Few-Shot Classification with Feature Map Reconstruction Networks
This work addresses the problem of improving few-shot classification accuracy and computational efficiency for researchers and practitioners working with limited data.
This paper redefines few-shot classification as a latent space reconstruction problem, where a network's ability to reconstruct a query feature map from support features predicts class membership. The proposed Feature Map Reconstruction Networks achieve consistent and substantial accuracy gains on four fine-grained benchmarks and are competitive on mini-ImageNet and tiered-ImageNet.
In this paper we reformulate few-shot classification as a reconstruction problem in latent space. The ability of the network to reconstruct a query feature map from support features of a given class predicts membership of the query in that class. We introduce a novel mechanism for few-shot classification by regressing directly from support features to query features in closed form, without introducing any new modules or large-scale learnable parameters. The resulting Feature Map Reconstruction Networks are both more performant and computationally efficient than previous approaches. We demonstrate consistent and substantial accuracy gains on four fine-grained benchmarks with varying neural architectures. Our model is also competitive on the non-fine-grained mini-ImageNet and tiered-ImageNet benchmarks with minimal bells and whistles.