Search papers, labs, and topics across Lattice.
1
0
3
Distributing SciML models with hardware and physics awareness slashes latency and energy consumption by over 8x and 33x, respectively, while paradoxically *improving* reconstruction fidelity.