Researcher

Wei Dong

Wei Dong is a computer scientist whose work spans large image datasets and nearest-neighbor search. He co-wrote the ImageNet paper, which appeared at the 2009 IEEE Conference on Computer Vision and Pattern Recognition. Two 2008 papers tuned locality-sensitive hashing and estimated distances with sketches. In 2011 he completed a computer science PhD at Princeton University, and his NN-Descent method, presented at WWW 2011, builds approximate k-nearest neighbor graphs for arbitrary similarity measures. Ann Arbor Algorithms, which he founded in 2014, develops KGraph, a library for approximate nearest neighbor search.

Works
63
Citations
62,455
h-index
18

Publication numbers from OpenAlex, as of 15 October 2026.

Timeline

  1. 2008

    Modeling LSH for performance tuning

    He writes Modeling LSH for performance tuning, a paper on locality-sensitive hashing published at CIKM in 2008, with Zhe Wang, William Josephson, Moses Charikar and Kai Li.

  2. 2008

    Asymmetric distance estimation with sketches

    With Moses Charikar and Kai Li, he writes a SIGIR 2008 paper on asymmetric distance estimation with sketches.

  3. 2009

    Co-writing the ImageNet paper

    He co-writes the ImageNet paper, which appears at the 2009 IEEE Conference on Computer Vision and Pattern Recognition with five co-authors.

  4. 2011

    A Princeton computer science PhD

    He completes a computer science PhD at Princeton University in 2011.

  5. 2011

    NN-Descent for nearest neighbors

    With Moses Charikar and Kai Li, he presents NN-Descent at WWW 2011, a method for approximate k-nearest neighbor graphs with arbitrary similarity measures.

  6. 2014

    Founding Ann Arbor Algorithms

    He founds Ann Arbor Algorithms and starts doing business under that name in 2014.

  7. Since 2014

    Developing KGraph and Donkey

    The firm develops KGraph, a library for approximate nearest neighbor search, and Donkey, a feature-vector database.

Sources and credits

Photo: We have no published photo of Wei Dong yet. Send us a free-licensed photo.

  1. Modeling LSH for performance tuning - Princeton University collaborate.princeton.edu. Checked 15 October 2026.
  2. Asymmetric distance estimation with sketches for similarity search in high-dimensional spaces - Princeton University collaborate.princeton.edu. Checked 15 October 2026.
  3. ImageNet: A Large-scale Hierarchical Image Database citation ruofeidu.com. Checked 15 October 2026.
  4. Ann Arbor Algorithms aaalgo.com. Checked 15 October 2026.
  5. Efficient K-nearest neighbor graph construction for generic similarity measures - Princeton University collaborate.princeton.edu. Checked 15 October 2026.