Search papers, labs, and topics across Lattice.
Sungkyunkwan University
3
0
6
Action-aware reinforcement learning can boost code localization performance, achieving up to 8x better results than larger models.
LT-Mem achieves consistent object identity across sessions, enabling robots to answer complex queries about object histories without succumbing to temporal amnesia.
Existing factual consistency metrics fall short on real-world code, but ReFEree closes the gap with a reference-free, fine-grained approach that better aligns with human judgment.
