Meta-free few-shot learning via representation learning with weight averaging

TL;DR

This paper introduces a transfer learning approach called MFRL that improves few-shot learning accuracy and uncertainty calibration without episodic meta-learning, and demonstrates the effectiveness of weight averaging and temperature scaling.

Contribution

The paper proposes MFRL, a novel transfer learning method for few-shot tasks that avoids episodic meta-learning and enhances uncertainty estimation.

Findings

MFRL achieves state-of-the-art accuracy on few-shot benchmarks.

MFRL provides well-calibrated uncertainty estimates.

Weight averaging and temperature scaling improve existing meta-learning methods.

Abstract

Recent studies on few-shot classification using transfer learning pose challenges to the effectiveness and efficiency of episodic meta-learning algorithms. Transfer learning approaches are a natural alternative, but they are restricted to few-shot classification. Moreover, little attention has been on the development of probabilistic models with well-calibrated uncertainty from few-shot samples, except for some Bayesian episodic learning algorithms. To tackle the aforementioned issues, we propose a new transfer learning method to obtain accurate and reliable models for few-shot regression and classification. The resulting method does not require episodic meta-learning and is called meta-free representation learning (MFRL). MFRL first finds low-rank representation generalizing well on meta-test tasks. Given the learned representation, probabilistic linear models are fine-tuned with few-shot samples to obtain models with well-calibrated uncertainty. The proposed method not only achieves the highest accuracy on a wide range of few-shot learning benchmark datasets but also correctly quantifies the prediction uncertainty. In addition, weight averaging and temperature scaling are effective in improving the accuracy and reliability of few-shot learning in existing meta-learning algorithms with a wide range of learning paradigms and model architectures.