Building Softbots for UNIX (Preliminary Report)

AI is moving away from "toy tasks" such as block stacking towards real-world problems. This trend is positive, but the amount of preliminary groundwork required to tackle a real-world task can be staggering, particularly when developing an integrated agent architecture. To address this problem, we advocate real-world software environments, such as operating systems or databases, as domains for agent research. The cost, effort, and expertise required to develop and experiment with software agents is relatively low. Furthermore, software environments circumvent many thorny, but peripheral, research issues that are inescapable in other environments. Thus, software environments enable us to test agents in a real world yet focus on core AI research issues. To support this claim, we describe our project to develop UNIX 1 softbots (software robots)--complete intelligent agents that interact with UNIX. Our fully-implemented softbot is able to accept a diverse set of high-level goals, generate and execute plans to achieve these goals in real time, and recover from errors when necessary.

[1]  Pattie Maes,et al.  Learning Interface Agents , 1993, AAAI.

[2]  Oren Etzioni,et al.  Planning with Incomplete Information in the UNIX Domain , 1993 .

[3]  Allen Newell,et al.  Learning 10, 000 Chunks: What's It Like Out There? , 1992, AAAI.

[4]  Oren Etzioni,et al.  Embedding Decision-Analytic Control in a Learning Architecture , 1991, Artif. Intell..

[5]  David Chapman,et al.  Planning for Conjunctive Goals , 1987, Artif. Intell..

[6]  Oren Etzioni,et al.  The First Law of Robotics (A Call to Arms) , 1994, AAAI.

[7]  Raymond J. Mooney,et al.  Changing the Rules: A Comprehensive Approach to Theory Refinement , 1990, AAAI.

[8]  Ralph E. Droms,et al.  Access to heterogeneous directory services , 1990, Proceedings. IEEE INFOCOM '90: Ninth Annual Joint Conference of the IEEE Computer and Communications Societies@m_The Multiple Facets of Integration.

[9]  Richard B. Segal.  Softbots as Testbeds for Machine Learning , 1992 .

[10]  Marcel Schoppers,et al.  Universal Plans for Reactive Robots in Unpredictable Environments , 1987, IJCAI.

[11]  David A. McAllester,et al.  Systematic Nonlinear Planning , 1991, AAAI.

[12]  Oren Etzioni,et al.  Explanation-Based Learning: A Problem Solving Perspective , 1989, Artif. Intell..

[13]  L. Morgenstern,et al.  Foundations of a logic of knowledge, action, and communication , 1988 .

[14]  Michael J. Pazzani,et al.  A Knowledge-intensive Approach to Learning Relational Concepts , 1991, ML.

[15]  Robert C. Moore.  A Formal Theory of Knowledge and Action , 1984 .

[16]  Yoav Shoham,et al.  Agent-Oriented Programming , 1992, Artif. Intell..

[17]  Oren Etzioni,et al.  Tractable Closed World Reasoning with Updates , 1994, KR.

[18]  Richard Fikes,et al.  STRIPS: A New Approach to the Application of Theorem Proving to Problem Solving , 1971, IJCAI.

[19]  Tom M. Mitchell,et al.  A Personal Learning Apprentice , 1992, AAAI.

[20]  Daniel S. Weld,et al.  UCPOP: A Sound, Complete, Partial Order Planner for ADL , 1992, KR.

[21]  Oren Etzioni,et al.  An Approach to Planning with Incomplete Information , 1992, KR.

[22]  P. Schönemann.  On artificial intelligence , 1985, Behavioral and Brain Sciences.

[23]  Daniel S. Weld,et al.  Xii: Planning with Universal Quantification and Incomplete Information , 1994 .

[24]  Maria Gini,et al.  Deferred Planning and Sensor Use , 1990 .

[25]  Allen Newell,et al.  Intelligent control of external software systems , 1993, Artif. Intell. Eng..

[26]  Oren Etzioni,et al.  OS Agents: Using AI Techniques in the Operating System Environment , 1993 .