1999 http://www.msci.memphis.edu/~cmattie/Sparse_Distributed_Memory_as_a_tool_for_Conscious_Software_Agents/Sparse_Distributed_Memory_as_a_tool_for_Conscious_Software_Agents.html Sparse Distributed Memory 1.mht SDM (Sparse Distributed Memory) is a content addressable, associative memory technique which relies on close memory items tending to be clustered together, with some abstraction and blurring of details. We use the auto-associative version of SDM as an associative memory in the conscious software agent, CMattie. SDM retrieves the behaviors and emotions associated with an incoming percept. This association relies on similar percepts having occured in the past and having been associated with some behaviors and emotions. So, observing some percept later on should trigger into attention the previous behaviors taken and emotions aroused when similar percepts were observed in the past. Each perception register contains some seminar key value like seminar organizer, speaker, date, location, etc… The results obtained so far are good and promising. Some other possible use for Sparse Distributed Memory in CMattie is the disambiguation of each perception register by removal of some inherent noise or misspells. sparse distributed memory no The International Journal on Artificial Intelligence Tools. River Edge, NJ: World Scientific Publishing Company. future http://csrg.cs.memphis.edu/csrg/assets/papers/ConAg%20-%20a%20reusable%20framework%20for%20developing.pdf ConAg.pdf ConAg is a reusable framework, written in Java, for creating “conscious” software agents. Its particular focus is on these agents’ “consciousness” mechanism. A “conscious” software agent is an cognitive agent that integrates numerous artificial intelligence mechanisms to implement Bernard Baars’ global workspace theory, a psychological theory of mind. This article gives overviews of “conscious” software agents and the need for a software reuse strategy for these agents’ development. It describes the architectural style of “conscious” software agents, from which ConAg is built. ConAg is then discussed, including the design patterns and components utilized in this framework. architecture no Dauthenhahn, Kerstin (Ed.). Human Cognition and Social Agent Technology. Amsterdam: John Benjamins Publishing Company. 2000 http://www.msci.memphis.edu/~cmattie/Consciousness_and_Conceptual_Learning_in_a_Socially_Situated_Agent/Consciousness_and_Conceptual_Learning_In_A_Socially_Situated_Agent.html Consciousness and Conceptual Learning.doc This paper describes the architecture of a "conscious" software agent, Conscious Mattie, who is socially situated. We’ll focus on consciousness and conceptual learning. CMattie gathers seminar information via email from humans, composes an announcement of the next week’s seminars, and mails it to members of a mailing list that she maintains. Through further interaction with seminar organizers, CMattie learns new variations to seminars, such as a colloquium, that must also be announced, but is to be handled differently. In this paper, we discuss in detail the functionality of CMattie’s perception and consciousness modules that bring about socially situated intelligence. Using a limited capacity global workspace that coordinates with the perception module, CMattie communicates with human seminar organizers. Through such interaction CMattie gathers missing information and learns new concepts for use in composing the announcements. Here a "conscious" software architecture and its mechanisms are proposed as a method for designing and implementing useful, socially situated, software agents. architecture no Proc. Workshop on Behavior Planning for Life-Like Characters and Avatars, i3 Spring Days '99, March, Sitges, Spain 1999 no Consciousness and Cognition future http://www.msci.memphis.edu/~cmattie/A_Software_Agent_Model_of_Consciousness/A_Software_Agent_Model_of_Consciousness.html Models Of Consciousness.mht Baars has proposed a psychological theory of consciousness, called global workspace theory. The present study describes a software agent implementation of that theory, called Conscious Mattie (CMattie). CMattie operates in a clerical domain from within a UNIX operating system, sending messages and interpreting messages in natural language that organize seminars at a university. CMattie fleshes out global workspace theory with a detailed computational model that integrates contemporary architectures in cognitive science and artificial intelligence. Baars lists the psychological "facts that any complete theory of consciousness must explain" in his appendix to In the Theater of Consciousness (1997); global workspace theory was designed to explain these "facts." The present article discusses how the design of CMattie accounts for these facts and thereby the extent to which it implements conscious software. The success of CMattie hopefully will stir up the general question of whether it is possible to achieve consciousness in a computer. architecture no Proceedings of the IEEE Conference on Systems, Man and Cybernetics, 2646-2651 http://www.msci.memphis.edu/~cmattie/IDA_ACognitiveAgentArchitecture.pdf IDA_ACognitiveAgentArchitecture.pdf 1998 For most of its four decades of existence artificial intelligence has devoted its attention primarily to studying and emulating individual functions of intelligence. During the last decade, researchers have expanded their efforts to include systems modeling a number of cognitive functions (Albus, 1991, 1996; Ferguson, 1995; Hayes-Roth, 1995; Jackson, 1987; Johnson and Scanlon, 1987; Laird, Newall, and Rosenbloom, 1987; Newell, 1990; Pollack, ????; Riegler, 1997; Sloman, 1995). There’s also been a movement in recent years towards producing systems situated within some environment (Akman, 1998; Brooks, 1990; Maes, 1990b). Some recent work of