Tag
This paper presents SAIL, a method for disentangling and representing human skill as interpretable multi-dimensional embeddings to improve behavior prediction and AI coaching, focusing on robustness and generalizability.