Search papers, labs, and topics across Lattice.
The University of Texas at Austin
3
0
5
4
Forget hand-tuning: AutoScout automates ML system configuration, delivering up to 3x speedups over expert settings by jointly optimizing structural and execution parameters.
Stop hand-writing CUDA kernels: CUCo's agent-driven approach co-optimizes computation and communication, slashing LLM training/inference latency by up to 1.57x.
Forget hand-crafted benchmarks: this paper shows how LLMs can continuously generate relevant evaluation datasets for enterprise AI agents from just a few semi-structured documents.
