Moderated and Drifting Linear Dynamical Systems

Jinyan Guan, Kyle Simek, Ernesto Brau, Clayton Morrison, Emily Butler, Kobus Barnard
Proceedings of the 32nd International Conference on Machine Learning, PMLR 37:2473-2482, 2015.

Abstract

We consider linear dynamical systems, particularly coupled linear oscillators, where the parameters represent meaningful values in a domain theory and thus learning what affects them contributes to explanation. Rather than allow perturbations of latent states, we assume that temporal variation beyond noise is explained by parameter drift, and variation across coupled systems is a function of moderating variables. This change of focus reduces opportunities for efficient inference, and we propose sampling procedures to learn and fit the models. We test our approach on a real dataset of physiological measures of heterosexual couples engaged in a conversation about a potentially emotional topic, with body mass index (BMI) being considered as a moderator. We evaluate several models on their ability to predict future conversation dynamics (the last 20% of the data for each test couple), with shared parameters being learned using held out data. As proof of concept, we validate the hypothesis that BMI affects the conversation dynamic in the experimentally chosen topic.

Cite this Paper

BibTeX
@InProceedings{pmlr-v37-guan15, title = {Moderated and Drifting Linear Dynamical Systems}, author = {Guan, Jinyan and Simek, Kyle and Brau, Ernesto and Morrison, Clayton and Butler, Emily and Barnard, Kobus}, booktitle = {Proceedings of the 32nd International Conference on Machine Learning}, pages = {2473--2482}, year = {2015}, editor = {Bach, Francis and Blei, David}, volume = {37}, series = {Proceedings of Machine Learning Research}, address = {Lille, France}, month = {07--09 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v37/guan15.pdf}, url = {https://proceedings.mlr.press/v37/guan15.html}, abstract = {We consider linear dynamical systems, particularly coupled linear oscillators, where the parameters represent meaningful values in a domain theory and thus learning what affects them contributes to explanation. Rather than allow perturbations of latent states, we assume that temporal variation beyond noise is explained by parameter drift, and variation across coupled systems is a function of moderating variables. This change of focus reduces opportunities for efficient inference, and we propose sampling procedures to learn and fit the models. We test our approach on a real dataset of physiological measures of heterosexual couples engaged in a conversation about a potentially emotional topic, with body mass index (BMI) being considered as a moderator. We evaluate several models on their ability to predict future conversation dynamics (the last 20% of the data for each test couple), with shared parameters being learned using held out data. As proof of concept, we validate the hypothesis that BMI affects the conversation dynamic in the experimentally chosen topic.} }
Endnote
%0 Conference Paper %T Moderated and Drifting Linear Dynamical Systems %A Jinyan Guan %A Kyle Simek %A Ernesto Brau %A Clayton Morrison %A Emily Butler %A Kobus Barnard %B Proceedings of the 32nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2015 %E Francis Bach %E David Blei %F pmlr-v37-guan15 %I PMLR %P 2473--2482 %U https://proceedings.mlr.press/v37/guan15.html %V 37 %X We consider linear dynamical systems, particularly coupled linear oscillators, where the parameters represent meaningful values in a domain theory and thus learning what affects them contributes to explanation. Rather than allow perturbations of latent states, we assume that temporal variation beyond noise is explained by parameter drift, and variation across coupled systems is a function of moderating variables. This change of focus reduces opportunities for efficient inference, and we propose sampling procedures to learn and fit the models. We test our approach on a real dataset of physiological measures of heterosexual couples engaged in a conversation about a potentially emotional topic, with body mass index (BMI) being considered as a moderator. We evaluate several models on their ability to predict future conversation dynamics (the last 20% of the data for each test couple), with shared parameters being learned using held out data. As proof of concept, we validate the hypothesis that BMI affects the conversation dynamic in the experimentally chosen topic.
RIS
TY - CPAPER TI - Moderated and Drifting Linear Dynamical Systems AU - Jinyan Guan AU - Kyle Simek AU - Ernesto Brau AU - Clayton Morrison AU - Emily Butler AU - Kobus Barnard BT - Proceedings of the 32nd International Conference on Machine Learning DA - 2015/06/01 ED - Francis Bach ED - David Blei ID - pmlr-v37-guan15 PB - PMLR DP - Proceedings of Machine Learning Research VL - 37 SP - 2473 EP - 2482 L1 - http://proceedings.mlr.press/v37/guan15.pdf UR - https://proceedings.mlr.press/v37/guan15.html AB - We consider linear dynamical systems, particularly coupled linear oscillators, where the parameters represent meaningful values in a domain theory and thus learning what affects them contributes to explanation. Rather than allow perturbations of latent states, we assume that temporal variation beyond noise is explained by parameter drift, and variation across coupled systems is a function of moderating variables. This change of focus reduces opportunities for efficient inference, and we propose sampling procedures to learn and fit the models. We test our approach on a real dataset of physiological measures of heterosexual couples engaged in a conversation about a potentially emotional topic, with body mass index (BMI) being considered as a moderator. We evaluate several models on their ability to predict future conversation dynamics (the last 20% of the data for each test couple), with shared parameters being learned using held out data. As proof of concept, we validate the hypothesis that BMI affects the conversation dynamic in the experimentally chosen topic. ER -
APA
Guan, J., Simek, K., Brau, E., Morrison, C., Butler, E. & Barnard, K.. (2015). Moderated and Drifting Linear Dynamical Systems. Proceedings of the 32nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 37:2473-2482 Available from https://proceedings.mlr.press/v37/guan15.html.

Related Material