We describe a language and protocol intended to support interoperability among intelligent agents in a distributed application. Examples of applications envisioned include intelligent multi-agent design systems as well as intelligent planning, scheduling and replanning agents supporting distributed transportation planning and scheduling applications. The language, KQML for Knowledge Query and Manipulation Language, is part of a larger DARPA-sponsored Knowledge Sharing effort focused on developing techniques and tools to promote the sharing on knowledge in intelligent systems. We will define the concepts which underly KQML and attempt to specify its scope and provide a model for how it will be used. Notice of DRAFT Status. This document presents the current draft of a specification under consideration by the DARPA Knowledge Sharing Effort. It is provided for information purposes, and should be treated as representing only the current status of discussions. It should not be interpreted a...

We have been engaged in a series of projects in which Cassie, the SNePS cognitive agent, has been incorporated into a hardware- or software-simulated cognitive robot. In this paper, we present an informal summary of our approach to anchoring the abstract symbolic terms that denote Cassie's mental entities in the lowerlevel structures used by embodied-Cassie to operate in the real (or simulated) world. We discuss anchoring in the domains of: perceivable entities and properties, actions, time, and language.

We describe a language and protocol intended to support interoperability among intelligent agents in a distributed application. Examples of applications envisioned include intelligentmulti-agent design systems as well as intelligent planning, scheduling and replanning agents supporting distributed transportation planning and scheduling applications. The language, KQML for Knowledge Query and Manipulation Language, is part of a larger DARPA-sponsored Knowledge Sharing e#ort focused on developing techniques and tools to promote the sharing on knowledge in intelligent systems. e will de#ne the concepts which underly KQML and attempt to specify its scope and provide a model for how it will be used. Please send comments to Tim Finin, Computer Science, University of Maryland, Baltimore MD 21228; #nin@cs.umbc.edu; 410-455-3522 or to Don Mckay,Paramax Systems Corporation, PO Box 517, Paoli PA 19301; mckay@prc.unisys.com; 215-648-2256. This work is partly supported byDARPA and Rome La...

SNePS, the Semantic Network Processing System, has been designed to be a system for representing the beliefs of a natural-language-using intelligent system (a "cognitive agent"). This paper expands on this motivation, discusses some of the system features that derived from this motivation, and presents four case studies of interactions with SNePS demonstrating some of these features. The features demonstrated in the case studies are: nonstandard connectives; the use of recursive rules; the Unique Variable Binding Rule, that says that two variables in a rule cannot be instantiated to the same term; and discussing sentences and propositions in natural language. 1 System Description SNePS, the Semantic Network Processing System [9, 15, 17], has been designed to be a system for representing the beliefs of a natural-language-using intelligent system (a "cognitive agent"). It has always been the intention that a SNePSbased "knowledge base" would ultimately be built, not by a programmer or k...

GLAIR (Grounded Layered Architecture with Integrated Reasoning) is a multi-layered cognitive architecture for embodied agents operating in real, virtual, or simulated environments containing other agents. The highest layer of the GLAIR Architecture, the Knowledge Layer (KL), contains the beliefs of the agent, and is the layer in which conscious reasoning, planning, and act selection is performed. The lowest layer of the GLAIR Architecture, the Sensori-Actuator Layer (SAL), contains the controllers of the sensors and effectors of the hardware or software robot. Between the KL and the SAL is the Perceptuo-Motor Layer (PML), which grounds the KL symbols in perceptual structures and subconscious actions, contains various registers for providing the agent’s sense of situatedness in the environment, and handles translation and communication between the KL and the SAL. The motivation for the development of GLAIR has been “Computational Philosophy”, the computational understanding and implementation of human-level intelligent behavior without necessarily being bound by the actual implementation of the human mind. Nevertheless, the approach has been inspired by human psychology and biology.

We report on a natural language agent, originally developed as a command driven interface, that was enhanced with time-dependence, contradiction tolerance, meta-linguistic abilities, and an overall meta-cognitive awareness. We show how these new capacities together can make an AI system’s natural language processing more robust and human-like. 1.

The GLAIR grounded layered architecture with integrated reasoning for cognitive robots and intelligent autonomous agents has been used in a series of projects in which Cassie, the SNePS cognitive agent, has been incorporated into hardware- or software-simulated cognitive robots. In this paper, we present an informal, but coherent, overview of the GLAIR approach to anchoring the abstract symbolic terms that denote an agent's mental entities in the lower-level structures used by the embodied agent to operate in the real (or simulated) world. We discuss anchoring in the domains of: perceivable entities and properties, actions, time, and language.