VisLunch/Fall2008

From VistrailsWiki
Jump to navigation Jump to search


Coming up next Friday (10/03)

VIS 2008 papers Session I

Tomorrow (Friday 10/03, 12pm) in our VisLunch session
we are going to see Carlos Scheidegger and David Koop
presenting their VIS 2008 papers:

Revisiting Histograms and Isosurface Statistics
Carlos Scheidegger, John Schreiner, Brian Duffy, Hamish Carr, Claudio Silva
http://www.sci.utah.edu/~cscheid/vis2008/histograms

Abstract
Recent results have shown a link between geometric properties of isosurfaces and statistical properties of the underlying sampled data. However, this has two defects: not all of the properties described converge to the same solution, and the statistics computed are not always invariant under isosurface-preserving transformations. We apply Federer's Coarea Formula from geometric measure theory to explain these discrepancies. We describe an improved substitute for histograms based on weighting with the inverse gradient magnitude, develop a statistical model that is invariant under isosurface-preserving transformations, and argue that this provides a consistent method for algorithm evaluation across multiple datasets based on histogram equalization. We use our corrected formulation to reevaluate recent results on average isosurface complexity, and show evidence that noise is one cause of the discrepancy between the expected figure and the observed one.

VisComplete: Automating Suggestions for Visualization Pipelines
David Koop, Carlos E. Scheidegger, Steven P. Callahan, Juliana Freire, Claudio T. Silva

Abstract
Building visualization and analysis pipelines is a large hurdle in the adoption of visualization and workflow systems by domain scientists. In this paper, we propose techniques to help users construct pipelines by consensus—automatically suggesting completions based on a database of previously created pipelines. In particular, we compute correspondences between existing pipeline subgraphs from the database, and use these to predict sets of likely pipeline additions to a given partial pipeline. By presenting these predictions in a carefully designed interface, users can create visualizations and other data products more efficiently because they can augment their normal work patterns with the suggested completions. We present an implementation of our technique in a publicly-available, open-source scientific workflow system and demonstrate efficiency gains in real-world situations.

Sessions

08/29: Open Discussion and Semester Planning

Lauro's suggestion:

Seeing in Four Dimensions

The article http://www.sciencenews.org/view/generic/id/35740/title/Seeing_in_four_dimensions

The site http://www.dimensions-math.org/

Valerio Pascucci talked about his work of computing Reeb graphs.

The original plan for the session:

Our plan to this Friday's session is to have an open
discussion of any interesting ideas that you want to
bring in and to plan/schedule presentations for the
following Fridays.

A common practice for VisLunch is to use some of its
sessions as a mean to let people know about the work
of the new people around: new faculties, new post docs,
new PhD. students . As there are lots of new faces
around, we hope to schedule some of these presentations
in this session.

This semester Lauro Lins will be responsible for
organizing the VisLunch sessions. Here is his contact
informations:

Lauro Lins
Room: 4887
Phone: 581-8061
vislunch@sci.utah.edu

And here is the wikipage where we want to keep
all the information related to this semesters'
VisLunch:

http://www.vistrails.org/index.php/VisLunch/Fall2008

09/05 - Summer Internships One

In this Vislunch session we are going to see what some
PhD. Students did in their summer internships.
This week we will have Carson Brownlee,
Erik Anderson, Mark Kim and David Koop
talking about...

Carson Brownlee (15 min)
Worked at LANL for the summer with Patrick McCormick
on an analytic visualization tool called SCOUT writing
volume renderers and animation systems.
He did do some work on visualizing cosmological
datasets (dark halos) and worked on simulating
Sclieren visualization which is an old optical
technique used to look at invisible differences
in inhomogeneous data (such as shockwaves or
heat dissipation).

Erik Anderson (15 min)
Worked at Los Alamos National Lab (LANL) with
Jim Ahrens and Patrick McCormick. Iplemented a
method for discovering correlations in extremely
high dimensional datasets such as Oceanographic
data (3200 x 2400 x 42 x 114 x 52 (x,y,z,d,time)).
Additionally researched and implemented various
metrics to assist in distance visualization
prioritization and transmission.

Mark Kim (15 min)
Worked at LANL under Pat McCormick on a vis tool
called Scout (just like Carson). He did some mundane
stuff like a volume renderer and spent some time
visualizing lyman alpha lines. He also wrote a
particle-mesh n-body simulator for Scout and in CUDA.

