Reproduce “O2N2: Efficient Doubly Oblivious Approximate Nearest Neighbor Search” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Zhechen Li, Zimai Guo, Mingxun Zhou, Chang Shi, Cheng Hong
O2N2 searches for approximate nearest neighbors while hiding memory-access patterns outside and inside a trusted execution environment. It follows one path in each independently rotated k-d tree, then selects distinct candidates with fixed-schedule operations. On SIFT1M 10-nearest-neighbor search at 90% recall, it takes 1.3 ms versus 161.3 ms for a doubly oblivious baseline, a result at one operating point.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.