Most search and recommender systems deal with large amounts of natural language data, hence an effective system requires a deep understanding of text semantics. Recently, deep learning based natural language processing (deep NLP) models have generated promising results.
In this talk, we will introduce DeText, a state-of-the-art open source NLP framework for text understanding. DeText is a flexible framework with BERT/CNN/LSTM encoders for text data processing, designed for efficient industry use cases. It has been applied in many productions at LinkedIn, such as search ranking, query auto completion, query intent prediction, etc. We will discuss:
Overview on DeText
Technical architect and design of DeText
Illustrate how neural ranking is designed and developed in DeText
More on DeText: Open Sourcing DeText
leads the NLP team at LinkedIn. He has broad interests in NLP and Information Retrieval.
Senior software engineer in the NLP team at LinkedIn where she focuses on applying state-of-the-art NLP algorithms to power and improve LinkedIn products.
senior software engineer in the NLP team at LinkedIn.
leads the AI Algorithms Foundation team at LinkedIn.