Kiri L. Wagstaff

My background is in Computer Science, Planetary Science, and Geology. I am most interested in problems that lie at the interfaces between these fields, such as automated methods to investigate science questions using planetary data.

I am a researcher at the Jet Propulsion Laboratory in Pasadena, CA. After receiving my Ph.D. in Computer Science from Cornell University in 2002, I worked for a year in the JHU Applied Physics Lab's Space Department. My two major projects dealt with space weather prediction and the fault protection system for the MESSENGER spacecraft. I am now a member of JPL's Machine Learning and Instrument Autonomy Group, investigating ways that machine learning can be used to increase the autonomy of space missions. I also teach classes at Cal State L.A. in the Computer Science Department. My projects at JPL have included:

Tracking the north polar ice caps (water and CO₂) on Mars
Automatic landmark identification and change detection in Mars orbital images (dark slope streaks, dust devil tracks, etc.)
Predicting county-level crop yield from Earth orbital images
Modeling user preferences for sets, rather than individual items (like music playlists or rover image downlink sets)
Modeling flight software with state charts and using automatic code generation to convert them into C/C++ (for implementation) or Promela (for model checking)
Analyzing the sensitivity of machine learning algorithms to high-radiation environments

News and upcoming events:

I was selected to receive the 2008 Lew Allen Award for Excellence in Research.
I will receive my M.S. in Geological Sciences from USC in December, 2008. My thesis is titled "Biogenicity Analysis for Stromatolite Structures." Here is the abstract (also shown as a wordle) and the full text from the USC library:

The origin of stromatolites (laminated, dome-shaped geologic structures thought to constitute some of the oldest fossils on Earth) is controversial. Stromatolites are now known to form via biologic and abiologic processes. Separating the two can improve our understanding of the evolution of life on Earth (and potentially other planets). Four information-theoretic image features (gzip compression, PNG compression, entropy, and Wold energy ratio) are proposed to help differentiate biogenic from abiotic samples. A biogenic separability metric is defined to permit a quantitative assessment of these features. The most effective feature was the PNG compression ratio, which provided good biogenic separability for macroscopic and microscopic views of stromatolites. Increasing the size of the analysis sub-window and/or increasing magnification also improved separability. While conclusive judgments about biogenicity are unlikely to be made solely from these features, they can provide a "first cut" estimate of the importance of a follow-up search for other biosignatures.
I will be presenting a paper on our method for predicting county-level crop yield from remote sensing data at the 4th International Workshop on Mining Complex Data in Pisa, Italy, on December 14, 2008:
- "Multiple-Instance Regression with Structured Data," by Kiri L. Wagstaff, Terran Lane, and Alex Roper