DABS: A Domain-Agnostic Benchmark for Self-Supervised Learning

Self-supervised learning algorithms, including BERT and SimCLR, have enabled significant strides in fields like natural language processing, computer vision, and speech processing. However, these algorithms are domain-specific, meaning that new self-supervised learning algorithms must be developed for each new setting, including myriad healthcare, scientific, and multimodal domains. To catalyze progress toward domain-agnostic methods, we introduce DABS: a Domain-Agnostic Benchmark for Self-supervised learning. To perform well on DABS, an algorithm is evaluated on seven diverse domains: natural images, multichannel sensor data, English text, speech recordings, multilingual text, chest x-rays, and images with text descriptions. Each domain contains an unlabeled dataset for pretraining; the model is then scored based on its downstream performance on a set of labeled tasks in the domain. We also present e-Mix and ShED: the first domainagnostic algorithms evaluated on such a wide range of modalities. While e-Mix and ShED outperform a no-pretraining baseline, these improvements are uneven, demonstrating that significant progress is needed before self-supervised learning is an out-of-the-box solution for arbitrary domains. Code for the benchmark datasets and algorithms is available at https://github.com/alextamkin/dabs.