Predicting Congressional Swing Voters
April 2012 - June 2012
This is a project we did for the course CS145 at Caltech. The goal of the project was to predict the chance of passing a bill in the Congress, and if a bill is on the borderline of passing, determine the swing voters that a lobbyist should target for the bill to pass. Instead of analyzing the content of the given bill to determine congressmen's opinions, we completely disregarded the bill content and chose to analyze the mutual influence between congressmen as probabilistic graphical models. We used machine learning algorithms to determine the graphical model from the past Congress voting record and then used the model to predict bill outcomes and identify swing congressmen that are "cheap" to lobby. Here is our final report, and our poster that was presented in the CS145 poster session.
Crowdsourcing for Event Classification in NOνA Particle Physics Experiment
June 2011 - September 2011
This is a SURF project I did under Professor Ryan Patterson. The NuMI Off-axis νe Appearance Experiment (NOνA) aims to probe the nature of neutrinos through the phenomenon of neutrino oscillation. In NOνA, a crucial task is to determine the types of neutrino interactions in the detector, which produces different patterns. Some of these patterns are difficult to distinguish by computers and the goal of this project is to use crowdsourcing to identify these patterns both accurately and within an acceptable cost. We built a prototype platform for such crowdsourcing and tested the performance of the crowdsourcing scheme. Here is the final report, and the slides for SURF final presentation.