the first author and his colleagues have combined these two trends by experimenting with cognitive agents (Bogner, Ramamurthy, and Franklin to appear; Franklin and Graesser forthcoming; McCauley and Franklin, to appear; Song and Franklin , forthcoming; Zhang, Franklin and Dasgupta, 1998; Zhang et al, 1998). This paper briefly describes the architecture of one such agent architecture no Papers from the AAAI Fall Symposium entitled Emotional and Intelligent: The Tangled Knot of Cognition. Menlo Park, CA: AAAI Press, 122 - 127 1998 http://www.msci.memphis.edu/~cmattie/ArchitectureForEmotion.pdf ArchitectureForEmotion.pdf The addition of emotions may be the key to producing rational behavior in autonomous agents. For situated agents, a different perspective on learning is proposed which relies on the agent’s ability to react in an emotional way to its dynamically changing environment. Here an architecture of mind is presented with the ability to display adaptive emotional states of varying types and intensities, and an implementation, “Conscious” Mattie (CMattie), of this architecture is discussed. Using this architecture, CMattie will be able to interact with her environment in a way that includes emotional content at a basic level. In addition she will learn more complex emotions which will enable her to react to her situation in a more complex manner. A general description is given of the emotional mechanisms of the architecture and its effects on learning are explained. architecture no Proceedings of the ISCA 5th International Conference (Denver, Colorado, June 1999). Raleigh NC: International Society for Computers and Their Applications - ISCA 1999 no Fifth International Conference of the Society for Adaptive Behavior-SAB98. Zurich, Switzerland, MIT Press 1998 http://www.msci.memphis.edu/~cmattie/Learning_Concepts_in_Software_Agents/Learning_Concepts_in_Software_Agents.html Learning Concepts in Software Agents.mht This concept-paper explores issues related to learning new concepts in software agents which inhabit dynamic domains. We argue that agents learn based on what they already know and agents solve new problems which they encounter by making analogies to previously solved problems of similar type. We explore these issues within the scope of the Cognitive Agent Architecture and Theory (CAAT) [Franklin, 1997]. The architecture of a multi-agent system, CMattie which is based on the CAAT strategy and inhabits an email-based dynamic domain, is described. CMattie gathers information from humans, composes announcements of next week's seminars, maintains a mailing list, mails the weekly seminar announcement to members of that mailing list, and learns new variations to seminars and other such events which have to be announced to the members of the mailing list as her domain changes with new types of seminar-like events. Learning mechanisms being implemented in this system which enable CMattie to adapt to her changing domain are described. Through her learning, CMattie acquires new domain-specific concepts, thus adapting to her dynamic domain. no Proc. of The Second Asia Pacific Conference on Simulated Evolution and Learning. Canberra, Australia, 24-27 Nov. 1998 1998 http://www.msci.memphis.edu/~cmattie/Conscious_Learning_In_An_Adaptive_Software_Agent/Conscious_Learning_In_An_Adaptive_Software_Agent.ps Conscious_Learning_1.ps This paper describes the "conscious" learning mechanism of an adaptive software agent, Conscious Mattie (CMattie). This conscious software agent takes the role of a departmental seminar organizer. The agent gathers seminar information via email from humans, composes an announcement of the next week's seminars, and mails it to members of a mailing list which she maintains. Through further interaction with seminar organizers, CMattie learns to handle new variations of seminars, such as colloquia. In this paper, the agent's perception and consciousness modules are described. Using a limited capacity global workspace that coordinates with the perception module, CMattie communicates with human seminar organizers. Through such interaction, CMattie gathers missing information and learns new concepts for use in composing the announcements no Proceedings of the Fifteenth National Conference on Artificial Intelligence, Madison, Wisconsin. Cambridge: MIT Press, 83-88. 1998 http://www.msci.memphis.edu/~cmattie/MetacognitionInSoftwareAgentsUsingClassifierSystems Comp 7282 Fall,1997.mht Software agents "living" and acting in a real world software environment, such as an operating system, a network, or a database system, can carry out many tasks for humans. Metacognition is very important for humans. It guides people to select, evaluate, revise, and abandon cognitive tasks, goals, and strategies. Thus, metacognition plays an important role in human-like software agents. Metacognition includes metacognitive knowledge, metacognitive monitoring, and metacognitive regulation. Conscious Mattie (CMattie), "living" in a Unix machine, automatically reads and understands email concerning seminars (in natural language), and composes and distributes weekly seminar schedule announcements. CMattie implements Baar’s global workspace theory of consciousness and some other cognitive theories concerning metacognition, episodic memory, emotions, and learning. Thus, the CMattie project has its cognitive science side (cognitive modeling) as well as its computer science side (intelligent software). This paper describes a case study of the design and implementation of modeling metacognition in software agents like CMattie by using a classifier system. no Proc. IEEE Joint Symposia on Intelligence and Systems, Rockville, Maryland, 374-81 1998 http://www.msci.memphis.edu/~cmattie/NaturalLanguageSensingforAutonomousAgents Natural Language Sensing for Autonomous Agents.mht