A Survey on Dynamic Neural Networks for Natural Language Processing
This is an incremental survey paper that organizes existing research on dynamic neural networks for NLP practitioners facing model scaling issues.
This survey examines dynamic neural networks as a solution to the computational scaling challenges of large Transformer models in NLP, summarizing progress on three approaches (skimming, mixture of experts, early exit) that enable sub-linear increases in computation while handling trillions of parameters.
Effectively scaling large Transformer models is a main driver of recent advances in natural language processing. Dynamic neural networks, as an emerging research direction, are capable of scaling up neural networks with sub-linear increases in computation and time by dynamically adjusting their computational path based on the input. Dynamic neural networks could be a promising solution to the growing parameter numbers of pretrained language models, allowing both model pretraining with trillions of parameters and faster inference on mobile devices. In this survey, we summarize progress of three types of dynamic neural networks in NLP: skimming, mixture of experts, and early exit. We also highlight current challenges in dynamic neural networks and directions for future research.