Search papers, labs, and topics across Lattice.
Affiliation:
4
0
7
2
SALR is proposed, a schema-anchored latent reasoning method for LF construction that performs multi-step reasoning by generating continuous thoughts in the model's hidden states, thereby delaying the explicit commitment to LF decisions.
Explicitly learning spatial priorities boosts infrared target detection performance, achieving up to 89.00 F1 score with real-gaze supervision.
A simple data recipe can outperform complex reward engineering in enhancing long-context reasoning for large language models.
By explicitly modeling motion with a retina-inspired mechanism, MI-DETR achieves a remarkable +26.35 mAP@50 improvement over the best multi-frame baseline for infrared small target detection.
