Conversational Question Answering

Project 2022-23


Project Title: Conversational Question Answering for Scientific Domains



Professor: Hamed Zamani


Lab/Research Group: Center for Intelligent Information Retrieval (CIIR)



Project Description

Intelligent assistants, such as Amazon Alexa, Apple Siri, and Google Assistant, have attracted much attention in industry. However, currently they are only able to successfully respond to a few types of requests. For instance, they are incapable of addressing several information seeking tasks. In this project, we focus on conversational question answering for the scholarly domain.

The assigned team would be responsible to learn how to use Macaw, an open-source platform for conversational information seeking research, and further improve the platform by implementing state-of-the-art solutions for a number of conversational question answering tasks. The outcome of this project will be released as an open-source extension to Macaw and may be published. Programming experiences in Python is a requirement for this project. Basic knowledge of machine learning is a plus.

For more information, please see the github page for the Macaw project: https://github.com/microsoft/macaw

Learning Objectives:

  1. Learning about question answering and conversational AI systems.

  2. Learning to work with deep learning libraries, such as TensorFlow and HuggingFace

  3. Learning about cutting edge research problems in the field of Information Retrieval