Homepage for Russell G. Almond

Russell Almond is a statistician with extensive training and interest in artificial intelligence and human--computer interaction. His research interests include:

  • Representations of uncertainty in artificial intelligence (especially belief functions and graphical models),
  • Uses of artificial intelligence techniques in data analysis,
  • Human factor design for data analysis and education, and
  • Risk analysis.

Education and Employment

Russell received his undergraduate degree from Caltech in Mathematics in 1983 and received his Ph.D. from Harvard University in Statistics in 1990.

Russell is currently working for the Research Division of Educational Testing Service in the Research Statistics Group. (ETS is still establishing a Web presence. These links may not be available to people outside ETS yet.)

Before coming to ETS he was a research scientist at StatSci. and a visitor at the UW statistics department. He continues to be involved in collaborative research with David Madigan.

Russell’s complete c.v. is available on line, including links to downloadable technical reports and preprints.

Graphical-Belief Project

Russell Almond’s most recent research has been centered around using graphical models (related to influence diagrams and Bayesian networks) to model risk using both probability and belief functions as the principle representation of uncertainty. He is the principle designer of the Graphical-Belief software for manipulating these models. He has also recently published a book on these models:

Russell maintains a listing of software for manipulating probabilistic and belief function graphical models (WWW Page.)

Russell is on the program comittee for the 6th International Workshop on Artificial Intelligence and Statistics. This is usally a lot of fun and very informatative, so check it out.

Software Packages

Russell is responsible for designing and coding the following software:

BELIEF

Program to manipulate graphical belief function models written in Common Lisp. It is available through the: CMU Artificial Intelligence Library

ElToY

Program for eliciting Bayesian conjugate prior distributions through constraint based dynamic graphics. Written in XLISP-STAT. Copies available via statlib: Instructions , Shar Archive. [Note there is a small bug in the XLISP-STAT 3.44 which causes ElToy to fail. This should be fixed with the next release of XLISP-STAT.]

GRAPHICAL-BELIEF

Revised and extended version of BELIEF with many new capabilities and graphical user interface. Currently only available as a prototype system. To find out more about GRAPHICAL-BELIEF, visit the GRAPHICAL-BELIEF home page.

Affiliations

Russell has actively participated in several organizations related to reasoning with and about uncertainty including:

Russell has also been developing with the Garnet user interface development environment. Russell is currently a member of:

Other Interests

Russell is an active early musician and plays the curtal (a sort of early bassoon), the shawm and the recorder. He is also working on settings for various renaissance dances. You can see some of his work on his Renaissance Dance Page.

Hu!

These web pages maintained by:

almond@acm.org

Last modified: Thu Jun 7 15:53:14 2001