Legislator Representation Learning with Social Context and Expert Knowledge

TL;DR

This paper introduces a novel framework for learning legislator representations by integrating social context and expert knowledge through heterogeneous networks and relational graph neural networks, improving performance in political modeling tasks.

Contribution

It presents a new approach that combines social context and expert knowledge for more holistic legislator representation learning, surpassing existing text and voting record methods.

Findings

Outperforms state-of-the-art in downstream political tasks

Learned representations correlate with socio-political factors

Social context and expert knowledge are crucial for accurate modeling

Abstract

Modeling the ideological perspectives of political actors is an essential task in computational political science with applications in many downstream tasks. Existing approaches are generally limited to textual data and voting records, while they neglect the rich social context and valuable expert knowledge for holistic evaluation. In this paper, we propose a representation learning framework of political actors that jointly leverages social context and expert knowledge. Specifically, we retrieve and extract factual statements about legislators to leverage social context information. We then construct a heterogeneous information network to incorporate social context and use relational graph neural networks to learn legislator representations. Finally, we train our model with three objectives to align representation learning with expert knowledge, model ideological stance consistency, and simulate the echo chamber phenomenon. Extensive experiments demonstrate that our learned representations successfully advance the state-of-the-art in three downstream tasks. Further analysis proves the correlation between learned legislator representations and various socio-political factors, as well as bearing out the necessity of social context and expert knowledge in modeling political actors.