Kang Yang

This work proposes GrayTrack, which fuses weak anonymous events with sparse direct observations using a road-constrained particle filter to extend tracking capabilities and demonstrates that GrayTrack can effectively exploit weak indirect observations to extend tracking capabilities.

SA-RSQ achieves a remarkable balance between compact storage and high-quality representation, leading to significant performance boosts in real-world recommender systems.

Uncertainty-Enhanced Collaborative Perception outperforms existing methods by effectively decoupling perception quality from detection noise, leading to more reliable autonomous driving systems.

Shuffling activations, a popular defense in secure Transformer inference, crumbles under a new alignment attack that recovers model weights for just $1.