Search papers, labs, and topics across Lattice.
Australian National University
4
0
9
LLMs can pinpoint mental states but falter at predicting dialogue trajectories, revealing a critical gap in their reasoning capabilities.
Achieve near-perfect robot grasping (87% success) by giving robots a "second look" – a closed-loop system that compares pre- and post-action states and uses text feedback to correct errors.
Domain-aware federated learning unlocks high-quality time series foundation models, even when training on highly heterogeneous datasets.
Stop letting foreground attention shifts sabotage your VLM prompt tuning: FVG-PT adaptively guides attention to boost performance.
