Representation Learning and NLP
Zhiyuan Liu, Yankai Lin, Maosong Sun · 2020
Abstract Natural languages are typical unstructured information. Conventional Natural Language Processing (NLP) heavily relies on feature engineering, which requires careful design and considerable expertise. Representation learning aims to learn representations of raw data as useful information for further classification or prediction. This chapter presents a brief introduction to representation learning, including its motivation and basic idea, and also reviews its history and recent advances in both machine learning and NLP.