Learning from Few Samples: A Survey

TL;DR

This survey reviews and categorizes recent meta-learning techniques in computer vision aimed at enabling models to learn effectively from very few samples, highlighting their methods, evaluations, and future directions.

Contribution

It provides a comprehensive taxonomy and comparison of existing few-shot learning methods, analyzing their approaches and benchmarks in the computer vision domain.

Findings

Meta-learning techniques are categorized into data-augmentation, embedding, optimization, and semantics-based methods.

Benchmark results on Omniglot and MiniImagenet datasets show varied performance across techniques.

Future research directions include improving generalization and surpassing human-level performance.

Abstract

Deep neural networks have been able to outperform humans in some cases like image recognition and image classification. However, with the emergence of various novel categories, the ability to continuously widen the learning capability of such networks from limited samples, still remains a challenge. Techniques like Meta-Learning and/or few-shot learning showed promising results, where they can learn or generalize to a novel category/task based on prior knowledge. In this paper, we perform a study of the existing few-shot meta-learning techniques in the computer vision domain based on their method and evaluation metrics. We provide a taxonomy for the techniques and categorize them as data-augmentation, embedding, optimization and semantics based learning for few-shot, one-shot and zero-shot settings. We then describe the seminal work done in each category and discuss their approach towards solving the predicament of learning from few samples. Lastly we provide a comparison of these techniques on the commonly used benchmark datasets: Omniglot, and MiniImagenet, along with a discussion towards the future direction of improving the performance of these techniques towards the final goal of outperforming humans.