Announcing a new Meetup for Bay Area Artificial Intelligence Meetup Group!
What: Normative Artificial Intelligence Requirements
When: Sunday, May 30, 2010 12:00 PM
120 Independence Dr.
Menlo Park, CA 94025
This is an extra MeetUp in addition to our regular second and fourth Sunday patterns. Bob Kirby will give a presentation with the above title and the abstract below. No video recording will be made.
Normative Artificial Intelligence proposes that computers should help people deal with everyday problems rather than explain human behavior. This presentation gives background and motivation for a reductionist approach to normative Artificial Intelligence. Five top-level issues (unboundedness, contexts, uncertainty, importance, and efficiency) are identified along with technologies that may address the issues (reductionism, high-order logic, probability, utility, and parallelism). Several example use cases come from existing and proposed uses of normative Artificial Intelligence. The relationship of the technologies to the use cases guides requirements. The technologies needed for the use cases suggest development dependencies.
This presentation is a sanity test for development.
Bob Kirby received his masters and Ph.D. degrees in Computer Science from the University of Maryland, College Park. For the masters degree, he implemented and distributed a Lisp interpreter and compiler. He designed, led implementation of, and distributed the first comprehensive image processing package for Unix. His Ph.D. thesis established a probabilistic basis for more efficient labeling with Waltz relaxation, a parallel processing technique that preceded simulated annealing, current neural nets, and genetic algorithms.
He has used Expert System and related database technologies to address several problems. He designed and led implementation of a Expert System using probability and Decision Analysis to model military force structures identified from features of Synthetic-aperture radar (SAR) imagery. He used KEE to identify inconsistencies in map data. He coordinated some Postgres development. He designed and led implementation of a data fusion display of text processing results, spatial data, and SQL database records. He used Jess (http://www.jessrules.com/
) to develop a tool (http://www.cisco.com/go/pix
) to convert firewall configurations. He worked as a Computer Scientist and Software Engineer at Advanced Decision Systems, which Booz Allen Hamilton acquired; ESL, a subsidiary of TRW, which Northrup Grumman acquired; and Cisco Systems. He helped design and led implementation of security appliance portions of Cisco Monitoring Director (http://www.cisco.com/go/cmd)
. He helped develop the ASDM (http://www.cisco.com/go/asdm
) and PDM device managers for Cisco security products. He develops in Lisp, Java, C++, and with other tools and languages. He taught and mentored Lisp and object-oriented software development in C++.
Since leaving Cisco, Bob Kirby has learned Hadoop/MapReduce and demonstrated skills in Ant and JUnit while solving one of the hardest Facebook puzzles, "Facebull" (http://www.facebook.com/careers/puzzles.php?puzzle_id=1)
. He continues to maintain a homework project in Knowledge Representation of natural language semantics.http://www.linkedin.com/in/bobkirby
Learn more here:http://www.ai-meetup.org/calendar/13550266/