STUNT: Few-shot Tabular Learning with Self-generated Tasks from Unlabeled Tables
This addresses the challenge of high annotation costs and data scarcity in industrial tabular machine learning applications, offering a novel approach for few-shot learning in this domain.
The paper tackles the problem of few-shot learning for tabular data, where labeled samples are scarce, by proposing STUNT, a framework that self-generates tasks from unlabeled tables and uses meta-learning, resulting in significant performance gains across benchmarks compared to prior methods.
Learning with few labeled tabular samples is often an essential requirement for industrial machine learning applications as varieties of tabular data suffer from high annotation costs or have difficulties in collecting new samples for novel tasks. Despite the utter importance, such a problem is quite under-explored in the field of tabular learning, and existing few-shot learning schemes from other domains are not straightforward to apply, mainly due to the heterogeneous characteristics of tabular data. In this paper, we propose a simple yet effective framework for few-shot semi-supervised tabular learning, coined Self-generated Tasks from UNlabeled Tables (STUNT). Our key idea is to self-generate diverse few-shot tasks by treating randomly chosen columns as a target label. We then employ a meta-learning scheme to learn generalizable knowledge with the constructed tasks. Moreover, we introduce an unsupervised validation scheme for hyperparameter search (and early stopping) by generating a pseudo-validation set using STUNT from unlabeled data. Our experimental results demonstrate that our simple framework brings significant performance gain under various tabular few-shot learning benchmarks, compared to prior semi- and self-supervised baselines. Code is available at https://github.com/jaehyun513/STUNT.