In sufficiently narrow domains, natural language understanding may be achieved via an analysis of surface features without the use of a traditional symbolic parser. Here we illustrate this notion by describing the natural language sensing in Virtual Mattie (VMattie), an autonomous clerical agent. VMattie "lives" in a UNIX system, communicates with humans via email in natural language with no agreed upon protocol, and autonomously carries out her tasks without human intervention. In particular, she keeps a mailing list to which she emails seminar anouncements once a week. VMattie's various tasks include gathering information from seminar organizers, reminding organizers to send seminar information, updating her mailing list in response to human requests, composing next week’s seminar schedule announcement, and sending out the announcement to all the people on her mailing list in a timely fashion. VMattie’s limited domain requires her to deal with only nine distinct message types, each with preditable content. This allows for surface level natural language processing. VMattie's language understanding module has been implemented as a Copycat-like architecture though her understanding takes place differently. The mechanism includes a slipnet storing domain knowledge and a pool of codelets (processors) specialized for specific jobs, along with templates for building and verifying understanding. Together they constitute an integrated sensing system for the autonomous agent VMattie. With it she's able to recognize, categorizize and understand. Here we describe in detail the design and implementation of natural language sensing for autonomous agents such as VMattie who communicate (sense) via email. Assessment of Vmattie’s performance on one hundred email messages was very encouraging. Vmattie assigned 100% of the messages into the correct categories and correctly sensed 99% of the content slots. no Proceedings of the ISCA Conference on Intelligent Systems, Reno Nevada, June 1996 1996 http://www.msci.memphis.edu/~franklin/sumpy.html SUMPY.ps SUMPY is a software agent "living" in and helping to maintain a UNIX file system for better disk space utilization by compressing and backing up. Built using subsumption architecture, SUMPY displays a "plug and play" property. A new UNIX maintenance task can be added to SUMPY's repertoire without modification of existing layers. One of SUMPY's layers sports a fuzzy control mechanism enabling it to achieve its goals in a real-world manner. Another restricts SUMPY's activity to times of slow CPU use. An experiment in agent architecture and in the use of agents for such maintenance tasks, SUMPY promises to prove useful, and has added no significant problems to the test systems. no 1996 http://www.msci.memphis.edu/~franklin/vmattie.html Virtual Mattie.mht One important role for autonomous agents is to assume tasks previously performed by humans. Such tasks often require communication with humans, and the coordination of multiple activities. What sort of agent architecture will empower an agent to collaborate with humans in the process of autonomously carrying out an every day clerical task? What if this task requires the coordination of several different activities, including the composition of simple documents? Here we propose such an architecture and include a preliminary report on its first implementation. This architecture enables a softbot, Virtual Mattie, to actively gather information from humans, compose announcements of next week's seminars, and mail them each week to a list that she keeps updated, all without the supervision of a human. VMattie's architecture combines Maes' behavior net architecture and Hofstadter and Mitchell's Copycat architecture and significantly extends them. ``Living'' in a UNIX environment, VMattie's communication with humans is entirely via email with no agreed upon protocol. Thus natural language processing in a narrow domain is required. She also warns seminar organizers of impending time or place conflicts. In future versions, VMattie will learn the habits of organizers in order to more effectively dun them for information. With her low level operations based entirely on codelets, VMattie can also be viewed as a multiagent system. The implementation is in Java and Perl. VMattie is also an interdisciplinary project, leaning heavily on a psychological analysis of a corpus of the real Mattie's messages as a basis for natural language understanding. no Journal of Consciousness Studies, 4 (4) 1997, 322-234 1997 Cybernetics and Systems, special issue on Epistomological Issues in Embodied AI, 28:6 1997 499-520 1997 http://csrg.cs.memphis.edu/csrg/assets/papers/Autonomous%20Agents%20as%20Embodied%20AI.htm Autonomous Agents as Embodied AI.mht This paper is primarily concerned with answering two questions: What are necessary elements of embodied architectures? How are we to proceed in a science of embodied systems? Autonomous agents, more specifically cognitive agents, are offered as the appropriate objects of study for embodied AI. The necessary elements of the architectures of these agents are then those of embodied AI as well. A concrete proposal is presented as to how to proceed with such a study. This proposal includes a synergistic parallel employment of an engineering approach and a scientific approach. It also proposes the exploration of design space and of niche space. A general architecture for a cognitive agent is outlined and discussed. no Proceedings of the International Workshop on Affect in Interactions Towards a New Generation of Interfaces, Siena, Italy, October 21-22, 1999. New York: Springer-Verlag. 