Direct Post-Training Preference Alignment for Multi-Agent Motion Generation Model Using Implicit Feedback from Pre-training Demonstrations.

Thomas Tian, Kratarth Goel

Browse the full ICLR paper archive.