Social Systems and Machine Learning: From Book Publishing to Fairness
Big data analysis and machine learning techniques have allowed us to explore multiple problems through a more quantitative approach that reshapes social science and impacts every aspect of our daily life. In this thesis proposal, I propose to explore complex social systems phenomena with the aid of data analysis and machine learning. The first project is on the early prediction of the success of cultural products. The remaining projects aim to analyze bias in complex social systems, a prevailing yet important phenomenon. The second project aims to analyze and explain the gender bias against female artists in various aspects, including population, exhibition, and auction sales. The third and final project aims to develop an auditing tool that can automatically investigate data bias in real-world data, with the potential to influence policy-making and target the machine learning fairness problem from a different perspective.