1999 http://www.msci.memphis.edu/~mccaulet/EmotionBasedConsciousAgentArch.pdf EmotionBasedConsciousAgentArch.pdf Consciousness and Cognition 8:285-305 1999 http://csrg.cs.memphis.edu/csrg/assets/papers/A%20Software%20Agent%20Model%20of%20Consciousness.doc A Software Agent Model of Consciousness.doc Baars has proposed a psychological theory of consciousness, called global workspace theory. The present study describes a software agent implementation of that theory, called "Conscious" Mattie (CMattie). CMattie operates in a clerical domain from within a UNIX operating system, sending messages and interpreting messages in natural language that organize seminars at a university. CMattie fleshes out global workspace theory with a detailed computational model that integrates contemporary architectures in cognitive science and artificial intelligence. Baars lists the psychological "facts that any complete theory of consciousness must explain" in his appendix to In the Theater of Consciousness (1997); global workspace theory was designed to explain these "facts." The present article discusses how the design of CMattie accounts for these facts and thereby the extent to which it implements global workspace theory. no 1999 International Conference Genetic and Evolutionary Computation (GECCO). July 1999 1999 http://csrg.cs.memphis.edu/csrg/assets/papers/Using%20Genetic%20Algorithms%20for%20Sparse%20Distributed.doc Using Genetic Algorithms for Sparse Distributed.doc In this paper, we describe the use of Genetic Algorithms to initialize a set of hard locations that makes up the storage space of Sparse Distributed Memory (SDM). SDM is an associative memory technique that uses binary spaces, and relies on close memory items tending to be clustered together, with some level of abstraction and with details blurring. An important factor in the physical implementation of SDM is how many hard locations are used, which greatly affects the memory capacity. It is also dependent somehow on the dimension of the binary space used. For the SDM system to function appropriately, the hard locations should be uniformly distributed over the binary space. We represented a set of hard locations of SDM as population members, and employed GA to search for the best (fittest) distribution of hard locations over the vast binary space. Fitness is based on how far each hard location is from all other hard locations, which measures the uniformity of the distribution. The preliminary results are very promising, with the GA significantly outperforming random initialization used in most existing SDM implementations. This use of GA, which is similar to the Michigan approach, differs from the standard approach in that the object of the search is the entire population. no 2000 Connection Science 12:21–44 Proceedings of the International Conference on Cognitive Modeling, Groningen, NL, March 2000, ed. N. Taatgen 2000 http://csrg.cs.memphis.edu/csrg/assets/papers/Modeling%20Consciousness%20and%20Cognition%20in.pdf Modeling%20Consciousness_1.pdf Here we describe the architectures of two “conscious” software agents and the relatively comprehensive conceptual and computational models derived from them. Modules for perception, working memory, associative memory, “consciousness,” emotions, action selection, deliberation, and metacognition are included. The mechanisms implementing the agents are mostly drawn from the “new AI,” but include both symbolic and connectionist-like features. These models promise to yield testable hypotheses for cognitive science and cognitive neuroscience. We also explore briefly the relationship of these architectures to several recent theories of cognition including embodied cognition and the dynamical systems approach. no Workshop on Development and Learning, ed. J. Wang. Michigan State University; East Lansing, Michigan, USA: NSF; DARPA; April 5-7, 2000 2000 http://csrg.cs.memphis.edu/csrg/assets/papers/Learning%20in%20Conscious%20Software%20Agents.doc Learning in Conscious Software Agents.doc Here we describe mechanisms for a half-dozen or so different types of learning to be implemented in "conscious" software agents, and speculate briefly on their implications for human learning and development. In particular, we're concerned with the role of attention, that is, bringing specific content to consciousness, in learning. We offer computational mechanisms for such learning. We believe that the "conscious" software agent technology described below, when it matures, will allow the automation of all sorts of human information and decision making agents. no Neural Network World 10:505-521 2000 http://csrg.cs.memphis.edu/csrg/assets/papers/Deliberation%20and%20Voluntary%20Action%20in%20Conscious.pdf Deliberation%20and%20Voluntary%20Action_1.pdf Here we describe briefly the architecture and mechanisms of a “conscious” software agent named IDA who is to act as a type of employment agent. We go on to show how this architecture and these mechanisms allow IDA to deliberate much as we humans do and to make voluntary choices of actions. The technology described, if successful, should allow the automation of the work of many human information agents. The IDA model also makes, hopefully testable, hypotheses about how human deliberation and volition. no Artificial Intelligence Communications 13:183-193 2000 http://csrg.cs.memphis.edu/csrg/assets/papers/Building%20Life-Like%20Conscious%20Software%20Agents.pdf Building%20Life-Like%20Conscious_1.pdf Here we'll briefly describe action selection and language generation mechanisms in two "lifelike" software agents, CMattie and IDA, and discuss issues that bear on the topics of architectures for behavior control, interdependencies between emotions and goal-based behaviors, and coordination of scripted and improvised behaviors. These agents are life-like in the sense of interacting with humans via email in natural language. They are "conscious" only in the sense of implementing a psychological theory of consciousness (Baars 1988, 