Unsupervised Projection Models

Unsupervised Projection Models these methods fracture the input feature space using randomly drawn hyperplanes independently of the data distribution. Given the random nature of the learning procedure they are the fastest models at training time but suffer from the disadvantage of requiring long hashcodes and many hashtables to attain a reasonable level of retrieval effectiveness.

Name	Codebook	Architecture	Optimisation
, .	Binary	Shallow	Leech Lattice
, .	Binary	Shallow	Randomly Rotated Cross-Polytopes
, .	Binary	Shallow	Random Hyperplanes
, .	Binary	Shallow	Bit Sampling
, .	Binary	Shallow	Random Hyperplanes
, .	Binary	Shallow	Random Hyperplanes
, .	Binary	Shallow	Random Hyperplanes
, .	Binary	Shallow	Random Fourier Features
, .	Binary	Shallow	Random Permutations