CaltechAUTHORS
  A Caltech Library Service

Context-Aware Deep Sequence Learning with Multi-View Factor Pooling for Time Series Classification

Das Bhattacharjee, Sreyasee and Tolone, William J. and Elshambakey, Mohammed and Cho, Isaac and Mahabal, Ashish and Djorgovski, George (2018) Context-Aware Deep Sequence Learning with Multi-View Factor Pooling for Time Series Classification. In: 2018 IEEE International Conference on Big Data (Big Data). IEEE , Piscataway, NJ, pp. 959-966. ISBN 9781538650356. https://resolver.caltech.edu/CaltechAUTHORS:20190131-131445274

[img] PDF - Published Version
See Usage Policy.

1195Kb

Use this Persistent URL to link to this item: https://resolver.caltech.edu/CaltechAUTHORS:20190131-131445274

Abstract

In this paper, we propose an effective, multi-view, multivariate deep classification model for time-series data. Multi-view methods show promise in their ability to learn correlation and exclusivity properties across different independent information resources. However, most current multi-view integration schemes employ only a linear model and, therefore, do not extensively utilize the relationships observed across different view-specific representations. Moreover, the majority of these methods rely exclusively on sophisticated, handcrafted features to capture local data patterns and, thus, depend heavily on large collections of labeled data. The multi-view, multivariate deep classification model for time-series data proposed in this paper makes important contributions to address these limitations. The proposed model derives a LSTM-based, deep feature descriptor to model both the view-specific data characteristics and cross-view interaction in an integrated deep architecture while driving the learning phase in a data-driven manner. The proposed model employs a compact context descriptor to exploit view-specific affinity information to design a more insightful context representation. Finally, the model uses a multi-view factor-pooling scheme for a context-driven attention learning strategy to weigh the most relevant feature dimensions while eliminating noise from the resulting fused descriptor. As shown by experiments, compared to the existing multi-view methods, the proposed multi-view deep sequential learning approach improves classification performance by roughly 4% in the UCI multi-view activity recognition dataset, while also showing significantly robust generalized representation capacity against its single-view counterparts, in classifying several large-scale multi-view light curve collections.


Item Type:Book Section
Related URLs:
URLURL TypeDescription
https://doi.org/10.1109/bigdata.2018.8622253DOIArticle
ORCID:
AuthorORCID
Mahabal, Ashish0000-0003-2242-0244
Djorgovski, George0000-0002-0603-3087
Additional Information:© 2018 IEEE. Funding for this research was provided by the National Science Foundations (NSF) Data Infrastructure Building Blocks (DIBBs) Progam under award #1640818.
Funders:
Funding AgencyGrant Number
NSFOAC-1640818
Subject Keywords:LSTM, RNN, Multi-view Classification, Deep Learning, Time-Series Data, Bilinear Pooling, Matrix Factorization
Record Number:CaltechAUTHORS:20190131-131445274
Persistent URL:https://resolver.caltech.edu/CaltechAUTHORS:20190131-131445274
Official Citation:S. Das Bhattacharjee, W. J. Tolone, M. Elshambakey, I. Cho, A. Mahabal and G. Djorgovski, "Context-Aware Deep Sequence Learning with Multi-View Factor Pooling for Time Series Classification," 2018 IEEE International Conference on Big Data (Big Data), Seattle, WA, USA, 2018, pp. 959-966. doi: 10.1109/BigData.2018.8622253
Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:92548
Collection:CaltechAUTHORS
Deposited By: George Porter
Deposited On:31 Jan 2019 23:30
Last Modified:03 Oct 2019 20:45

Repository Staff Only: item control page