1997). At this writing we are exploring the transition from scripted language production to more improvised speech generation. We are also investigating deliberative behavior selection mechanisms whereby alternative scenarios are produced and evaluated, and one of them chosen and acted upon. ida cmattie no Proceedings of the Symposium on Designing a Functioning Mind, Birmingham, England, April 2000, ed. A. Sloman 2000 http://csrg.cs.memphis.edu/csrg/assets/papers/A%20Consciousness%20Based%20Architecture.pdf A%20Consciousness%20Based%20Architecture.pdf Here we describe an architecture designed to accommodatemultiple aspects of human mental functioning. In a roughly star-shaped configuration centered on a “consciousness” module, the architecture accommodates perception, associative memory, emotions, action-selection, deliberation, language generation, behavioral and perceptual learning, self-preservation and metacognition modules. The various modules (partially) implement several different theories of these various aspects of cognition. The mechanisms used in implementing the several modules have been inspired by a number of different “new AI” techniques. One software agent embodying much of the architecture is in the debugging stage (Bogner et al. in press). A second, intending to include all of the modules of the architecture is well along in the design stage (Franklin et al. 1998). The architecture, together with the underlying mechanisms, comprises a fairly comprehensive model of cognition (Franklin & Graesser 1999). The most significant gap is the lack of such human-like senses as vision and hearing, and the lack of realworld physical motor output. The agents interact with their environments mostly through email in natural language. no SSGRR-2001 International Conference on Advances in Infrastructure for e-Business, e-Education and e-Science on the Internet L'Aquila, Italy. August 6-12, 2001 2001 http://csrg.cs.memphis.edu/csrg/assets/papers/Learning%20Mechanisms%20for%20Intelligent%20Systems.pdf Learning%20Mechanisms_1.pdf In this paper, we describe mechanisms for several different types of learning being implemented in “conscious” software agents. We argue that in complex, dynamic domains, such learning mechanisms are essential for software agents to adapt and effectively ‘live’ in those domains. We further believe that a development period is required for knowledge acquisition. Particularly, in complex, dynamic domains where knowledge engineering is expensive, the development period provides a simple, but cost effective solution to knowledge acquisition. We speculate on the implications of these mechanisms for the evolving, complex electronic world, and for human learning and development. no Proceedings Of Sixth International Symposium on Artificial Life and Robotics (AROB-01). 2001 http://csrg.cs.memphis.edu/csrg/assets/papers/Deliberative%20Decision%20Making%20in%20Conscious.pdf Deliberative Decision Making in Conscious.pdf When we humans are faced with a problem to solve, we often create in our mind different strategies or possible solutions. We imagine the effects of executing each strategy or trial solution without actually doing so. It’s a kind of internal virtual reality. Eventually, we decide upon one strategy or trial solution and try solving the problem using it. This whole process is called deliberation. During the deliberation process several, possibly conflicting, ideas compete to be selected as the strategy or solution of the problem. One such is chosen voluntarily. In 1890 William James proposed a model that describes this voluntary decision-making calling it the ideo-motor theory. In this theory the mind is considered to be the seat of many ideas related to each other either favorably or antagonistically. Whenever an idea prompts an action by becoming conscious, antagonistic ideas may object to it, also by becoming conscious, and try to block that action. Or, other favorable ideas may become conscious to support it, trying to push its selection. While this conflict is going on among several ideas we are said to be "deliberating". Software agents, so equipped, should be able to make voluntary decisions much as we humans do. This paper describes a computational mechanism for this deliberation process, including James’ ideo-motor theory of voluntary action. It also describes an implementation of the mechanism in a software agent. Some preliminary experimentation is also reported. no CogSci2001: Proceedings of the 23rd Annual Conference of the Cognitive Science Society, ed. J. D. Moore, and K. Stenning. Mahwah, NJ: Lawrence Erlbaum Associates; August 1-4, 2001 2001 http://csrg.cs.memphis.edu/csrg/assets/papers/Modeling%20Cognition%20with%20Software%20Agents.pdf Modeling Cognition with Software Agents.pdf We propose the use of autonomous software agents as cognitive models that generate testable hypotheses about human cognition. While such agents are typically produced to automate practical human tasks, they can be designed within the constraints of a psychological theory. As an example we describe an agent designed within global workspace theory that accommodates several other theories as well. We discuss various resulting hypotheses, including a new interpretation of the readiness potential data of Libet. no Soft Computing Agents: New Trends for Designing Autonomous Systems, ed. V. Loia, and S. Sessa. Berlin: Springer (Physica-Verlag), 1 - 46 2001 Practical Applications of Intelligent Agents, ed. Z. Chen, and L. C. Jain. Berlin: Springer-Verlag 2001 http://csrg.cs.memphis.edu/csrg/assets/papers/Automating%20Human%20Information%20Agents.pdf Automating Human Information Agents.pdf no AAAI Spring Symposium on Robust Autonomy; American Association for Artificial Intelligence; Stanford, CA; March 2001 2001 