Abstract
Pallavi Patil, Kriti Myer, Ronak Zala, Arpit Singh, Sheshera Mysore, Andrew McCallum, Adrian Benton, Amanda Stent. Proceedings of the 23rd Conference on Computational Natural Language Learning (CoNLL). 2019.
Highlights
Roll call votes are official records of how politicians vote on bills in the United States House of Representatives and Senate
We evaluate the models under two settings: (Setting 1) in this setting all politicians are present in both training and test data; and (Setting 2) in this setting, votes from 5% of the politicians, chosen at random for each session of Congress, are removed from the training set5, resulting in a reduction of around 7,000 politician-bill-vote examples from each session of Congress
Our experiments attempt to answer several questions: (1) Setting 1: Does politician-related knowledge augmentation improve predictive performance when voting records of all politicians are observed (§4.0.1)? (2) Setting 2: Does politicianrelated knowledge augmentation improve performance when voting records of some politicians are not observed (§4.0.2)? (3) Does knowledge augmentation from a manually curated knowledge base (KB) improve model performance compared to knowledge augmentation using unstructured text (§4.0.3)? (4) for which politicians are our knowledge augmented models more effective than predicting a party majority (PARTYM) vote (§4.0.4)?
Summary
Roll call votes are official records of how politicians vote on bills (potential laws) in the United States House of Representatives and Senate. Prior work has used politicians’ voting records as a means to study their ideological stances (Poole and Rosenthal, 1985; Clinton et al, 2004), as well as roll call votes combined with the text of the corresponding bills to predict votes on newly drafted bills (Gerrish and Blei, 2012; Kraft et al, 2016; Kornilova et al, 2018) These approaches fail to make good predictions for the votes of politicians whose records are not established, such as new candidates for office – a time when this information can be most useful for the electorate. Information in a KB is likely to be more restricted, but more reliable, than information extracted from news articles
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
