Deep Learning (DL) based models have fundamentally changed how we model natural language. This talk will cover our efforts relating to scenarios focusing on improving (a) retrieval and ranking of web documents, (b) question-answering over web-scale unstructured textual data, and (c) template free natural language generation for Cortana based QnA scenario. At the end of the talk we will also present some exploratory work along the lines of universal grammar to have a single semantic representation across different languages.
Saurabh Tiwary leads the deep learning efforts for NLP at Microsoft AI & Research. His team works on various aspects of search and question-answering for web and its applications to enterprise domain (Microsoft Search). Deep learning models from his team are today serving 100s of millions of users. His team is the driver for FPGA based deep learning model inferencing effort of Microsoft. He has previously worked at Cadence Research Labs and Google. At Google, he worked on pagerank, synonyms, and intent detection for web pages as part of the search quality team. He also incubated a project which ended up as Google Cloud Talent Solution. He has a Bachelor degree from IIT Kanpur. He obtained his Masters and PhD from Carnegie Mellon University. His PhD work was nominated for the SRC best dissertation award.