http://csrg.cs.memphis.edu/csrg/assets/papers/An%20Agent%20Architecture%20Potentially%20Capable%20of.pdf An Agent Architecture Potentially Capable of.pdf Robust autonomy on the part of software agents requires, at least in part, the ability to deal intelligently with novel and unexpected situations. According to global workspace theory, dealing with such situations is one of the primary functions of consciousness in humans. Below we briefly describe two software agents that implement this psychological theory, and discuss their resulting potential for robust autonomy. no Connection Science 14:371-385 2002 http://csrg.cs.memphis.edu/csrg/assets/papers/Multi-AgentIDAJrnl.pdf Multi-agentIda.pdf In the US Navy, at the end of each sailor's tour of duty, he or she is assigned to a new job. The Navy employs some 280 people, called detailers, full time to effect these new assignments. The IDA (Intelligent Distribution Agent) prototype was designed and built to automate, in a cognitively plausible manner, the job of the human detailers. That model is being redesigned to function as a multi-agent system. This is not a trivial matter due to the fact that there would need to be approximately 350,000 individual agents. There are also many issues relating to how the agents interact and how all entities involved, including humans, exercise their autonomy. This paper describes both the IDA prototype and the Multi-Agent IDA system being created from it. We will also discuss several of the major issues regarding the design, interaction, and autonomy of the various agents involved. no Cognitive Science Quarterly 2:363-386 2002 http://csrg.cs.memphis.edu/csrg/assets/papers/CSQ-AS-ForConsciousAgents-Final.pdf CSQ-AS-ForConsciousAgents-Final.pdf In this paper we describe an action selection mechanism for a "conscious" software agent. We discuss briefly the main cognitive modules of our agent architecture. Our focus is on the operational/functional details of the “consciousness” module and the action selection mechanism and how these two work together. We describe how events come to “consciousness,” how “conscious” events prime and bind to relevant actions/action-plans, and how the most relevant behavior/action is selected and executed. Our mechanisms provide the flexibility to interleave actions that operate under different tasks. We consider the descriptions in this paper a listing of the various hypotheses about human cognition resulting from our design of the agent described. no Editorial Review: Science & Consciousness Review. 2003 http://csrg.cs.memphis.edu/csrg/assets/papers/crick-koch-editorial-review.pdf crick-koch-editorial-review.pdf Recently Crick and Koch offered a “Framework for Consciousness” (2003). Pradeep Mutalik’s review of that article in SCR (Mutalik 2003) asserts that “Crick and Koch describe ten aspects of a framework that they believe offers a coherent scheme for explaining the neural correlates of consciousness.” no 2003 http://csrg.cs.memphis.edu/csrg/assets/papers/IDAonFramworikJournal.doc IDAonFramworikJournal.doc Crick and Koch recently introduced a Framework for Consciousness intended as a means of facilitating the explanation of the neural correlates of consciousness. The IDA model of Global Workspace Theory is both a conceptual and computational model of consciousness and of much of cognition. Here we systematically compare the ten aspects of the Framework with hypotheses derived from the IDA model. On the seven aspects to which the IDA model relates there's a surprising amount of agreement. In addition the IDA model serves to flesh out the Framework with detailed hypotheses about the functionality of the aspect, and often of its structure as well. no Seventh Conference of the Association for the Scientific Study of Consciousness - ASSC7 (Memphis, May 30 - June 2, 2003) 2003 Human memory seems to come in myriad of versions: sensory memory, working memory, transient episodic memory, autobiographical memory, semantic memory and procedural memory. This research attempts to computationally model these memory systems and their relationship to consciousness in the context of an existing software agent, IDA (Intelligent Distribution Agent). IDA implements Baars' global workspace theory of consciousness. We implement some of the memory systems in IDA with Kanerva's Sparse Distributed Memory architecture. Using the IDA model, we are attempting to computationally test several hypotheses that serve to simplify relationships between the various types of memory and their relationship to consciousness. We hypothesize that Working Memory contents cue retrieval of local associations from Transient Episodic Memory and Long-Term Memory; an attention mechanism selects the contents of consciousness from the contents of Working Memory and these local associations; and only conscious contents are stored in Transient Episodic Memory and hence, Long Term Memory, due to the consolidation process from Transient Episodic Memory to Long Term Memory. This research is aimed at better understanding of memory systems, and the relationship between memory and consciousness. no Trends in Cognitive Sciences Vol7 No.4 April 2003 2003 http://csrg.cs.memphis.edu/csrg/assets/papers/TICSarticle.pdf TICSarticle.pdf Active components of classical working memory are conscious, but traditional theory does not account for this fact. Global Workspace theory suggests that consciousness is needed to recruit unconscious specialized networks that carry out detailed working memory functions. The IDA model provides a fine-grained analysis of this process, specifically of two classical workingmemory tasks, verbal rehearsal and the utilization of a visual image. In the process, new light is shed on the interactions between conscious and unconscious aspects of working memory. no Journal of Consciousness Studies 10:47-66 2003 