Scalable Representational Structures for Complex Multivariate Time Series

Abstract

Prof. Fox proposes to investigate the evolving complexities faced when performing time series analysis with multiple data sources that introduce measures of uncertainty, integration across numerous heterogeneous data sources, filling in for missing data due to data sensor dropouts, and several other processing challenges. Several specific research questions to be addressed include: (a) Dimensionality - how to define models that capture intricate and possibly evolving relationships between the components of a highly multivariate time series, (b) Causality - learning causal links, especially at potentially high-order lags, (c) Mixed-Type Data - how to properly leverage generative modeling abilities to fuse information of disparate types, such as video and text with neural recordings, (d) Mixed-Frequency Data- how to fuse information from data streams recorded at different time scales (e.g., daily, weekly, quarterly, etc.).

Document Details

Document Type
DoD Grant Award
Publication Date
Mar 23, 2016
Source ID
FA95501610038

Entities

People

  • Emily B. Fox

Organizations

  • Air Force Office of Scientific Research
  • United States Air Force
  • University of Washington

Tags

Fields of Study

  • Computer science

Readers

  • Mathematics or Statistics
  • Neural Network Machine Learning.

Technology Areas

  • AI & ML