Ambiguity Modeling in Latent Spaces

Carl Henrik Ek, Jon Rihan, Philip H. S. Torr, Gregory Rogez, Neil D. Lawrence
Machine Learning for Multimodal Interaction (MLMI 2008), Springer-Verlag :62-73, 2008.

Abstract

We are interested in the situation where we have two or more representations of an underlying phenomenon. In particular we are interested in the scenario where the representation are complementary. This implies that a single individual representation is not sufficient to fully discriminate a specific instance of the underlying phenomenon, it also means that each representation is an ambiguous representation of the other complementary spaces. In this paper we present a latent variable model capable of consolidating multiple complementary representations. Our method extends canonical correlation analysis by introducing additional latent spaces that are specific to the different representations, thereby explaining the full variance of the observations. These additional spaces, explaining representation specific variance, separately model the variance in a representation ambiguous to the other. We develop a spectral algorithm for fast computation of the embeddings and a probabilistic model (based on Gaussian processes) for validation and inference. The proposed model has several potential application areas, we demonstrate its use for multi-modal regression on a benchmark human pose estimation data set.

Cite this Paper


BibTeX
@InProceedings{Ek:ambiguity08, title = {Ambiguity Modeling in Latent Spaces}, author = {Ek, Carl Henrik and Rihan, Jon and Torr, Philip H. S. and Rogez, Gregory and Lawrence, Neil D.}, booktitle = {Machine Learning for Multimodal Interaction (MLMI 2008)}, pages = {62--73}, year = {2008}, editor = {Popescu-Belis, Andrei and Stiefelhagen, Rainer}, series = {LNCS}, publisher = {Springer-Verlag}, doi = {10.1007/978-3-540-85853-9_6}, pdf = {https://inverseprobability.com/publications/files/mlmi2008.pdf}, url = {http://inverseprobability.com/publications/ek-ambiguity08.html}, abstract = {We are interested in the situation where we have two or more representations of an underlying phenomenon. In particular we are interested in the scenario where the representation are complementary. This implies that a single individual representation is not sufficient to fully discriminate a specific instance of the underlying phenomenon, it also means that each representation is an ambiguous representation of the other complementary spaces. In this paper we present a latent variable model capable of consolidating multiple complementary representations. Our method extends canonical correlation analysis by introducing additional latent spaces that are specific to the different representations, thereby explaining the full variance of the observations. These additional spaces, explaining representation specific variance, separately model the variance in a representation ambiguous to the other. We develop a spectral algorithm for fast computation of the embeddings and a probabilistic model (based on Gaussian processes) for validation and inference. The proposed model has several potential application areas, we demonstrate its use for multi-modal regression on a benchmark human pose estimation data set.} }
Endnote
%0 Conference Paper %T Ambiguity Modeling in Latent Spaces %A Carl Henrik Ek %A Jon Rihan %A Philip H. S. Torr %A Gregory Rogez %A Neil D. Lawrence %B Machine Learning for Multimodal Interaction (MLMI 2008) %C LNCS %D 2008 %E Andrei Popescu-Belis %E Rainer Stiefelhagen %F Ek:ambiguity08 %I Springer-Verlag %P 62--73 %R 10.1007/978-3-540-85853-9_6 %U http://inverseprobability.com/publications/ek-ambiguity08.html %X We are interested in the situation where we have two or more representations of an underlying phenomenon. In particular we are interested in the scenario where the representation are complementary. This implies that a single individual representation is not sufficient to fully discriminate a specific instance of the underlying phenomenon, it also means that each representation is an ambiguous representation of the other complementary spaces. In this paper we present a latent variable model capable of consolidating multiple complementary representations. Our method extends canonical correlation analysis by introducing additional latent spaces that are specific to the different representations, thereby explaining the full variance of the observations. These additional spaces, explaining representation specific variance, separately model the variance in a representation ambiguous to the other. We develop a spectral algorithm for fast computation of the embeddings and a probabilistic model (based on Gaussian processes) for validation and inference. The proposed model has several potential application areas, we demonstrate its use for multi-modal regression on a benchmark human pose estimation data set.
RIS
TY - CPAPER TI - Ambiguity Modeling in Latent Spaces AU - Carl Henrik Ek AU - Jon Rihan AU - Philip H. S. Torr AU - Gregory Rogez AU - Neil D. Lawrence BT - Machine Learning for Multimodal Interaction (MLMI 2008) DA - 2008/08/09 ED - Andrei Popescu-Belis ED - Rainer Stiefelhagen ID - Ek:ambiguity08 PB - Springer-Verlag DP - LNCS SP - 62 EP - 73 DO - 10.1007/978-3-540-85853-9_6 L1 - https://inverseprobability.com/publications/files/mlmi2008.pdf UR - http://inverseprobability.com/publications/ek-ambiguity08.html AB - We are interested in the situation where we have two or more representations of an underlying phenomenon. In particular we are interested in the scenario where the representation are complementary. This implies that a single individual representation is not sufficient to fully discriminate a specific instance of the underlying phenomenon, it also means that each representation is an ambiguous representation of the other complementary spaces. In this paper we present a latent variable model capable of consolidating multiple complementary representations. Our method extends canonical correlation analysis by introducing additional latent spaces that are specific to the different representations, thereby explaining the full variance of the observations. These additional spaces, explaining representation specific variance, separately model the variance in a representation ambiguous to the other. We develop a spectral algorithm for fast computation of the embeddings and a probabilistic model (based on Gaussian processes) for validation and inference. The proposed model has several potential application areas, we demonstrate its use for multi-modal regression on a benchmark human pose estimation data set. ER -
APA
Ek, C.H., Rihan, J., Torr, P.H.S., Rogez, G. & Lawrence, N.D.. (2008). Ambiguity Modeling in Latent Spaces. Machine Learning for Multimodal Interaction (MLMI 2008), in LNCS:62-73 doi:10.1007/978-3-540-85853-9_6 Available from http://inverseprobability.com/publications/ek-ambiguity08.html.

Related Material