http://csrg.cs.memphis.edu/csrg/assets/papers/IDA-ConsciousArtifact.doc IDA-ConsciousArtifact.doc After discussing various types of consciousness, several approaches to machine consciousness, software agent, and global workspace theory, we describe a software agent, IDA, that is "conscious" in the sense of implementing that theory of consciousness. IDA perceives, remembers, deliberates, negotiates, and selects actions, sometimes "consciously." She uses a variety of mechanisms, each of which is briefly described. It's tempting to think of her as a conscious artifact. Is such a view in any way justified? The remainder of the paper considers this question. no Human Interaction with Autonomous Systems in Complex Environments: Papers from 2003 AAAI Spring Symposium, ed. D. Kortenkamp, and M. Freed. Palo Alto: AAAI 2003 http://csrg.cs.memphis.edu/csrg/assets/papers/An%20Autonomous%20Software%20Agent%20for%20Navy%20Personnel%20Work.doc An Autonomous Software Agent_1.doc IDA is an autonomous software agent whose task is to assign a sailor to a new tour of duty at the end of the old. Her task input is most often initiated by an email from the sailor; thereafter she acts completely autonomously. The task requires IDA to access databases, to deliberate, to perform complex constraint satisfaction, and to negotiate with the sailor in natural language. IDA's architecture and mechanisms are motivated by a variety of computational paradigms and implement a number of cognitive models including the Global Workspace Theory of consciousness. IDA is able to interact in depth with sailors in a relatively complex real-world environment involving job requirements, Navy policy, location, travel, training, dates, costs, preferences, etc. Here we briefly describe the IDA technology and her interactions as a case study of the interaction of sophisticated software agents with humans. no IEEE International Conference on Systems, Man and Cybernetics - SMC2004 2004 http://www.cs.memphis.edu/~ramamurt/UR%20IEEE%20SMC2004%20Paper%20FinalJuly12.pdf UR IEEE SMC2004 Paper FinalJuly12.pdf This paper presents a Modified Sparse Distributed Memory architecture for use in software agents with natural language processing capabilities. We have modified Kanerva’s Sparse Distributed Memory (SDM) into an architecture with a ternary memory space. This enables the memory to be used in IDA, the Intelligent Distribution Agent built for the U.S. Navy. IDA implements Baars’ global workspace theory, a psychological theory of consciousness. As a result, it can react to novel and problematic situations in a more flexible, more human-like way than traditional AI systems. IDA performs a function, namely billet assignment, heretofore reserved for humans. We argue that such flexibility requires advanced memory systems such as transient episodic memory and auto-biographical memory. Here, we present the architecture, tests and results of this modified SDM system which can be used as a transient episodic memory in suitable software agents. no AAAI Fall Symposium on Dialogue Systems for Health Communication; American Association for Artificial Intelligence; Washington,DC, USA. October 22-24, 2004 2004 http://csrg.cs.memphis.edu/csrg/assets/papers/TRIAGENT%20information%20agent%20based%20on%20IDA%20technology.pdf TRIAGENT information agent based on IDA technology.pdf Busy hospital emergency rooms are concerned with shortening the waiting times of patients, with relieving overburdened physicians, and with reducing the number of mistakes made by triage nurses. Here we propose a software agent, dubbed TIA (Triage Information Agent) that, via dialogue in English, would gather both logistical and medical information from a patient for later use by the triage nurse. TIA would also give tentative, possible diagnoses to the triage nurse, along with recommendations for non-physician care. The IDA Technology makes a software agent such as TIA feasible, at least in principle. no AAAI Spring Symposia Technical Series Technical Reports SS-04-02 2004 http://csrg.cs.memphis.edu/csrg/assets/papers/2004-fe-ml.pdf 2004-fe-ml.pdf The use of feelings and emotions as primary motivators and facilitators of several types of learning within the IDA architecture is described. no First World Congress on Lateral-Computing (WCLC2004) (Bangalore, India, December 17-19, 2004). 2004 http://www.cs.memphis.edu/~ramamurt/UR-SF%20Self-PreservationPaperFinalOct20.pdf UR-SF Self-PreservationPaperFinalOct20.pdf Humans and other animals have a sense of selfpreservation that motivates them to take appropriate actions to preserve themselves. Intelligent software systems must have built-in or learnt functionality to detect emergencies in their domains, and must be able to act appropriately to save their state and data, and shutdown cleanly if necessary. They must also be able to restart and restore themselves to their most recent state once their computing environment comes back online. We have designed such mechanisms for an Intelligent Distribution Agent (IDA) built for the U.S. Navy. IDA implements Baars’ Global Workspace Theory of consciousness. As a result, she can react to novel and problematic situations in a more flexible, more human-like way than traditional AI systems. We hypothesize that software agents like IDA must have self-preservation mechanisms to adapt and survive in their domains. This paper presents a design for self-preservation mechanisms consistent with IDA’s cognitive cycle which can be implemented in intelligent software systems. no Visions of Mind, ed. D. Davis. Hershey, PA: IDEA Group, Inc 2005 http://csrg.cs.memphis.edu/csrg/assets/papers/conscious%20based%20architecture%20for%20a%20functioning%20mind.pdf conscious based architecture_1.pdf Here we describe an architecture for an autonomous software agent designed to model a broad