Search papers, labs, and topics across Lattice.
3
0
6
0
Achieve strong zero-shot performance in semantic column type detection by fine-tuning LLMs on synthetically generated pseudo-tables tailored to specific domains.
By selectively injecting teacher demonstrations only during failure, HAPO overcomes the limitations of both pure RL and mixed-policy optimization in sparse-reward RLVR, enabling models to surpass static teacher forcing.
Trajectory prediction models can now adapt to new environments far more effectively thanks to a meta-learning approach that dynamically adjusts learning rates based on online data characteristics.
