Musical query-by-description as a multiclass learning problem

Authors:
  • BWB. Whitman
    B. Whitman
  • RRR. Rifkin
    R. Rifkin

Abstract
We present the query-by-description (QBD) compo- nent of "Kandem," a time-aware music retrieval system. The QBD system we describe learns a relation between descriptive text con- cerning a musical artist and their actual acoustic output, making such queries as "Play me something loud with an electronic beat" possible by merely analyzing the audio content of a database. We show a novel machine learning technique based on Regularized Least-Squares Classification (RLSC) that can quickly and effi- ciently learn the non-linear relation between descriptive language and audio features by treating the problem as a large number of possible output classes linked to the same set of input features. We show how the RLSC training can easily eliminate irrelevant labels. In this article, we present a query-by-description system for music, using the tenets of language processing, informa- tion retrieval and machine learning. Our system treats the re- lation between words and audio content as a 'severe multi- class' learning problem: given audio content with thousands of known labels (descriptions), only some of which are even rele- vant (reflect anything in the underlying audio data), we train a machine for each label. We discuss a novel machine learning technique using regularized least-squares classification (RLSC) that makes such multi-class problems tractable and show how we can eliminate incorrect classifications and increase query- by-description accuracy. II. BACKGROUND

This publication has 6 references indexed in Scilit: