Unlocking the Answers with A.I. and Natural Language Processing | Online

About this event

With early methods dating back to the 1960s, question answering (QA) is one of the oldest research areas in artificial intelligence. Traditional QA systems typically rely on semantic parsing to translate natural language into a formal meaning representation that can be stored and queried in a knowledge base. Recently, neural network (NN) based QA methods have also shown great promise, especially on reading comprehension tasks.

The goal of this talk is two-fold. The first part provides a comprehensive overview of the QA landscape, both for symbolic methods with semantic parsing and end-to-end learning with NNs. The second part introduces two current implementations of both types of QA systems: SippyCup, a simple semantic parser written in Python, and DrQA, a scalable system for reading comprehension applied to open-domain question answering.