Computer Vision meets Mechanism Design: Augmenting OCR to Reverse Engineer Historical Matching Algorithms
This project focuses on applying computer vision and OCR (Optical Character Recognition) techniques to analyze printed data outputs generated by a college allocation program. The data, structured according to the Deferred Acceptance algorithm, provides a unique opportunity to innovate within the fields of computer vision and mechanism design.
The research assistant will work on developing and implementing a script that leverages traditional OCR methods, augmented by an understanding of the underlying algorithm and data structure, to accurately read and interpret the printed output. The ultimate goal is to replicate the data-generating code and improve data collection efficiency. This project will explore the intersection of computer vision and economic theory, offering a novel approach to reverse engineering historical matching algorithms.
The RA will work approximately 10 hours per week for 12 weeks.
Requisite Skills and Qualifications
The ideal candidate should have:
- A strong background in computer science, particularly in areas related to computer vision and OCR.
- Familiarity with programming languages such as Python, and experience with associated libraries.
- An interest in economic theory, particularly in algorithmic game theory or mechanism design, is a plus.
- Ability to work independently and contribute to innovative problem-solving.