Helping Machines Be Conversational

by Michael McLaughlin August 23, 2018

written by Michael McLaughlin August 23, 2018

Stanford University researchers have released the Conversational Question Answering (CoQA) dataset to help machines better gather and provide information in conversations with humans. The dataset includes 127,000 questions from 8,000 different conversations. These conversations are from seven different types of text, including children’s stories, high school English exams, and Reddit. AI models often struggle to answer questions across different domains (i.e. news stories vs. English exams), and the researchers found that humans significantly outperformed reading comprehension models in answering the questions.

Get the data.

Michael McLaughlin

Michael McLaughlin is a research analyst at the Center for Data Innovation. He researches and writes about a variety of issues related to information technology and Internet policy, including digital platforms, e-government, and artificial intelligence. Michael graduated from Wake Forest University, where he majored in Communication with Minors in Politics and International Affairs and Journalism. He received his Master’s in Communication at Stanford University, specializing in Data Journalism.

Helping Machines Be Conversational

Using Twitter to Visualize Polarization

Event Recap: How Countries Are Preparing for the Global AI Race

You may also like