Search papers, labs, and topics across Lattice.
3
0
7
Online reinforcement fine-tuning can measurably improve the safety and efficiency of diffusion-based multi-agent driving planners, even when starting from strong pre-trained models.
Instead of blindly imputing missing data, ProMMA first evaluates the importance of each modality, leading to state-of-the-art multimodal sentiment analysis even when data is incomplete.
LLMs can now translate natural language commands into provably safe UAV flight plans, even repairing infeasible specifications on the fly.
