Interpretable Few-shot Learning with Online Attribute Selection

TL;DR

This paper introduces an interpretable few-shot learning model that uses online attribute selection and attribute augmentation to improve accuracy and interpretability, aligning better with human understanding.

Contribution

The paper presents a novel interpretable FSL model with online attribute filtering and automatic attribute augmentation, addressing limitations of previous approaches.

Findings

Achieves comparable accuracy to black-box models on four datasets.

Improves interpretability through attribute selection and alignment with human understanding.

Outperforms existing methods in decision interpretability metrics.

Abstract

Few-shot learning (FSL) presents a challenging learning problem in which only a few samples are available for each class. Decision interpretation is more important in few-shot classification due to a greater chance of error compared to traditional classification. However, the majority of the previous FSL methods are black-box models. In this paper, we propose an inherently interpretable model for FSL based on human-friendly attributes. Previously, human-friendly attributes have been utilized to train models with the potential for human interaction and interpretability. However, such approaches are not directly extendible to the few-shot classification scenario. Moreover, we propose an online attribute selection mechanism to effectively filter out irrelevant attributes in each episode. The attribute selection mechanism improves accuracy and helps with interpretability by reducing the number of attributes that participate in each episode. We further propose a mechanism that automatically detects the episodes where the pool of available human-friendly attributes is insufficient, and subsequently augments it by engaging some learned unknown attributes. We demonstrate that the proposed method achieves results on par with black-box few-shot learning models on four widely used datasets. We also empirically evaluate the level of decision alignment between different models and human understanding and show that our model outperforms the comparison methods based on this criterion.