An Efficient Posterior Regularized Latent Variable Model for Interactive Sound Source Separation

Nicholas Bryan, Gautham Mysore
Proceedings of the 30th International Conference on Machine Learning, PMLR 28(3):208-216, 2013.

Abstract

In applications such as audio denoising, music transcription, music remixing, and audio-based forensics, it is desirable to decompose a single-channel recording into its respective sources. One of the current most effective class of methods to do so is based on non-negative matrix factorization and related latent variable models. Such techniques, however, typically perform poorly when no isolated training data is given and do not allow user feedback to correct for poor results. To overcome these issues, we allow a user to interactively constrain a latent variable model by painting on a time-frequency display of sound to guide the learning process. The annotations are used within the framework of posterior regularization to impose linear grouping constraints that would otherwise be difficult to achieve via standard priors. For the constraints considered, an efficient expectation-maximization algorithm is derived with closed-form multiplicative updates, drawing connections to non-negative matrix factorization methods, and allowing for high-quality interactive-rate separation without explicit training data.

Cite this Paper

BibTeX
@InProceedings{pmlr-v28-bryan13, title = {An Efficient Posterior Regularized Latent Variable Model for Interactive Sound Source Separation}, author = {Bryan, Nicholas and Mysore, Gautham}, booktitle = {Proceedings of the 30th International Conference on Machine Learning}, pages = {208--216}, year = {2013}, editor = {Dasgupta, Sanjoy and McAllester, David}, volume = {28}, number = {3}, series = {Proceedings of Machine Learning Research}, address = {Atlanta, Georgia, USA}, month = {17--19 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v28/bryan13.pdf}, url = {https://proceedings.mlr.press/v28/bryan13.html}, abstract = {In applications such as audio denoising, music transcription, music remixing, and audio-based forensics, it is desirable to decompose a single-channel recording into its respective sources. One of the current most effective class of methods to do so is based on non-negative matrix factorization and related latent variable models. Such techniques, however, typically perform poorly when no isolated training data is given and do not allow user feedback to correct for poor results. To overcome these issues, we allow a user to interactively constrain a latent variable model by painting on a time-frequency display of sound to guide the learning process. The annotations are used within the framework of posterior regularization to impose linear grouping constraints that would otherwise be difficult to achieve via standard priors. For the constraints considered, an efficient expectation-maximization algorithm is derived with closed-form multiplicative updates, drawing connections to non-negative matrix factorization methods, and allowing for high-quality interactive-rate separation without explicit training data.} }
Endnote
%0 Conference Paper %T An Efficient Posterior Regularized Latent Variable Model for Interactive Sound Source Separation %A Nicholas Bryan %A Gautham Mysore %B Proceedings of the 30th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2013 %E Sanjoy Dasgupta %E David McAllester %F pmlr-v28-bryan13 %I PMLR %P 208--216 %U https://proceedings.mlr.press/v28/bryan13.html %V 28 %N 3 %X In applications such as audio denoising, music transcription, music remixing, and audio-based forensics, it is desirable to decompose a single-channel recording into its respective sources. One of the current most effective class of methods to do so is based on non-negative matrix factorization and related latent variable models. Such techniques, however, typically perform poorly when no isolated training data is given and do not allow user feedback to correct for poor results. To overcome these issues, we allow a user to interactively constrain a latent variable model by painting on a time-frequency display of sound to guide the learning process. The annotations are used within the framework of posterior regularization to impose linear grouping constraints that would otherwise be difficult to achieve via standard priors. For the constraints considered, an efficient expectation-maximization algorithm is derived with closed-form multiplicative updates, drawing connections to non-negative matrix factorization methods, and allowing for high-quality interactive-rate separation without explicit training data.
RIS
TY - CPAPER TI - An Efficient Posterior Regularized Latent Variable Model for Interactive Sound Source Separation AU - Nicholas Bryan AU - Gautham Mysore BT - Proceedings of the 30th International Conference on Machine Learning DA - 2013/05/26 ED - Sanjoy Dasgupta ED - David McAllester ID - pmlr-v28-bryan13 PB - PMLR DP - Proceedings of Machine Learning Research VL - 28 IS - 3 SP - 208 EP - 216 L1 - http://proceedings.mlr.press/v28/bryan13.pdf UR - https://proceedings.mlr.press/v28/bryan13.html AB - In applications such as audio denoising, music transcription, music remixing, and audio-based forensics, it is desirable to decompose a single-channel recording into its respective sources. One of the current most effective class of methods to do so is based on non-negative matrix factorization and related latent variable models. Such techniques, however, typically perform poorly when no isolated training data is given and do not allow user feedback to correct for poor results. To overcome these issues, we allow a user to interactively constrain a latent variable model by painting on a time-frequency display of sound to guide the learning process. The annotations are used within the framework of posterior regularization to impose linear grouping constraints that would otherwise be difficult to achieve via standard priors. For the constraints considered, an efficient expectation-maximization algorithm is derived with closed-form multiplicative updates, drawing connections to non-negative matrix factorization methods, and allowing for high-quality interactive-rate separation without explicit training data. ER -
APA
Bryan, N. & Mysore, G.. (2013). An Efficient Posterior Regularized Latent Variable Model for Interactive Sound Source Separation. Proceedings of the 30th International Conference on Machine Learning, in Proceedings of Machine Learning Research 28(3):208-216 Available from https://proceedings.mlr.press/v28/bryan13.html.

Related Material