Dynamic Factorization Tests: Applications to Multi-modal Data Association

Michael R. Siracusa, John W. Fisher III
Proceedings of the Eleventh International Conference on Artificial Intelligence and Statistics, PMLR 2:508-515, 2007.

Abstract

The goal of a dynamic dependency test is to correctly label the interaction of multiple observed data streams and to describe how this interaction evolves over time. To this end, we propose the use of a hidden factorization Markov model (HFactMM) in which a hidden state indexes into a finite set of possible dependence structures on observations. We show that a dynamic dependency test using an HFactMM takes advantage of both structural and parametric changes associated with changes in interaction. This is contrasted both theoretically and empirically with standard sliding window based dependence analysis. Using this model we obtain state-ofthe-art performance on an audio-visual association task without the benefit of labeled training data.

Cite this Paper

BibTeX
@InProceedings{pmlr-v2-siracusa07a, title = {Dynamic Factorization Tests: Applications to Multi-modal Data Association}, author = {Siracusa, Michael R. and III, John W. Fisher}, booktitle = {Proceedings of the Eleventh International Conference on Artificial Intelligence and Statistics}, pages = {508--515}, year = {2007}, editor = {Meila, Marina and Shen, Xiaotong}, volume = {2}, series = {Proceedings of Machine Learning Research}, address = {San Juan, Puerto Rico}, month = {21--24 Mar}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v2/siracusa07a/siracusa07a.pdf}, url = {https://proceedings.mlr.press/v2/siracusa07a.html}, abstract = {The goal of a dynamic dependency test is to correctly label the interaction of multiple observed data streams and to describe how this interaction evolves over time. To this end, we propose the use of a hidden factorization Markov model (HFactMM) in which a hidden state indexes into a finite set of possible dependence structures on observations. We show that a dynamic dependency test using an HFactMM takes advantage of both structural and parametric changes associated with changes in interaction. This is contrasted both theoretically and empirically with standard sliding window based dependence analysis. Using this model we obtain state-ofthe-art performance on an audio-visual association task without the benefit of labeled training data.} }
Endnote
%0 Conference Paper %T Dynamic Factorization Tests: Applications to Multi-modal Data Association %A Michael R. Siracusa %A John W. Fisher III %B Proceedings of the Eleventh International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2007 %E Marina Meila %E Xiaotong Shen %F pmlr-v2-siracusa07a %I PMLR %P 508--515 %U https://proceedings.mlr.press/v2/siracusa07a.html %V 2 %X The goal of a dynamic dependency test is to correctly label the interaction of multiple observed data streams and to describe how this interaction evolves over time. To this end, we propose the use of a hidden factorization Markov model (HFactMM) in which a hidden state indexes into a finite set of possible dependence structures on observations. We show that a dynamic dependency test using an HFactMM takes advantage of both structural and parametric changes associated with changes in interaction. This is contrasted both theoretically and empirically with standard sliding window based dependence analysis. Using this model we obtain state-ofthe-art performance on an audio-visual association task without the benefit of labeled training data.
RIS
TY - CPAPER TI - Dynamic Factorization Tests: Applications to Multi-modal Data Association AU - Michael R. Siracusa AU - John W. Fisher III BT - Proceedings of the Eleventh International Conference on Artificial Intelligence and Statistics DA - 2007/03/11 ED - Marina Meila ED - Xiaotong Shen ID - pmlr-v2-siracusa07a PB - PMLR DP - Proceedings of Machine Learning Research VL - 2 SP - 508 EP - 515 L1 - http://proceedings.mlr.press/v2/siracusa07a/siracusa07a.pdf UR - https://proceedings.mlr.press/v2/siracusa07a.html AB - The goal of a dynamic dependency test is to correctly label the interaction of multiple observed data streams and to describe how this interaction evolves over time. To this end, we propose the use of a hidden factorization Markov model (HFactMM) in which a hidden state indexes into a finite set of possible dependence structures on observations. We show that a dynamic dependency test using an HFactMM takes advantage of both structural and parametric changes associated with changes in interaction. This is contrasted both theoretically and empirically with standard sliding window based dependence analysis. Using this model we obtain state-ofthe-art performance on an audio-visual association task without the benefit of labeled training data. ER -
APA
Siracusa, M.R. & III, J.W.F.. (2007). Dynamic Factorization Tests: Applications to Multi-modal Data Association. Proceedings of the Eleventh International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 2:508-515 Available from https://proceedings.mlr.press/v2/siracusa07a.html.

Related Material