Search papers, labs, and topics across Lattice.
3
0
4
12
Annotations can be transformed into oracle rollouts, drastically improving the efficiency and scalability of reinforcement learning for video MLLMs.
P2Fusion achieves state-of-the-art fusion quality by adaptively regulating modal competition through learnable dynamic regulators, outperforming existing methods in 14 out of 20 evaluation metrics.
Multimodal queries in UAV imagery can significantly reduce visual-query ambiguity and improve target localization across diverse scenarios.