STA-ViT: A Spatiotemporal Self-Attention Vision Transformer for Learning Behavior Recognition and Intervention. IJCAI. 2025;49(7). doi:10.31449/inf.v49i7.8653