BDL.NET: Bayesian dictionary learning in Infer.NET

Tom Diethe; Niall Twomey; Peter Flach

doi:10.1109/MLSP.2016.7738851

BDL.NET: Bayesian dictionary learning in Infer.NET

Tom Diethe, Niall Twomey, Peter Flach

Research output: Chapter in Book/Report/Conference proceeding › Conference Contribution (Conference Proceeding)

2 Citations (Scopus)

271 Downloads (Pure)

Abstract

We introduce and analyse a flexible and efficient implementation of Bayesian dictionary learning for sparse coding. By placing Gaussian-inverse-Gamma hierarchical priors on the coefficients, the model can automatically determine the required sparsity level for good reconstructions, whilst also automatically learning the noise level in the data, obviating the need for heuristic methods for choosing sparsity levels. This model can be solved efficiently using Variational Message Passing (VMP), which we have implemented in the Infer.NET framework for probabilistic programming and inference. We analyse the properties of the model via empirical validation on several accelerometer datasets. We provide source code to replicate all of the experiments in this paper.

Original language	English
Title of host publication	2016 IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2016 - Proceedings
Subtitle of host publication	Proceedings of a meeting held 13-16 September 2016, Vietri sul Mare (Salerno), Italy
Publisher	Institute of Electrical and Electronics Engineers (IEEE)
Number of pages	6
ISBN (Electronic)	9781509007462
ISBN (Print)	9781509007479
DOIs	https://doi.org/10.1109/MLSP.2016.7738851
Publication status	Published - Dec 2016
Event	26th IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2016 - Proceedings - Vietri sul Mare, Salerno, Italy Duration: 13 Sep 2016 → 16 Sep 2016

Publication series

Name	Proceedings of the IEEE International Workshop on Machine Learning for Signal Processing (MLSP)
Publisher	Institute of Electrical and Electronics Engineers (IEEE)
ISSN (Print)	1551-2541

Conference

Conference	26th IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2016 - Proceedings
Country	Italy
City	Vietri sul Mare, Salerno
Period	13/09/16 → 16/09/16

Structured keywords

Jean Golding

Keywords

Accelerometers
Bayesian
Dictionary Learning
Sparse Coding

Access to Document

10.1109/MLSP.2016.7738851

Full-text PDF (accepted author manuscript)
This is the accepted author manuscript (AAM). The final published version (version of record) is available online via IEEE at https://doi.org/10.1109/MLSP.2016.7738851 . Please refer to any applicable terms of use of the publisher.
Accepted author manuscript, 821 KB

Link to publication in Scopus

Fingerprint Dive into the research topics of 'BDL.NET: Bayesian dictionary learning in Infer.NET'. Together they form a unique fingerprint.

Projects

1 Finished

SPHERE (EPSRC IRC)

Craddock, I. J., Coyle, D. T., Flach, P. A., Kaleshi, D., Mirmehdi, M., Piechocki, R. J., Stark, B. H., Ascione, R., Ashburn, A. M., Burnett, M. E., Aldamen, D., Gooberman-Hill, R. J. S., Harwin, W. S., Hilton, G., Holderbaum, W., Holley, A. P., Manchester, V. A., Meller, B. J., Stack, E. & Gilchrist, I. D.

1/10/13 → 30/09/18

Project: Research, Parent

Cite this

Diethe, T., Twomey, N., & Flach, P. (2016). BDL.NET: Bayesian dictionary learning in Infer.NET. In 2016 IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2016 - Proceedings: Proceedings of a meeting held 13-16 September 2016, Vietri sul Mare (Salerno), Italy [7738851] (Proceedings of the IEEE International Workshop on Machine Learning for Signal Processing (MLSP)). Institute of Electrical and Electronics Engineers (IEEE). https://doi.org/10.1109/MLSP.2016.7738851

Diethe, Tom ; Twomey, Niall ; Flach, Peter. / BDL.NET : Bayesian dictionary learning in Infer.NET. 2016 IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2016 - Proceedings: Proceedings of a meeting held 13-16 September 2016, Vietri sul Mare (Salerno), Italy. Institute of Electrical and Electronics Engineers (IEEE), 2016. (Proceedings of the IEEE International Workshop on Machine Learning for Signal Processing (MLSP)).

