Working with Uncertainty Workshop: Representation, Quantification, Propagation, Visualization, and Communication
Harbor at Providence, Rhode Island

What is not surrounded by uncertainty cannot be the truth.
- Richard Feynman

Knowledge is an unending adventure at the edge of uncertainty.
- Jacob Bronowski


Important Dates
  • Papers Due: September 2, 2011
  • Notification of Papers: September 16, 2011
  • Final Papers Due: September 30, 2011
  • Revised Paper for IJUQ Due: December 31, 2011

At IEEE VisWeek 2011, Providence, Rhode Island
Monday, October 24 from 8:30 a.m. - 5:55 p.m.

Uncertainty is a certainty when working with real world data. This is true whether one is grappling with exascale data sets or squeezing out information from every bit of sparse data. This is also true whether one is working with spatio-temporal data or with multivariate relational data. Tools, techniques and methodologies are needed in every facet of dealing with uncertainty from representation, quantification, propagation, and visualization. The domain of expertise and applications that have a stake in addressing uncertainty is not limited to the visualization community. In fact, if we examine the ``uncertainty pipeline'', we need to consider how uncertainty is represented e.g. as a scalar quantity such as standard deviation, as a pair of scalar quantities such as min/max range, as a multivariate representing higher order statistics, or perhaps use the data distribution itself. Then, we need to consider how uncertainties are quantified e.g. is it via fuzzy logic, evidence theory, Bayesian methods, polynomial chaos theory, etc.? An important aspect in working with uncertainty is not only in identifying and quantifying the different types and sources of uncertainty, but also in tracking how those quantities propagate in numerical simulations or tree/graph based reasoning. A basic problem is how to do arithmetic operations on variables and quantities that have an associated uncertainty. Techniques can range from incorporating uncertainty into numerical models e.g. stochastic PDE's to reasoning about uncertainty in belief networks. Then, to make sense out of all these, uncertainty visualization plays a key role in depicting both data and their associated uncertainty in a clear, unbiased fashion. Depending on who the target audience is, the visualization task may also extend to risk communication e.g. for health concerns, for severe weather warnings, etc.

This workshop will bring together researchers and practitioners from different fields who have a strong interest for the proper treatment of uncertainty. It will provide a venue for describing and identifying open problems, current best practices, and discussions on challenges and long term directions.