Machine Learning for Information Visualization
This tutorial will introduce a wide variety of machine learning techniques, focusing on methods that are particularly relevant to visualization and analysis of complex and high dimensional data. Case studies of applying these methods to real-world problems will be presented and discussed. The audience will gain understanding and insight of how to apply machine learning techniques to different information visualization tasks.
The first part of the tutorial will describe basic concepts and paradigms in machine learning, laying the foundation for the rest of the tutorial. The second part will cover state-of-the-art techniques commonly used for information visualization: clustering, density estimation, classification, regression, and dimensionality reduction. The third part will discuss various application examples and practical issues. The fourth part presents a brief discussion of more advanced topics in machine learning.
» Sunday, Morning
Applying Color Theory to Visualization
We highlight the visual impact of specific color combinations and provide practical suggestions on digital color mixing for visualization. The successful application of color theory is a key component in the design of digital media for interactive visual discovery, time series animation, and other visual analytics efforts. Various artists' and scientists' theories of color and how to apply these theories to creating your own digital media work will be reviewed.
We include a hands on session that teaches you how to build and evaluate color schemes with Adobe's Kuler, Color Scheme Designer, and Color Brewer tools, each of which are available online.
Please bring various small JPEG examples of your visualizations for doing color analyses. We will also share our own personal failures and successes with applying these color theories and tools to actual visualization projects.
Before the tutorial, please consider registering at http://kuler.adobe.com/ for an account to access Adobe's Kuler tool.
» Sunday, Afternoon
Visualizing Data in R
R is an open-source statistical programming environment. It is widely used by academic statisticians and has become increasingly popular in many applied domains. In this tutorial, you'll learn about: the strengths and weaknesses of this tool that is employed in diverse fields from biology to psychology to politcal science; approaches to visualisation from a different tradition that embraces the command line interface and expects that users will have some programming knowledge; and how you can connect to R to take advantage of cutting edge statistical and machine learning models.
We'll begin with an brief introduction to the R language, continuing with a discussion of how you can use it with your existing tools. You'll also learn the basic data structures and some of the tools most important for fluent R use.
You'll learn how to create a wide variety of basic graphics using the R package ggplot2, and enhance them with aesthetics and facetting. You'll also learn a new strategy for dealing with large data.
» Monday, All day
DIY Vis Applications
|Timothy M. Shead|
Every year, researchers present many new wonderful visualization and analysis algorithms. However, many of these algorithms are not transitioned to receptive researchers in a timely manner. Algorithm developers typically build lightweight prototypes to demonstrate their ideas and research to the community, and building full-featured visualization applications is hard work. This tutorial covers some of the most popular open-source frameworks whose aim is to simplify the development and deployment of visualization algorithms to high quality software applications:
- ParaView, an open-source turnkey application for analyzing and visualizing scientific data sets.
- VTK, a very popular toolkit for building scientific visualization and informatics applications.
- Titan, extending VTK to provide analytics functionality.
- VisTrails, a scientific workflow and provenance management system.
- VisMashups, providing an easy way to deploy VisTrails workflows over the Web.
- Voreen, an interactive visualization environment.
We will focus on building desktop- and Web-based visualization applications using these open-source tools, covering a variety of application types.
» Monday, Afternoon
Large Vector-Field Visualization
The study of vector fields resulting from simulation and measurement has a rich tradition in Scientific Visualization, because of the ubiquity of vector fields across science, engineering, and medicine. While visualization techniques have in the past often focused on fluid flows, other application domains such as astrophysics, geodynamics, life sciences, and high-energy physics are gaining prominence. Across all these domains, numerical simulations produce large, time-varying, and highly complex vector fields. While the visual investigation of these four-dimensional fields creates obvious issues of depiction and perception, the exploding data size coupled with the growing significance of Lagrangian methods raises unique challenges in data management and computational scalability.
This tutorial provides an overview of modern approaches and discusses their suitability for large vector field visualization. The included topics encompass fundamentals of vector field theory and computation, as well as application examples involving large vector fields from the areas of biomedicine and aeronautics.
» Tuesday, Morning
The PhD in Visualization Starter Kit (PVSK)
|Organizer:||Robert S. Laramee|
Writing a PhD is difficult, and those just starting a PhD in visualization have not usually acquired all of the key skills necessary for completion. For example, how does a researcher navigate through the vast amounts of previously published literature related to their topic? For some, this may be their first time implementing a larger, long-term project. Developing a large software application requires more knowledge than implementing a small one. How can bugs and problems be tracked down and eliminated in a large visualization application? And what is a good starting point when writing a research paper?
We present some of the essential skills that a PhD candidate in Visualization needs during their study including (1) reading and (2) writing research papers, and (3) implementing and (4) debugging software.
» Tuesday, Afternoon
Tensors in Visualization
|Organizers:||Gordon L. Kindlmann|
|M. Alex O. Vasilescu|
Tensor fields arise in several scientific applications, such as diffusion-weighted magnetic resonance imaging and fluid flows. Organization of data into data tensors can provide a useful mathematical tool even for processing and visualizing simple scalar volume datasets. This tutorial will present a coherent and coordinated explanation of these topics with particular emphasis on topics for additional research.
The first half of the tutorial will focus on second-order tensors. An indispensible step in the visualization of tensor fields is to select a part of the data for display, in order to avoid visual clutter.
The second half will start with treating tensors of higher orders in the context of High Angular Resolution Diffusion Imaging (HARDI). We will describe the main threads of HARDI research and point out differences in their interpretation.
Finally, we turn to a larger class of volume datasets: Data tensors or multi-way arrays are often encountered when we have a collection of multivariate data which can be organized into a data tensor based on their causal factors.