Weifeng Zhong on Machine Learning and the Policy Change Index

On this episode of Virtual Sentiments, Kristen interviews Weifeng Zhong of the Mercatus Center on his work with the Policy Change Index (PCI), a series of open-source machine learning projects that predict authoritarian regimes’ major policy moves by “reading” their propaganda publications.  Weifeng explains how his shocking revelation about the Tianneman Square massacre inspired him to create the PCI and details the ways in which it has evolved over the years, particularly as a means of "watching the watchers."  Additionally, he gives a brief overview of China's recent policy changes, specifically concerning when its liberalization began to reverse course.  Later, Kristen and Weifeng discuss the problems associated with machine learning algorithms, including whether bias is an automatic part of any machine learning process, and talk about what can be done to mitigate the current problems associated with machine learning.

About Virtual Sentiments

Virtual Sentiments is a new podcast from the Hayek Program in which Kristen Collins interviews scholars and practitioners grappling with pressing problems in political economy with an eye to the past.