Very Fast Nearest Neighbor Retrieval in High-Dimensional Domains 7a.029.TUT

Project Start Date: Aug 1, 2018
Research Areas: Analytics, Analytics - Machine Learning, Analytics - Signal Processing, Data Management, Data Management - Big Data Platforms
Funding: Member Funded
Project Tags: ,

Project Summary

This project investigates variants of approximate nearest neighbor (ANN) search algorithms based on random projection trees.

The objectives are to develop and evaluate: i) scalable algorithms for ultrahigh-dimensional data by exploiting sparsity; ii) incremental algorithms for dynamic (streaming) data; iii) robust (fault-tolerant) parallelization techniques on GPU, cluster and cloud platforms.

Applications involve text, audio, image and other signal data.