Cloudera Fast Forward Labs’ latest applied machine learning research report is about boosting natural language processing (NLP) with transfer learning. Organizations large and small have volumes of valuable data stored as free-form text yet the scale of data combined with the complexities of language processing makes using it to drive insight and automation a challenge.

While deep learning applied to NLP produces models powerful enough to automatically answer questions, translate languages, detect emotion, and even generate human-like language, these models are complex and unwieldy, making them too costly to build and incorporate into real-world systems without vast data, skilled human experts, and expensive infrastructure.

This report introduces transfer learning, a method of training models that solves these problems through knowledge reuse. You’ll see how deep learning for NLP and transfer learning are combined to improve the accuracy and robustness of models and reduce the cost of using advanced techniques, making NLP more practical, accessible and powerful than ever before.