Worked on multi-agent simulation powered by generative AI. Built agents with backstory memories, textured histories, diverse hobbies and interests, and a series of distinct attributes. Co-authored a paper published in ACM UMAP '24.
>> Towards Multimodality: Comparing Quantifications of Movement Coordination
Extracted body movement data using pose-estimation method (MediaPipe) as time seriesExplored three methods to analyze time series similaries in order to quantify movement coordination
>> Computer Vision: Litter Dection in Coastal Maine
Created and annotated a 1k-image drone dataset for litter detection in coastal Maine, achieving ~92% model accuracy
Visualized litter using geolocation metadata to help local recycling processes and promote sustainability