Project Details
Project Description
University of California San Diego - Monash University
The massive increase in the quantity of online information regarding politics and public policies presents both challenges and opportunities regarding the evaluation of political representatives and the policies they deliver. The quantity of this information makes it unfeasible to analyze it with traditional research methods, and the trustworthiness of some of the information is questionable. However, if harnessed correctly, information empowers citizens to hold political representatives to account and policymakers to identify the effectiveness of public policies.
The long-term vision of this program is to build an interuniversity and multidisciplinary capacity to understand and improve the quality of representative democracy and public policies. Political Science enables us to understand the processes of politics and policymaking that we seek to improve. In this first stage of the program, we focus on democratic representation through which citizens’ demands are responded to with effective public policies. Computer Science develops and applies technology for the analysis of large amounts of information. We focus on the Artificial Intelligence subfields of machine learning and natural language processing.
In addition to its social relevance to the quality of democracy and public policies, the program is of academic importance. In Political Science, traditional qualitative methods for studying text and speech during election campaigns, such as the qualitative coding of campaign promises and adverts, are becoming problematic as the quantity of information in campaigns increases. The technology we aim to develop and apply will enable more effective analyses. In Computer
Science, advances are made partly through the development of novel applications, such as our proposed application to text from election campaigns. This is expected to lead to advances in natural language processing, for instance in supervised text classification methods to identify the sentiment (positive, neutral or negative) in relation to specific units of text (e.g. specific campaign promises).
The massive increase in the quantity of online information regarding politics and public policies presents both challenges and opportunities regarding the evaluation of political representatives and the policies they deliver. The quantity of this information makes it unfeasible to analyze it with traditional research methods, and the trustworthiness of some of the information is questionable. However, if harnessed correctly, information empowers citizens to hold political representatives to account and policymakers to identify the effectiveness of public policies.
The long-term vision of this program is to build an interuniversity and multidisciplinary capacity to understand and improve the quality of representative democracy and public policies. Political Science enables us to understand the processes of politics and policymaking that we seek to improve. In this first stage of the program, we focus on democratic representation through which citizens’ demands are responded to with effective public policies. Computer Science develops and applies technology for the analysis of large amounts of information. We focus on the Artificial Intelligence subfields of machine learning and natural language processing.
In addition to its social relevance to the quality of democracy and public policies, the program is of academic importance. In Political Science, traditional qualitative methods for studying text and speech during election campaigns, such as the qualitative coding of campaign promises and adverts, are becoming problematic as the quantity of information in campaigns increases. The technology we aim to develop and apply will enable more effective analyses. In Computer
Science, advances are made partly through the development of novel applications, such as our proposed application to text from election campaigns. This is expected to lead to advances in natural language processing, for instance in supervised text classification methods to identify the sentiment (positive, neutral or negative) in relation to specific units of text (e.g. specific campaign promises).
| Status | Finished |
|---|---|
| Effective start/end date | 1/08/19 → 1/11/20 |