A Strong Baseline for Semi-Supervised Incremental Few-Shot Learning
This addresses a realistic but incremental scenario in few-shot learning for machine learning researchers, focusing on improving model adaptation with mixed data.
The paper tackles the problem of semi-supervised incremental few-shot learning (S2 I-FSL), a complex setting combining limited labeled data, unlabeled examples, and performance on both base and novel classes, and proposes a method that achieves effectiveness across multiple benchmarks.
Few-shot learning (FSL) aims to learn models that generalize to novel classes with limited training samples. Recent works advance FSL towards a scenario where unlabeled examples are also available and propose semi-supervised FSL methods. Another line of methods also cares about the performance of base classes in addition to the novel ones and thus establishes the incremental FSL scenario. In this paper, we generalize the above two under a more realistic yet complex setting, named by Semi-Supervised Incremental Few-Shot Learning (S2 I-FSL). To tackle the task, we propose a novel paradigm containing two parts: (1) a well-designed meta-training algorithm for mitigating ambiguity between base and novel classes caused by unreliable pseudo labels and (2) a model adaptation mechanism to learn discriminative features for novel classes while preserving base knowledge using few labeled and all the unlabeled data. Extensive experiments on standard FSL, semi-supervised FSL, incremental FSL, and the firstly built S2 I-FSL benchmarks demonstrate the effectiveness of our proposed method.