Search papers, labs, and topics across Lattice.
4
0
6
0
A lightweight Q-value model can boost a 9B VLA's performance beyond that of a 27B model while reducing inference latency by 27%.
Tactile dynamics are crucial for contact-rich manipulation, and VT-WAM outperforms existing models by 26.67% to 35.84% by effectively integrating visual and tactile cues.
FORCE achieves a remarkable 79% increase in success rates for VLA models while eliminating the need for costly human interventions during training.
Intention-aware tool discovery can boost LLM agent performance by nearly 60% while slashing unnecessary complexity in tool management.
