We present FewShotTextGCN, a novel method designed to effectively utilize the properties of word-document graphs for improved learning in low-resource settings. We introduce K-hop Neighbourhood Regularization, a regularizer for heterogeneous graphs, and show that it stabilizes and improves learning when only a few training samples are available. We furthermore propose a simplification in the graph-construction method, which results in a graph that is ∼7 times less dense and yields better performance in little-resource settings while performing on par with the state of the art in high-resource settings. Finally, we introduce a new variant of Adaptive Pseudo-Labeling tailored for word-document graphs. When using as little as 20 samples for training, we outperform a strong TextGCN baseline with 17% in absolute accuracy on average over eight languages. We demonstrate that our method can be applied to document classification without any language model pretraining on a wide range of typologically diverse languages while performing on par with large pretrained language models.

The great majority of languages in the world are considered under-resourced for successful application of deep learning methods. In this work, we propose a meta-learning approach to document classification in low-resource languages and demonstrate its effectiveness in two different settings: few-shot, cross-lingual adaptation to previously unseen languages; and multilingual joint-training when limited target-language data is available during trai-ing. We conduct a systematic comparison of several meta-learning methods, investigate multiple settings in terms of data availability, and show that meta-learning thrives in settings with a heterogeneous task distribution. We propose a simple, yet effective adjustment to existing meta-learning methods which allows for better and more stable learning, and set a new state-of-the-art on a number of languages while performing on-par on others, using only a small amount of labeled data.