistic and challenging scenario in few-shot learning for image classification, though it is incremental as it builds on existing transductive methods.
The paper tackles the problem of nonuniform few-shot image classification, where query shots per class vary, and proposes a Task-Prior Conditional Variational Auto-Encoder (TP-VAE) that improves accuracy, including a 3% increase in 1-shot scenarios.
Transductive methods always outperform inductive methods in few-shot image classification scenarios. However, the existing few-shot methods contain a latent condition: the number of samples in each class is the same, which may be unrealistic. To cope with those cases where the query shots of each class are nonuniform (i.e. nonuniform few-shot learning), we propose a Task-Prior Conditional Variational Auto-Encoder model named TP-VAE, conditioned on support shots and constrained by a task-level prior regularization. Our method obtains high performance in the more challenging nonuniform few-shot scenarios. Moreover, our method outperforms the state-of-the-art in a wide range of standard few-shot image classification scenarios. Among them, the accuracy of 1-shot increased by about 3\%.