Search papers, labs, and topics across Lattice.
3
0
4
GFlowNets can be harnessed to learn optimal transport plans, bridging the gap between generative modeling and optimal transport theory.
Asynchronous Q-learning converges faster than you thought, with rates up to $n^{-1/6} \log^{4} (nS A)$ now proven for high-dimensional settings.
GFlowNets can provably learn shortest paths in arbitrary graphs by minimizing total flow, outperforming specialized methods on Rubik's Cube with less search.
