Shizhe Chen

“PointACT's architectural insight is clean: don't touch the pretrained language-vision backbone. Instead, build a separate 'point-action expert' that lets 3D geometry directly condition the action generation process at multiple spatial scales.”

“PointACT achieves 82.3% on RLBench-10Tasks versus 73.2% for EO1 (a strong 3B parameter 2D VLA), while running at 6.7 Hz versus EO1's 3.8 Hz (Table VIII, Appendix A).”