Search papers, labs, and topics across Lattice.
University of Cambridge
3
0
3
Action precision in autonomous biliary navigation reaches 91.96% by integrating scene-aware learning with instruction conditioning.
Self-play in vectorized simulators allows TerraTransfer to achieve superior driving performance without the costly burden of expert data, reshaping the landscape of autonomous driving training.
Ditch hard clipping: ratio-variance regularization offers a principled, "soft brake" approach to trust region policy optimization, unlocking substantial gains in sample efficiency and performance, especially for smaller LLMs.
