Designing Intelligent Systems


All artificial systems must be created, and as such they need good design.  Learning, genetic algorithms and autonomous development are useful techniques, but they don't spontaneously create working artificial systems.   In nature, evolution took billions of years to create the genetic scaffolding essential for animals with complex intelligence.  Natural human-like intelligence further requires many years of individual and social learning.  Artificial systems have to be designed and constructed before they can begin to do any learning, evolving or developing, or to perform any other function.

Behavior Oriented Design (BOD) is a development methodology for intelligent and cognitive systems.  BOD is a modular approach that extends object-oriented design to support agency.  Basic actions are provided by a reusable modular library developed for a particular platform & domain, e.g. a set of domestic robots, or characters for a virtual game world.  The process of specifying agency can also be seen as the process of system integration around a particular individual intelligence. Both are achieved by specifying the agent's priorities using POSH dynamic plans.  These plans facilitate intelligent action selection by providing priorised arbitration wherever behaviour generated by two or more different modules might otherwise be in conflict.

The BOD methodology has been used to develop software for robots, real-time virtual reality characters, intelligent environments and experimental platforms for increasing our understanding of natural intelligence.  For a list of working systems, see my research page.  I have also done some work on applying this methodology to designing services for the semantic web, creating intelligent tutoring systems and managing ubiquitous computing in intelligent environments.  These latter projects have not yet been fully developed, but I metion them here because they give a broader notion of what an intelligent system can be.

Most of my publications concern techniques for designing or developing AI.  See also my special lists of papers for action selection and robot ethics.  The best introductions to Behavior Oriented Design are:
  • Joanna J. Bryson Behavior-Oriented Design of Modular Agent Intelligence, (HTML, there is also a pdf version), Agent Technologies, Infrastructures, Tools, and Applications for e-Services, R. Kowalszyk, J. P. Müller, H. Tianfield and R. Unland, eds., pp. 61-76, Springer, 2003.
  • Joanna J. Bryson, Tristan J. Caulfield and Jan Drugowitsch, Integrating Life-Like Action Selection into Cycle-Based Agent Simulation Environments, in Proceedings of Agent 2005: Generative Social Processes, Models, and Mechanisms, Michael North, David L. Sallach and Charles Macal eds., pp. 67-81, Argonne National Laboratory 2006.  Associated software.
  • How to Make a Monkey Do Something Smart  A brief, `fun' document about Behavior Oriented Design (BOD) taken from my 2001 PhD dissertation.
    •  See also:
  • Bidan Huang, Jason L. Leake and I, Humanoid Robots and Cognitive Systems Research: An Epistemological Case Study Based on the iCub. Presented at a poster at ICDL - EPIROB in Frankfurt in August 2011.
  • Jakub Gemrot, Cyril Brom, Joanna Bryson & Michal Bida, How to compare usability of techniques for the specification of virtual agents behavior? An experimental pilot study with human subjects, in the AAMAS 2011 Workshop on the uses of Agents for Education, Games and Simulations.
  • A Role for Consciousness in Action Selection, in Proceedings of the AISB 2011 Symposium Machine Consciousness.
  • John Grey and I, Procedural Quests: A Focus for Agent Interaction in Role-Playing-Games, in Proceedings of the AISB 2011 Symposium AI & Games.
  • Philipp Rohlfshagen and Joanna J. Bryson, Flexible Latching: A Biologically-Inspired Mechanism for Improving the Management of Homeostatic Goals in Cognitive Computation 2(3):230-241).  Associated software comes with the standard python/jython distribution of BOD.
  • TALK: AI Architectures (or State Requirements for Human-like Action Selection), plenary presented to the euCognition meeting on Cognitive Architectures, 29 June 2007. Mine was the first, overview talk -- I focus on characteristics of AI architectures that worked, and mention a few that didn't work but got famous anyway.
  • Cyril Brom, Jakub Gemrot, Michal Bída, Ondrej Burkert, Sam J. Partington and I, POSH Tools for Game Agent Development by Students and Non-Programmers, in the proceedings of  CGAMES 2006.
  • Samuel J. Partington and Joanna J. Bryson, ``The Behavior Oriented Design of an Unreal Tournament Character'', The Fifth International Working Conference on Intelligent Virtual Agents, T. Panayiotopoulos, J. Gratch, R. Aylett, D. Ballin, P. Olivier and T. Rist, eds., pp. 466-477, Springer, 2005. Associated software.
  • Bruce Edmonds and I, The Insufficiency of Formal Design Methods - the necessity of an experimental approach for the understanding and control of complex MAS.  In the proceedings of AAMAS 2004.
  • Action Selection and Individuation in Agent Based Modelling (in HTML or PDF). In the proceedings of Agent 2003: Challenges of Social Simulation, David L. Sallach and Charles Macal eds.
  • Where Should Complexity Go? Cooperation in Complex Agents with Minimal Communication (in pdf). Discusses when to use communication between agents in a multi-agent system vs. when to use behavior arbitration between modules in a modular single agent. In the proceedings of the First GSFC/JPL Workshop on Radical Agent Concepts (WRAC).
  • Modularity and Design in Reactive Intelligence, with Lynn Andrea Stein. From IJCAI 2001. A 6 page summary of my dissertation, including a description of BOD.
  • Hypothesis Testing for Complex Agents in pdf (or postscript), with Will Lowe and Lynn Andrea Stein. In the proceedings of the NIST Workshop on Performance Metrics for Intelligent Systems (2000).
  • Architectures and Idioms: Making Progress in Agent Design in pdf (or postscript). Written with Lynn Andrea Stein. Presented at ATAL 2000, now a book chapter, the final version is © Springer-Verlag.
  • Cross-Paradigm Analysis of Autonomous Agent Architecture, in the Journal of Experimental and Theoretical Artificial Intelligence (JETAI) 12(2). Summary: article about trends in agent architectures and what they imply about optimal strategies for designing intelligence.
    • See further my work on the ethics of AI systems.
    Joanna  Bryson
    Last updated November, 2011