Search papers, labs, and topics across Lattice.
4
0
5
Explicitly learning spatial priorities boosts infrared target detection performance, achieving up to 89.00 F1 score with real-gaze supervision.
Visual reranking and active rejection in MMAgent-R$^2$ significantly boost retrieval accuracy in challenging KB-VQA tasks, outperforming traditional methods.
Online data selection can shift model behavior as much as explicit preference optimization, revealing a hidden layer of alignment influence.
Achieving a score of $2.68 \times 10^{-3}$ in a depth estimation challenge reveals the untapped potential of zero-shot learning in complex visual tasks.
