Summer 2013: Modern Statistical and Computational Methods for Analysis of Kepler Data: June 10-28, 2013

Location

This was held at SAMSI in Research Triangle Park, NC.

Description

This three week mini-research program focused on statistical and computational challenges associated with analyzing exoplanet data from NASA’s Kepler mission. The first day (Monday, June 10, 2013) consisted of invited talks (~20-30 minutes each + questions) chosen: 1) to highlight key problems that can be addressed using Kepler data, 2) to help participants understand the nature of Kepler data, and 3) to provide an introduction to selected statistical methods that are likely to be applied during the program.

On the morning of the second day (Tuesday, June 11, 2013), all participants who did not give an invited talk introduced themselves (~5-10 minutes + questions) to the group. Astronomers were encouraged to identify statistical challenges that they hope to make progress on during their visit and could become a focus for one of the working groups during this program. Statisticians were encouraged to identify tools that they think could be applied to one or more of the problems identified on the first day. Participants organized themselves into three working groups during the afternoon of the second day.

For the remainder of the mini-research program, each participant joined one or two of three working groups for intensive research collaboration among astronomers and statisticians. Most of the participant’s time was devoted to collaborative research. Each focus group met for ~1-1.5 hours daily to discuss progress and challenges.

On the final day (Friday, June 28, 2013), there was a one day event for program participants to present their results, as well as make plans for continued investigation or collaboration beyond the SAMSI mini-research program.

Working Groups

Three working groups:

  1. object detection and validation (e.g., searching the Kepler data for planets, moons, binary stars and/or other interesting astrophysical objects in the presence of measurement noise, instrumental systematics and other astrophysical signals; model comparison to establish that signals are due to planets rather than an astrophysical false positive),
  2. characterizing exoplanets and/or binary stars (e.g., efficient posterior sampling for measuring masses, orbits and their uncertainties using transit timing variations; Bayesian model comparison to quantify evidence for non-transiting planets), and
  3. population statistics (e.g., making inferences about the distribution of extrasolar planets along with their physical and orbital properties; likelihood-free methods and approximate Bayesian computing for population analyses with many model parameters).

Schedule and Supporting Media

Monday, June 10, 2013
at SAMSI, NISS Building, Room 104

Description Speaker Slides Videos
Welcome and Workshop Overview Eric Ford, University of Florida
Overview of Kepler Mission Thomas Barclay, NASA
Long Period Planets from the Kepler Data Matthew Payne, CfA
Searching for Planet Candidates in Kepler Data Thomas Barclay, NASA
Behind the Scenes of Kepler Eclipsing Binary Science Andrej Prsa, Villanova University
Validation of Kepler Planet Candidates Tim Morton, Caltech  
Data Mining for Astrophysical Transients Joseph Richards, UC-Berkeley
Characterizing Planets via Transit and/or Eclipse Timing Variations Joshua Carter, CfA
Characterizing the Distribution of Planetary Architectures with Kepler Eric Ford, University of Florida
Understanding New Systems of Multiple Planets from the Kepler Space Telescope Darin Ragozzine, University of Florida
Graphical Models on Graphics Processors Thomas Loredo, Cornell University
Hierarchical Inference for Characterizing Populations David Hogg, New York University
Efficient Bayesian Computation Paul Baines, UC-Davis

Tuesday, June 11, 2013
at SAMSI, NISS Building, Room 104

Description Speaker Slides Videos
Bayesian Adjustment for Multiplicity Jim Berger, Duke University
Model Selection, Interim Priors and Hierarchical Models Merlise Clyde, Duke University
Characterizing the Completeness of the Kepler Planet Candidate List Jessie Christiansen, NASA Ames
LARK: One Way to Quantify Uncertainty about Light Curves Robert Wolpert, Duke University
Approximate Bayesian Computing Jessi Cisewski, Carnegie Mellon University
Planet Hunters: Searching for Exoplanets with 500,000 Eyes Meg Schwamb, Yale University
Working Group Organization

Wednesday, June 12 – Friday, June 14, 2013
at SAMSI, NISS Building, Room 104

Description Speaker Slides Videos
Working Group #1
Working Group #2
Working Group #3
Working Group #4

Monday, June 17 – Friday, June 21, 2013
at SAMSI, NISS Building, Room 104

Description Speaker Slides Videos
Working Group #1
11:00-12:30 Working Group #2
Working Group #3
Working Group #4

Monday, June 23 – Thursday, June 27, 2013
at SAMSI, NISS Building, Room 104

Description Speaker Slides Videos
Working Group #1
Working Group #2
Working Group #3
Working Group #4

Friday, June 28, 2013
at SAMSI, NISS Building, Room 104

Description Speaker Slides Videos
Working Group Reports
Working Group Reports
Planning Session for Continuing Collaborations
Farewell