Learning on Probabilistic Modelling Using Grassmann Manifolds

Muhammad Ali, Junbin Gao

Research output: Book chapter/Published conference paperConference paper

Abstract

Probabilistic models are playing an important role as building blocks in more sophisticated geometric parametric models. Although probabilistic or parametric models are seemingly very straightforward for making Bayesian inferences from data that are assumed to be drawn from some model into consideration, the barrier to the Bayesian approach is the normalising constants naturally appearing with them. Our main focus in this paper is to make inferences on matrix variate parametric models via Maximum Likelihood Estimation (MLE) using a simple Bayesian approach. For calculating the value of the matrix based normalising constant we propose the Taylor expansion method for very high concentrated parameter. For inferences we have considered the matrix variate Bingham model on Grassman manifolds. The model is then tested for validity and performance on real World databases with best accuracy estimates.
Original languageEnglish
Title of host publicationProceedings of the 14th International Conference on Frontiers of Information Technology (FIT 2016)
Place of PublicationUnited States
PublisherIEEE, Institute of Electrical and Electronics Engineers
Pages241-246
Number of pages6
DOIs
Publication statusPublished - 21 Dec 2016
Event14th International Conference on Frontiers of Information Technology: FIT 2016 - Serena Hotel, Islamabad, Pakistan
Duration: 19 Dec 201621 Dec 2016
http://fit.edu.pk/index.php/conference-history

Conference

Conference14th International Conference on Frontiers of Information Technology
Abbreviated titleInternet of Things (IoT)
CountryPakistan
CityIslamabad
Period19/12/1621/12/16
Internet address

Fingerprint

Maximum likelihood estimation
Statistical Models

Cite this

Ali, M., & Gao, J. (2016). Learning on Probabilistic Modelling Using Grassmann Manifolds. In Proceedings of the 14th International Conference on Frontiers of Information Technology (FIT 2016) (pp. 241-246). United States: IEEE, Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/FIT.2016.051
Ali, Muhammad ; Gao, Junbin. / Learning on Probabilistic Modelling Using Grassmann Manifolds. Proceedings of the 14th International Conference on Frontiers of Information Technology (FIT 2016). United States : IEEE, Institute of Electrical and Electronics Engineers, 2016. pp. 241-246
@inproceedings{2a8a2b094677468a8e73004354de16b4,
title = "Learning on Probabilistic Modelling Using Grassmann Manifolds",
abstract = "Probabilistic models are playing an important role as building blocks in more sophisticated geometric parametric models. Although probabilistic or parametric models are seemingly very straightforward for making Bayesian inferences from data that are assumed to be drawn from some model into consideration, the barrier to the Bayesian approach is the normalising constants naturally appearing with them. Our main focus in this paper is to make inferences on matrix variate parametric models via Maximum Likelihood Estimation (MLE) using a simple Bayesian approach. For calculating the value of the matrix based normalising constant we propose the Taylor expansion method for very high concentrated parameter. For inferences we have considered the matrix variate Bingham model on Grassman manifolds. The model is then tested for validity and performance on real World databases with best accuracy estimates.",
author = "Muhammad Ali and Junbin Gao",
year = "2016",
month = "12",
day = "21",
doi = "10.1109/FIT.2016.051",
language = "English",
pages = "241--246",
booktitle = "Proceedings of the 14th International Conference on Frontiers of Information Technology (FIT 2016)",
publisher = "IEEE, Institute of Electrical and Electronics Engineers",
address = "United States",

}

Ali, M & Gao, J 2016, Learning on Probabilistic Modelling Using Grassmann Manifolds. in Proceedings of the 14th International Conference on Frontiers of Information Technology (FIT 2016). IEEE, Institute of Electrical and Electronics Engineers, United States, pp. 241-246, 14th International Conference on Frontiers of Information Technology, Islamabad, Pakistan, 19/12/16. https://doi.org/10.1109/FIT.2016.051

Learning on Probabilistic Modelling Using Grassmann Manifolds. / Ali, Muhammad; Gao, Junbin.

Proceedings of the 14th International Conference on Frontiers of Information Technology (FIT 2016). United States : IEEE, Institute of Electrical and Electronics Engineers, 2016. p. 241-246.

Research output: Book chapter/Published conference paperConference paper

TY - GEN

T1 - Learning on Probabilistic Modelling Using Grassmann Manifolds

AU - Ali, Muhammad

AU - Gao, Junbin

PY - 2016/12/21

Y1 - 2016/12/21

N2 - Probabilistic models are playing an important role as building blocks in more sophisticated geometric parametric models. Although probabilistic or parametric models are seemingly very straightforward for making Bayesian inferences from data that are assumed to be drawn from some model into consideration, the barrier to the Bayesian approach is the normalising constants naturally appearing with them. Our main focus in this paper is to make inferences on matrix variate parametric models via Maximum Likelihood Estimation (MLE) using a simple Bayesian approach. For calculating the value of the matrix based normalising constant we propose the Taylor expansion method for very high concentrated parameter. For inferences we have considered the matrix variate Bingham model on Grassman manifolds. The model is then tested for validity and performance on real World databases with best accuracy estimates.

AB - Probabilistic models are playing an important role as building blocks in more sophisticated geometric parametric models. Although probabilistic or parametric models are seemingly very straightforward for making Bayesian inferences from data that are assumed to be drawn from some model into consideration, the barrier to the Bayesian approach is the normalising constants naturally appearing with them. Our main focus in this paper is to make inferences on matrix variate parametric models via Maximum Likelihood Estimation (MLE) using a simple Bayesian approach. For calculating the value of the matrix based normalising constant we propose the Taylor expansion method for very high concentrated parameter. For inferences we have considered the matrix variate Bingham model on Grassman manifolds. The model is then tested for validity and performance on real World databases with best accuracy estimates.

U2 - 10.1109/FIT.2016.051

DO - 10.1109/FIT.2016.051

M3 - Conference paper

SP - 241

EP - 246

BT - Proceedings of the 14th International Conference on Frontiers of Information Technology (FIT 2016)

PB - IEEE, Institute of Electrical and Electronics Engineers

CY - United States

ER -

Ali M, Gao J. Learning on Probabilistic Modelling Using Grassmann Manifolds. In Proceedings of the 14th International Conference on Frontiers of Information Technology (FIT 2016). United States: IEEE, Institute of Electrical and Electronics Engineers. 2016. p. 241-246 https://doi.org/10.1109/FIT.2016.051