DR. TIM FREEMAN

The paper Using Compassion and Respect to Motivate an Artificial Intelligence said

This paper presents a decision procedure that can, in principle, observe people's behavior and from that infer what they want. Training the algorithm would require giving it past observations of the world coupled with an estimate of the perceptions and behaviors of the people that were being observed.

The output from the algorithm is a set of explanations of the observed phenomena; each explanation has an a-priori probability and enough information to infer an expected utility for each person. These expected utilities can be combined with a simple arithmetic to get a total utility function that could be used as input to a planner, if unlimited computing power were available. A different arithmetic would give rise to plans that plausibly fit the labels "compassionate" and "respectful".

Python source code is provided. This builds on past work with inductive inference by Marcus Hutter, Jürgen Schmidhuber, and Ray Solomonoff.

Tim Freeman, Ph.D. was the author of this paper and is based at Hewlett Packard. He designed and developed much of the client and server sides of HP Guide as a Python and Apache application on Linux and Windows XP, and ported a suitable piece of it to Vista for inclusion in HP Advisor and distribution on Spring 2007 HP commercial PCs.

He previously worked at Elemental Security, San Mateo, CA where he wrote Java code that functioned in a Tomcat server running on Linux, with clients in Python and an Oracle 9 database. Source control was done with Perforce and bugs were tracked with Bugzilla.

Tim developed Fungimol, an extensible system for designing atomic-scale objects. The intent is to eventually extend it to be a useful system for doing molecular nanotechnology design work. He coauthored DAGWOOD: A System for Manipulating Polynomial Given by Straight-Line Programs.

Tim earned his Ph.D. in Computer Science at Carnegie Mellon University, Pittsburgh, PA in 1994. His thesis topic was Refinement Types, a customizable type system for ML, a functional programming language. He earned his M.S. in Computer Science at Rensselaer Polytechnic Institute, Troy, NY in 1985. He earned his B.S. (Magna Cum Laude) in Mathematics, Physics, and Computer Science at Mary Washington College, Fredericksburg, VA in 1984.

Print bio!