Stance Detection and Ideology Prediction for Diverse Genres of Text 

RSVP for lunch requested.

Register to attend DS/CSS online events.

Abstract
News media often employ ideological language to sway their readers, e.g., criticizing people they disagree with, and praising those aligning with their values. In many cases, the media do not directly express their sentiment, but rather show their stances by narrating events and quoting others. In political news reporting, many events consist of individuals or groups who themselves are engaged in praise or blame. These seemingly objective decisions of choosing who to quote, and about what, may be shaped by ideological biases and have significant effects on readers. Therefore, it is important to examine the expressions of support and opposition in news articles and understand how media can bias readers via the selective inclusion of stances among entities. In this talk, I will present the first study on solving the task of entity-to-entity (E2E) stance detection in an end-to-end fashion. I will also demonstrate the usefulness of E2E stances for media stance characterization and entity-level ideology prediction.

For the second part of this talk, I will discuss how to effectively train general-purpose tools for measuring ideology that could be applied across different text genres. Importantly, we observe that media of different ideological leanings often choose different words and select different events for reporting the same news stories. Based on this observation, we design a new pretraining objective that compacts articles of similar ideologies while contrasting them with others of different leanings. Training on article clusters where each consists of pieces aligned on the same story, we produce a large model with competitive performance on ideology prediction and stance detection on varying genres of text.

Speaker bio

Lu Wang is an assistant professor in computer science and engineering at University of Michigan, Ann Arbor. Previously, she was an assistant professor in Khoury College of omputer Sciences at Northeastern University from 2015 to 2020. She received her PhD in computer science from Cornell University and her bachelor degrees in intelligent science and technology and economics from Peking University. Her research focuses on designing machine learning models for natural language processing tasks, including text summarization, language generation, reasoning, argument mining, and their applications in computational social science (e.g., media bias detection and narrative understanding). Lu received an outstanding paper award at ACL 2017 and a best paper nomination award at SIGDIAL 2012. She won the NSF CAREER award in 2021.