# Alignment Based Matching Networks for One-Shot Classification and Open-Set Recognition

## TL;DR

This paper introduces an alignment-based matching network for one-shot classification and open-set recognition, improving accuracy and interpretability by explicitly aligning images to reference exemplars, and enabling recognition of unseen classes.

## Contribution

It proposes a novel exemplar-based cross-alignment method that enhances one-shot learning and open-set recognition, with learned alignments not requiring similar reference objects.

## Key findings

- Reduced 5-way 1-shot error rate on Omniglot from 2.1% to 1.4%.
- Lowered error rate on MiniImageNet from 53.5% to 46.5%.
- Achieved F1-score above 0.5 for open-set recognition with 19 distractors.

## Abstract

Deep learning for object classification relies heavily on convolutional models. While effective, CNNs are rarely interpretable after the fact. An attention mechanism can be used to highlight the area of the image that the model focuses on thus offering a narrow view into the mechanism of classification. We expand on this idea by forcing the method to explicitly align images to be classified to reference images representing the classes. The mechanism of alignment is learned and therefore does not require that the reference objects are anything like those being classified. Beyond explanation, our exemplar based cross-alignment method enables classification with only a single example per category (one-shot). Our model cuts the 5-way, 1-shot error rate in Omniglot from 2.1% to 1.4% and in MiniImageNet from 53.5% to 46.5% while simultaneously providing point-wise alignment information providing some understanding on what the network is capturing. This method of alignment also enables the recognition of an unsupported class (open-set) in the one-shot setting while maintaining an F1-score of above 0.5 for Omniglot even with 19 other distracting classes while baselines completely fail to separate the open-set class in the one-shot setting.