Talks and Poster Presentations (with Proceedings-Entry):
H. Pálmason, B. Jónsson, L. Amsaleg, M. Schedl, P. Knees:
"On Competitiveness of Nearest-Neighbor Based Music Classification: A Methodological Critique";
Talk: 10th International Conference on Similarity Search and Applications,
- 2017-10-06; in: "Similarity Search and Applications",
C. Beeck, F. Borutta, P. Kröger, T. Seidl (ed.);
Lecture Notes in Computer Science, Springer,
The traditional role of nearest-neighbor classification in music classification research is that of a straw man opponent for the learning approach of the hour. Recent work in high-dimensional indexing has shown that approximate nearest-neighbor algorithms are extremely scalable, yielding results of reasonable quality from billions of high-dimensional features. With such efficient large-scale classifiers, the traditional music classification methodology of aggregating and compressing the audio features is incorrect; instead the approximate nearest-neighbor classifier should be given an extensive data collection to work with. We present a case study, using a well-known MIR classification benchmark with well-known music features, which shows that a simple nearest-neighbor classifier performs very competitively when given ample data. In this position paper, we therefore argue that nearest-neighbor classification has been treated unfairly in the literature and may be much more competitive than previously thought.
Music classification, Approximate nearest-neighbor classifiers, Research methodology
"Official" electronic version of the publication (accessed through its Digital Object Identifier - DOI)
Project Head Peter Knees:
Created from the Publication Database of the Vienna University of Technology.