David Koop (15 min.)
Worked at Microsoft Research on a workflow tool called
Trident (http://www.microsoft.com/mscorp/tc/trident.mspx).
Specifically, he worked on linking workflows with their
data inputs, and supporting general computational commands
like "process this data" or "re-run this workflow with
the latest data" using provenance, metadata, and semantics.
This involved digging into Windows Workflow Foundation
(http://msdn.microsoft.com/en-us/netframework/aa663328.aspx)
which Trident is based on, and figuring out how to extract
necessary metadata, inject input data, and capture
provenance information.

09/12 - Summer Internships Two

In this Vislunch session we are going to see what some
PhD. Students did in their summer internships.
This week we will have Emanuele Santos,
Kristi Potter and Abe Stephens
talking about...

Emanuele Santos (15 min)
"I worked on manipulating collections of workflows. We
(joint work with Lauro, Claudio, Juliana and Jim Ahrens)
wrote a paper in which we propose different similarity
measures for comparing workflows. I also worked on
"Workflow Medleys", which is a framework to manipulate and
interact with collections of workflows that can be used to
generate user-driven visualizations. I also helped on
extending VisTrails and other tools to facilitate the creation
of documents (papers, presentations and web pages) containing
visualizations."

Kristi Potter (15 min)
"My internship this summer was at Sandia national Labs in
Albuquerque, NM working with Andy Wilson. Our work focused
on the visualization of ensemble data in VTK. This type of data
consists of a collection of models run using various input
perturbations with the goal of getting a more accurate estimate
of the modeled phenomenon. Over the summer, we worked with
short-term weather forecast data. We have created a prototype
system to investigate methods for displaying this data, and
understanding the uncertainty and confidence levels associated
with the data."

Abe Stephens (30 min)
This year at SIGGRAPH NVIDIA demonstrated an interactive
ray tracer running on the GPU at HD resolution. The example
has practical implications for both the graphics and the
GPU computing communities. This talk will offer some
observations about these implications, the state of parallel
programming in CUDA, and discuss hybrid tracing and
rasterization. Abe Stephens is graduate student finishing
his PhD at SCI with Steven Parker. He spent part of the
spring and summer collaborating with NVIDIA Research in
Santa Clara and Salt Lake City.

09/19 - Poisson Editing: Past, Present, and Future

Friday 09/19, 12pm, in our VisLunch session
we are going to see Michael Bellem and Brian Summa
presenting:

"Poisson Editing: Past, Present, and Future

Operating in the gradient-domain has become a standard
method in image processing with applications in areas
such as image compositing and tone-mapping.

Tomorrow we will cover the general Poisson method for
editing images in the gradient domain. Also, we will
discuss the problems with this method when dealing
with gigapixel sized images. We will look at how to
extend and accelerate this method using quadtrees,
and the most recent work in Poisson editing: streaming
multigrid."

09/26 - Robust analysis of Vector Field Topology

Tomorrow (Friday 09/26, 12pm) in our VisLunch session
we are going to see Harsh Bhatia and Shreeraj Jadhav
presenting:

"Robust analysis of Vector Field Topology,

We will be speaking about the combinatorial ways of analyzing topology
of vector fields. More specifically, we will study the topology in terms of
directed graphs called Morse Connection Graphs (MCG), &
Entity Connection Graphs (ECG). We will do their comparative study,
and discuss their robustness. We will also discuss efficient algorithms to
generate the graphs.

The algorithms are suggested by Guoning Chen, Konstantin Mischaikow,
Robert S. Laramee, and Eugene Zhang in 'Efficient Morse Decompositions of Vector Fields',
published at IEEE TRANSACTIONS ON VISUALIZATION AND COMPUTER GRAPHICS,
VOL. 14, NO. 4, JULY/AUGUST 2008."

10/03 - Vis Conference: Carlos, Dave

10/10 - Vis Conference: Jens Jeff

10/17 - Fall Break (no lunch)

10/24 - VisWeek (no lunch)

10/31 - Vis Highlights

Mail List: vislunch@sci.utah.edu

People that are currently in the vislunch mail list (29 total):

Aaron Knoll              Emanuele Santos       Linh Ha                  
Adam Bargteil            Erik Anderson         Mark Kim                 
Berger                   Geoff Draper          Mathias Schott           
Carlos Scheidegger       Hao Wang              Siddarth Shankar         
Carson Browlee           Huy Vo                Steve Callahan           
Charles Hansen           Jens Kruger           Tiago Etiene             
Chems Toutai
Chris Johnson            Jianrong Shu          Valerio Pascucci         
Claudio Silva            John Schreiner        Wan Yong                 
Claurissa Tuttle         Josh Stratton    
David Koop               Lauro Lins