Yuheng Ji

contrastive learning (1) reinforcement learning (1) adversarial robustness (1) robotic manipulation (1) trajectory prediction (1) adversarial training (1) zero-shot generalization (1) multimodal large language model (1) task planning (1) large vision-language model (1) contrastive loss (1) minority class (1) performance disparity (1) gradient reweighting (1) spatiotemporal graph learning (1) zero-inflated distribution (1) gradient distribution (1) verifiable reward (1) embodied manipulation (1) affordance perception (1)