Analyzing Attention Mechanisms through Lens of Sample Complexity and Loss Landscape
Attention mechanisms have advanced state-of-the-art deep learning models in many machine learning tasks. Despite significant empirical gains, there is a lack of theoretical analyses on their effectiveness. In this paper, we address this problem by studying the sample complexity and loss landscape of attention-based neural networks. Our results show that, under mild assumptions, every local minimum of the attention model has low prediction error, and attention models require lower sample complexity than models without attention. Besides revealing why popular self-attention works, our theoretical results also provide guidelines for designing future attention models. Experiments on various datasets validate our theoretical findings.
PDF Abstract