_____________________________________________________________________
SUMMARY
Hubs are data points which persistently appear, unwanted, in the nearest neighbor lists of other data points. This effect is particularly problematic in similarity search algorithms, as the same objects are found over and over again. But it has also adverse effects for the many machine learning algorithms that make use of distance information.
The hubness problem has gained particular attention in the field of Music Information Retrieval (MIR) which is the interdisciplinary science of extracting information from music. The main goal of this project is to conduct an in-depth study of the hubness problem in the context of MIR with the aim of finding ways to avoid or at least attenuate its adverse effects.
_____________________________________________________________________
SOLUTIONS
- Discovering and delineating parameters of audio similarity which are less prone to hubness
- Transforming audio similarity spaces thereby avoiding the asymmetries that lead to hubness
- Considering audio similarity spaces as nearest neighbor graphs and using graph theoretic results to avoid hub nodes
_____________________________________________________________________
LINK: http://www.ofai.at/research/impml/projects/hubology.html
_____________________________________________________________________
