Skip to content
Binary Search Tree sparse matrix and vector
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.

Binary Search Tree matrix and vector classes

This implements the sparse matrix and vector classes mentioned in Ewin Tang's paper A quantum-inspired classical algorithm for recommendation systems. I learned of that paper from a blog post by Scott Aaronson.

The data structures are not original to Tang; apparently they were used earlier by Kerenidis and Prakash.

The vector class supports storing w sparse entries in an n-dimensional space in O(w log^2 n) space, reading/writing in O(log^2 n) time, computing the norm of the vector in O(1) time, and sampling the coordinates of a vector, weighted by the square norm of the entries, in O(log^2 n) time.

The matrix class supports similar operations; see Tang's paper for details.

You can’t perform that action at this time.