Search papers, labs, and topics across Lattice.
Affiliation:
4
0
6
Pruning visual tokens based on head alignment can retain nearly all performance while drastically reducing computational costs.
Click intent matters鈥擬ARCO achieves nearly perfect calibration and boosts conversions by 2.80% by treating clicks as distinct signals rather than uniform events.
Forget hand-tuning: VisPCO automatically finds optimal visual token pruning configurations in VLMs, outperforming predefined strategies across diverse benchmarks.
Forget memorizing surface patterns: RADAR leverages reinforcement learning to teach LLMs genuine relational reasoning, boosting knowledge graph performance by 5-6%.
