Model-Agnostic Graph Regularization for Few-Shot Learning
This work addresses few-shot learning challenges for domains with limited data by providing a simple, widely applicable regularization method, though it is incremental as it builds on existing graph-embedded approaches.
The paper tackles the problem of few-shot learning by incorporating knowledge graph information through a model-agnostic graph regularization approach, improving performance by up to 2% on Mini-ImageNet and 6.7% on ImageNet-FS compared to state-of-the-art methods.
In many domains, relationships between categories are encoded in the knowledge graph. Recently, promising results have been achieved by incorporating knowledge graph as side information in hard classification tasks with severely limited data. However, prior models consist of highly complex architectures with many sub-components that all seem to impact performance. In this paper, we present a comprehensive empirical study on graph embedded few-shot learning. We introduce a graph regularization approach that allows a deeper understanding of the impact of incorporating graph information between labels. Our proposed regularization is widely applicable and model-agnostic, and boosts the performance of any few-shot learning model, including fine-tuning, metric-based, and optimization-based meta-learning. Our approach improves the performance of strong base learners by up to 2% on Mini-ImageNet and 6.7% on ImageNet-FS, outperforming state-of-the-art graph embedded methods. Additional analyses reveal that graph regularizing models result in a lower loss for more difficult tasks, such as those with fewer shots and less informative support examples.