@inproceedings{3f2a189d2267441baeaa0b07b4e46cd2,

title = "BDL.NET: Bayesian dictionary learning in Infer.NET",

abstract = "We introduce and analyse a flexible and efficient implementation of Bayesian dictionary learning for sparse coding. By placing Gaussian-inverse-Gamma hierarchical priors on the coefficients, the model can automatically determine the required sparsity level for good reconstructions, whilst also automatically learning the noise level in the data, obviating the need for heuristic methods for choosing sparsity levels. This model can be solved efficiently using Variational Message Passing (VMP), which we have implemented in the Infer.NET framework for probabilistic programming and inference. We analyse the properties of the model via empirical validation on several accelerometer datasets. We provide source code to replicate all of the experiments in this paper.",

keywords = "Accelerometers, Bayesian, Dictionary Learning, Sparse Coding",

author = "Tom Diethe and Niall Twomey and Peter Flach",

year = "2016",

month = dec,

doi = "10.1109/MLSP.2016.7738851",

language = "English",

isbn = "9781509007479",

series = "Proceedings of the IEEE International Workshop on Machine Learning for Signal Processing (MLSP)",

publisher = "Institute of Electrical and Electronics Engineers (IEEE)",

booktitle = "2016 IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2016 - Proceedings",

address = "United States",

note = "26th IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2016 - Proceedings ; Conference date: 13-09-2016 Through 16-09-2016",

}

Diethe, T, Twomey, N & Flach, P 2016, BDL.NET: Bayesian dictionary learning in Infer.NET. in 2016 IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2016 - Proceedings: Proceedings of a meeting held 13-16 September 2016, Vietri sul Mare (Salerno), Italy., 7738851, Proceedings of the IEEE International Workshop on Machine Learning for Signal Processing (MLSP), Institute of Electrical and Electronics Engineers (IEEE), 26th IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2016 - Proceedings, Vietri sul Mare, Salerno, Italy, 13/09/16. https://doi.org/10.1109/MLSP.2016.7738851

BDL.NET : Bayesian dictionary learning in Infer.NET. / Diethe, Tom; Twomey, Niall; Flach, Peter.

2016 IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2016 - Proceedings: Proceedings of a meeting held 13-16 September 2016, Vietri sul Mare (Salerno), Italy. Institute of Electrical and Electronics Engineers (IEEE), 2016. 7738851 (Proceedings of the IEEE International Workshop on Machine Learning for Signal Processing (MLSP)).

Research output: Chapter in Book/Report/Conference proceeding › Conference Contribution (Conference Proceeding)

TY - GEN

T1 - BDL.NET

T2 - 26th IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2016 - Proceedings

AU - Diethe, Tom

AU - Twomey, Niall

AU - Flach, Peter

PY - 2016/12

Y1 - 2016/12

N2 - We introduce and analyse a flexible and efficient implementation of Bayesian dictionary learning for sparse coding. By placing Gaussian-inverse-Gamma hierarchical priors on the coefficients, the model can automatically determine the required sparsity level for good reconstructions, whilst also automatically learning the noise level in the data, obviating the need for heuristic methods for choosing sparsity levels. This model can be solved efficiently using Variational Message Passing (VMP), which we have implemented in the Infer.NET framework for probabilistic programming and inference. We analyse the properties of the model via empirical validation on several accelerometer datasets. We provide source code to replicate all of the experiments in this paper.

AB - We introduce and analyse a flexible and efficient implementation of Bayesian dictionary learning for sparse coding. By placing Gaussian-inverse-Gamma hierarchical priors on the coefficients, the model can automatically determine the required sparsity level for good reconstructions, whilst also automatically learning the noise level in the data, obviating the need for heuristic methods for choosing sparsity levels. This model can be solved efficiently using Variational Message Passing (VMP), which we have implemented in the Infer.NET framework for probabilistic programming and inference. We analyse the properties of the model via empirical validation on several accelerometer datasets. We provide source code to replicate all of the experiments in this paper.

KW - Accelerometers

KW - Bayesian

KW - Dictionary Learning

KW - Sparse Coding

UR - http://www.scopus.com/inward/record.url?scp=85002251286&partnerID=8YFLogxK

U2 - 10.1109/MLSP.2016.7738851

DO - 10.1109/MLSP.2016.7738851

M3 - Conference Contribution (Conference Proceeding)

AN - SCOPUS:85002251286

SN - 9781509007479

T3 - Proceedings of the IEEE International Workshop on Machine Learning for Signal Processing (MLSP)

BT - 2016 IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2016 - Proceedings

PB - Institute of Electrical and Electronics Engineers (IEEE)

Y2 - 13 September 2016 through 16 September 2016

ER -

Diethe T, Twomey N, Flach P. BDL.NET: Bayesian dictionary learning in Infer.NET. In 2016 IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2016 - Proceedings: Proceedings of a meeting held 13-16 September 2016, Vietri sul Mare (Salerno), Italy. Institute of Electrical and Electronics Engineers (IEEE). 2016. 7738851. (Proceedings of the IEEE International Workshop on Machine Learning for Signal Processing (MLSP)). https://doi.org/10.1109/MLSP.2016.7738851