Benjamin R. Goldstein

The paper classifies evidence types of iNaturalist mammal observations using machine learning, creating a dataset with 1.33 million annotated observations.

This paper develops an AI-driven pipeline to estimate animal density from camera traps by automatically extracting key parameters using computer vision and a hierarchical distance function.

The paper shows that environmental niches of North American mammals vary spatially and are not conserved, challenging the niche conservatism hypothesis and suggesting local adaptations or unmodeled interactions.