Search papers, labs, and topics across Lattice.
2
0
1
Achieving 3.8x higher recall with a co-designed framework that integrates graph construction, representation learning, and real-time serving could redefine large-scale recommendation systems.
A new shuffle-DP protocol achieves 10x better utility and 3x better robustness against data poisoning for distribution estimation, all while using a single message per user.
