Automated Learning Group

Group Leaders: Peter Cheeseman and David Wolpert

The goal of the Automated Learning Group (previously known as the Bayesian Model-Based Learning or Bayes Group) is to develop the basic theory and associated algorithms for various kinds of general data analysis techniques. The Automated Learning Group is a research and development effort within the Computational Sciences Division at NASA Ames Research Center.

Current Projects

Collective Intelligence (COIN)	Analyzing the science of large, inherently distributed artificial systems and implementing local strategies in such systems to augment their performance. (list of recent publications)
3-D Surface Modeling	A new approach to generation of sub-pixel resolution composites from multiple images with different alignments.

Previous Projects

AutoClass	A package for unsupervised classification using Bayesian methods.
EDIFIS: Engine Diagnostic Filter System	An automated system which determines the readiness-for-use of a rocket engine through exhaust plume analysis.
NP-complete problems	Predicting which problems are hard to solve.
Super-Resolution 2-D Surface Modeling	Generation of sub-pixel resolution composites from multiple nearly-aligned images.
Automated Data Analysis	Supervised classification and prediction using decision trees.
Model-Based Learning	A general-purpose toolkit for data analysis.

Resources

Mars Atlas & Viking Orbiter Image-finder -- a WWW browsable, zoomable and scrollable atlas of Mars, showing the locations (footprints) of thousands of high-resolution Viking Orbiter images.

Association for Uncertainty in Artificial Intelligence (AUAI) homepage

Probability Theory as Extended Logic (Larry Bretthorst)

International Society for Bayesian Analysis (ISBA)

A Bayesian Analysis E-Print Archive

Group Members

Automated Software Engineering and Data Analysis

Computational Science Division
Information Systems Directorate

Center of Excellence in Information Systems

NASA Ames Research Center

NASA Headquarters

Comments and Suggestions to: Will Taylor

Responsible NASA Official: Will Taylor


Computational Science Division	Information Systems Directorate	Center of Excellence in Information Systems	NASA Ames Research Center	NASA Headquarters

Automated Learning Group

Group Leaders: Peter Cheeseman and David Wolpert

Current Projects

Analyzing the science of large, inherently distributed artificial systems and implementing local strategies in such systems to augment their performance. (list of recent publications)

A new approach to generation of sub-pixel resolution composites from multiple images with different alignments.

Previous Projects

A package for unsupervised classification using Bayesian methods.

An automated system which determines the readiness-for-use of a rocket engine through exhaust plume analysis.

Predicting which problems are hard to solve.

Generation of sub-pixel resolution composites from multiple nearly-aligned images.

Supervised classification and prediction using decision trees.

A general-purpose toolkit for data analysis.

Resources

Mars Atlas & Viking Orbiter Image-finder -- a WWW browsable, zoomable and scrollable atlas of Mars, showing the locations (footprints) of thousands of high-resolution Viking Orbiter images.