spectrum of human cognitive and affective functioning. In addition to featuring “consciousness,” the architecture accommodates perception, several forms of memory, emotions, actionselection, deliberation, ersatz language generation, several forms of learning, and metacognition. One such software agent, IDA, embodying much of this architecture, is up and running. IDA’s “consciousness” module is based on global workspace theory, allowing it to select relevant resources with which to deal flexibly with both exogenous and endogenous stimuli. Within this architecture, emotions implement IDA’s drives, her primary motivations. Offering one possible architecture for a fully functioning artificial mind, IDA constitutes an early attempt at the exploration of design space and niche space. The design of the IDA architecture spawns hypotheses concerning human cognition and affect that can serve to guide the research of cognitive scientists and neuroscientists. One such hypothesis is that consciousness is discrete. Master’s Thesis, The University of Memphis 1999 http://www.msci.memphis.edu/~mccaulet/thesis.pdf thesis_McC.pdf Emotions are not generally considered to be an advantage when one wishes to build an artificial intelligence system that acts in a rational manner. However, the opposite is, in fact, the truth. Emotions are an essential part of intelligence and rational behavior. Recent evidence points to the fact that humans who have suffered damage to key areas of the brain responsible for communication to the limbic system, generally regarded as the seat of emotion in the brain, cannot function in a reasonable way either socially or logically. This evidence also gives us a glimpse into what emotions really do for cognition. From here it is possible to begin to create ways of giving emotional functionality to artificial systems. One such model of an emotional mechanism is detailed that allows an agent to represent a broad range of emotions. In addition, the mechanism gives the agent the ability to learn complex emotions from a set of primitives and to associate actions that will pursue or avoid particular emotional stimuli. Two implementations, called CMattie and IDA, are being designed with this emotion component. The role of learning in systems based on emotions must also be explored. The decades old paradigm of learning via an external evaluation function or human observer is fully replaced by internal emotional judgement. A portion of this judgement must be designed into the agent from its inception, and in such a way that the agent continues to act rationally after learning more complex emotions. IEEE Journal of Robotics and Automation, Vol. 2, No. 1, March 1986, pp. 14–23; also MIT AI Memo 864, September 1985 1985 http://people.csail.mit.edu/people/brooks/papers/AIM-864.pdf AIM-864.pdf no Connection Science 1:291-323 1989 http://citeseer.ist.psu.edu/rd/95589163%2C526757%2C1%2C0.25%2CDownload/http://citeseer.ist.psu.edu/cache/papers/cs/26140/ftp:zSzzSzpublications.ai.mit.eduzSzai-publicationszSz1000-1499zSzAIM-1180.pdf/maes89how.pdf maes89how.pdf no no no no no no no http://www.amazon.com/exec/obidos/tg/detail/-/0195147030/ref=pd_ecc_rvi_4/104-4524409-5374341?%5Fencoding=UTF8&v=glance http://www.amazon.com/exec/obidos/tg/detail/-/0262561093/ref=lpr_g_1/104-4524409-5374341?v=glance&s=books http://www.amazon.com/exec/obidos/ASIN/0262111322/104-4524409-5374341 Psi theory (Dörner) Belief-Desire-Intention http://citeseer.ist.psu.edu/cache/papers/cs/394/ftp:zSzzSzwww.aaii.oz.auzSzpubzSzaaii-technoteszSztechnote56.pdf/rao95bdi.pdf Pandemonium Theory a-brain b-brain (Minsky) Copycat Architecture sparse distributed memory slipnet Maes behavior net The MicroPsi Project http://www.cognitive-agents.org/ http://www.cognitive-agents.org/publications.php 2002 MicroPsi http://www.informatik.hu-berlin.de/~bach/publication/MicroPsiArchitectureICCM03.pdf Java meta-management yes LGPL The Cognition and Affect Project http://www.cs.bham.ac.uk/~axs/cog_affect/COGAFF-PROJECT.html http://www.cs.bham.ac.uk/research/cogaff/ 1985 H-CogAff http://courses.media.mit.edu/2003spring/mas963/sloman-aisb01.pdf POP-11 SimAgent perception action emotions planning memory (LTM, associative) motivation meta-management personae rule based Layered Conscious Software Research Group http://csrg.cs.memphis.edu/ http://csrg.cs.memphis.edu/csrg/html/papers.html Global Workspace Theory (Baars) Pandemonium Theory (Jackson) VMattie http://csrg.cs.memphis.edu/csrg/assets/papers/Virtual%20Mattie-an%20Intelligent%20Clerical%20Agent.htm http://www.msci.memphis.edu/~vmattie/ Java Perl http://csrg.cs.memphis.edu/csrg/assets/papers/Natural%20Language%20Sensing%20for%20Autonomous.htm CMattie http://csrg.cs.memphis.edu/csrg/assets/papers/A%20Software%20Agent%20Model%20of%20Consciousness.doc Java no ConAg http://csrg.cs.memphis.edu/csrg/assets/papers/An%20Architecture%20for%20Emotion.doc ? Holland’s classifier system Behaviour Net (Maes) IDA http://www.msci.memphis.edu/~ida/ http://csrg.cs.memphis.edu/csrg/assets/papers/conscious%20based%20architecture%20for%20a%20functioning%20mind.pdf http://csrg.cs.memphis.edu/csrg/assets/papers/IDA%20-%20A%20Cognitive%20Agent%20Architecture.doc ACT-R http://act-r.psy.cmu.edu/ http://act-r.psy.cmu.edu/publications/ ACT-R Lisp perception action declarative memory procedural memory motivation language comprehension Production System yes LGPL Soar http://sitemaker.umich.edu/soar http://sitemaker.umich.edu/soar/documentation_and_links 1996 Soar http://www.eecs.umich.edu/~soar/sitemaker/docs/manuals/Soar8Manual.pdf Soar Architecture Working Memory Long Term Memory motivation Problem Solving Production System yes BSD Oscar http://oscarhome.soc-sci.arizona.edu/ftp/OSCAR-web-page/oscar.html VERBOSE http://www.mesa.afmc.af.mil/html/Verbose.htm SimCog http://www.lti.pcs.usp.br/SimCog/