An Experimental Analysis of the Entanglement Problem in Neural-Network-based Music Transcription Systems

Kelz R, Widmer G. An Experimental Analysis of the Entanglement Problem in Neural-Network-based Music Transcription Systems. In: AES Conference: 2017 AES International Conference on Semantic Audio. Audio Engineering Society; 2017. Paper 5-1. Available from: https://aes.org/publications/elibrary-page/?id=18761

@inproceedings{Kelz2017_18761,
  author = {Kelz, Rainer and Widmer, Gerhard},
  title = {{An Experimental Analysis of the Entanglement Problem in Neural-Network-based Music Transcription Systems}},
  booktitle = {AES Conference: 2017 AES International Conference on Semantic Audio},
  note = {Paper 5-1},
  year = {2017},
  month = jun,
  publisher = {Audio Engineering Society},
  url = {https://aes.org/publications/elibrary-page/?id=18761}
}

TY  - CPAPER
TI  - An Experimental Analysis of the Entanglement Problem in Neural-Network-based Music Transcription Systems
AU  - Kelz, Rainer
AU  - Widmer, Gerhard
T2  - AES Conference: 2017 AES International Conference on Semantic Audio
M1  - Paper 5-1
PY  - 2017
DA  - 2017/06/06
UR  - https://aes.org/publications/elibrary-page/?id=18761
PB  - Audio Engineering Society
LA  - en
AB  - Several recent polyphonic music transcription systems have utilized deep neural networks to achieve state of the art results on various benchmark datasets, pushing the envelope on framewise and note-level performance measures. Unfortunately we can observe a sort of glass ceiling effect. To investigate this effect, we provide a detailed analysis of the particular kinds of errors that state of the art deep neural transcription systems make, when trained and tested on a piano transcription task. We are ultimately forced to draw a rather disheartening conclusion: the networks seem to learn combinations of notes, and have a hard time generalizing to unseen combinations of notes. Furthermore, we speculate on various means to alleviate this situation.
ER  -