Local Propagation for Few-Shot Learning
This work provides a more universally safe choice for few-shot inference for researchers and practitioners, reducing the need to select methods based on data quantity.
This paper tackles the problem of insufficient data in few-shot learning by combining local image representations with transductive inference. They introduce "local propagation," treating local image features as independent examples and building a graph to propagate features and labels, achieving improved accuracy over existing methods.
The challenge in few-shot learning is that available data is not enough to capture the underlying distribution. To mitigate this, two emerging directions are (a) using local image representations, essentially multiplying the amount of data by a constant factor, and (b) using more unlabeled data, for instance by transductive inference, jointly on a number of queries. In this work, we bring these two ideas together, introducing \emph{local propagation}. We treat local image features as independent examples, we build a graph on them and we use it to propagate both the features themselves and the labels, known and unknown. Interestingly, since there is a number of features per image, even a single query gives rise to transductive inference. As a result, we provide a universally safe choice for few-shot inference under both non-transductive and transductive settings, improving accuracy over corresponding methods. This is in contrast to existing solutions, where one needs to choose the method depending on the quantity of available data.