Strong Baselines for Neural Semi-supervised Learning under Domain Shift
This work addresses the difficulty in comparing models for domain shift in semi-supervised learning, showing that incremental improvements to classic methods can outperform recent neural approaches.
The paper tackled the problem of neural semi-supervised learning under domain shift by re-evaluating classic bootstrapping methods and proposing a novel multi-task tri-training approach, finding that classic tri-training with additions outperforms state-of-the-art methods on benchmarks, establishing a new state-of-the-art for sentiment analysis but not consistently across tasks.
Novel neural models have been proposed in recent years for learning under domain shift. Most models, however, only evaluate on a single task, on proprietary datasets, or compare to weak baselines, which makes comparison of models difficult. In this paper, we re-evaluate classic general-purpose bootstrapping approaches in the context of neural networks under domain shifts vs. recent neural approaches and propose a novel multi-task tri-training method that reduces the time and space complexity of classic tri-training. Extensive experiments on two benchmarks are negative: while our novel method establishes a new state-of-the-art for sentiment analysis, it does not fare consistently the best. More importantly, we arrive at the somewhat surprising conclusion that classic tri-training, with some additions, outperforms the state of the art. We conclude that classic approaches constitute an important and strong baseline.