Search papers, labs, and topics across Lattice.
7
0
10
0
Unlock uncertainty quantification for complex-valued neural networks with BayesCVNNs, achieving 4.5x-13x speedups on FPGAs while slashing power consumption.
Forget pre-built maps: this new navigation agent interprets signs like a human, achieving 80% success in complex indoor environments.
MLLMs are surprisingly robust to catastrophic forgetting during fine-tuning, needing only simple regularization or data-hybrid training to maintain performance.
Robots can now navigate cluttered spaces more efficiently by directly "seeing" and tolerating contact with movable objects, thanks to a vision-language model that reasons about contact in image space.
Robots that learn from their mistakes *while* navigating? SERP unlocks this by evolving the action model in-context during replanning, boosting success rates and cutting token costs.
Stop wasting compute: PonderLM-3 learns to spend extra inference FLOPs only on the tokens that actually need them, outperforming fixed-step pondering methods.
Key contribution not extracted.
