Neal Jean

Co-authored research on using Raman optical spectroscopy combined with deep learning for quick and accurate detection, identification, and antibiotic susceptibility testing of bacteria, crucial for combating infectious diseases.

Co-developed an unsupervised representation learning algorithm, Tile2Vec, to extend the distributional hypothesis to geospatial analysis, improving performance across various remote sensing tasks.

Co-authored work on leveraging geolocated Wikipedia articles to predict economic development, providing an innovative approach to socioeconomic indicator estimation.

Contributed to research on Semi-supervised Deep Kernel Learning (SSDKL), a regression model designed to leverage unlabeled data effectively by minimizing predictive variance, addressing challenges in data-scarce real-world problems.

Co-authored research demonstrating an accurate and scalable method to identify impoverished regions by applying machine learning to high-resolution satellite imagery, particularly in Africa.

Contributed to work exploring the use of transfer learning from deep features to enhance remote sensing capabilities and improve the accuracy of poverty mapping.

Authored an undergraduate honors thesis at Duke University examining the impact of guaranteed contracts on player performance in the NBA, specifically focusing on the 'contract year phenomenon'.