High-dimensional sequence transduction

TL;DR

This paper presents a probabilistic recurrent neural network model for high-dimensional sequence transduction, specifically for transcribing polyphonic audio into symbolic notation, achieving significant improvements over previous methods.

Contribution

It introduces a novel RNN-based probabilistic model and an efficient search algorithm for global mode, advancing polyphonic music transcription accuracy.

Findings

Outperforms previous state-of-the-art methods

Halves the test error rate on multiple datasets

Produces musically plausible transcriptions even with noise

Abstract

We investigate the problem of transforming an input sequence into a high-dimensional output sequence in order to transcribe polyphonic audio music into symbolic notation. We introduce a probabilistic model based on a recurrent neural network that is able to learn realistic output distributions given the input and we devise an efficient algorithm to search for the global mode of that distribution. The resulting method produces musically plausible transcriptions even under high levels of noise and drastically outperforms previous state-of-the-art approaches on five datasets of synthesized sounds and real recordings, approximately halving the test error rate.