Search papers, labs, and topics across Lattice.
4
0
9
9
By combining multi-step latent self-prediction with observation-level dynamics, OG-SPR achieves superior performance in visual control tasks, revealing the limitations of traditional predictive methods.
Leveraging large language models to generate synthetic transitions, ProDVI boosts sample efficiency in deep reinforcement learning without the need for extensive pre-collected data or simulators.
Hawk boosts NPU kernel generation accuracy by over 30% while doubling execution speed, revolutionizing how we approach hardware-specific programming.
Edge devices can now run MoEs in real-time thanks to a dynamic quantization scheme that prioritizes important experts and critical layers.
