Summary
The Coalition Search and Rescue Task Support (CoSAR-TS) is a DARPA DAML Program initiative designed to enhance search and rescue operations through advanced integration of organizational models, policies, and intelligent task support software By utilizing AIAI’s I-X planning technology, IHMC’s KAoS policy services, and various Semantic Web Services, the project enables the rapid dynamic composition of policy-constrained services, making it an ideal application for Semantic Web technologies Key contributors to this initiative include BBN Technologies, SPAWAR, AFRL, and Carnegie Mellon University.
At the beginning of the project, the joint AIAI/IHMC aims were:
Development of base technologies respectively I-X/I-Plan and KAoS Policy and Domain Services,
Deployment of the technology in a realistic CoAX agents demonstrator scenario,
Persuasion of closer integration of these two technologies with a perspective of a uniform tool release in the future.
These goals were achieved in the subsequent years of the project as follows:
In the first year, the development of distributed multi-agent systems was achieved, integrating them with the semantic web to enhance realistic coalition search and rescue operations This innovative work led to the creation of an Intelligent Systems Demonstrator for CoSAR-TS, showcased at AAAI-2004.
In Year 2, a web services composition and policy analysis tool for semantic web services, known as I-K-C, was developed This project resulted in the publication of an article in the IEEE Intelligent Systems journal and a presentation at the ISWC 2004 conference.
The results of the project can be accessed through various websites, including the CoSAR-TS Project website, the SemWebCentral site for DAML-program results, and the I-K-C project pages hosted by AIAI and IHMC For more details, please refer to Appendix C.
The software developed in this project can be downloaded from the specified web pages Additionally, the project generated a substantial collection of high-quality publications that effectively documented and disseminated the results to both the research and military communities.
The technology created by the project is currently being utilized in collaboration with JFCOM/JPRA for the Co-OPR initiative, which serves as a foundational element for DARPA’s Integrated Battle Command program.
Introduction
The project demonstrates the use of intelligent agents and artificial intelligence planning systems operating in a distributed manner, where dynamic policies from diverse groups and individuals dictate permissions and obligations These agents leverage semantic web services to efficiently access medical information and locate nearby rescue resources.
The goal of this research is to create a demonstrator for Task Support within a dynamic Coalition Search and Rescue environment This study at AIAI is integrated with IHMC's work on KAoS policy and domain services It utilizes OWL representations and OWL-S descriptions for agents and services, contributing valuable feedback to the OWL-S and Semantic Web Services development community.
The project facilitates collaboration between software and human agents through a shared, intelligible model encompassing tasks, processes, organizational structures, capabilities, agent status, secure communication, and policy management By leveraging existing ontologies from DAML/OWL and DAML-S/OWL-S, along with tools like the CMU Matchmaker and BBN SONAT Elements of National Power Knowledge Base, the work highlights the importance of semantically represented and shared models This technology is exemplified in a coalition search and rescue scenario, showcasing its practical application.
I-X Technology
I-X Process Panels
We “deliver” useful functionality based on the I-X and ontology via I-X Process Panels (I-P 2 ) These support a user or collaborative users in selecting and carrying out
I-X Process Panels serve as intelligent task lists that enhance user productivity When integrated with other users' panels, they transform into comprehensive workflows, facilitating coordination, reporting, and messaging This collaborative environment fosters more efficient teamwork and successful project outcomes, making I-X Process Panels essential for supporting user tasks and cooperation.
A panel offers users a personalized view of ongoing activities by displaying relevant items from the four sets of entities in the model This display, combined with the current collaboration context and state, dynamically generates support options within the tool For instance, specific activity nodes may present suggestions for executing tasks through established procedural expansions, invoking agents with corresponding capabilities, or delegating activities to other agents in the environment.
Figure 2: Anatomy of an I-X Process Panel
External capabilities (invoke or query/answer)
Reroute or delegate to other panels or agents (pass)
Plan and execute a composite of these capabilities (plan or expand)
Receives “progress” or “completion” reports and other event-related messages and, where possible, interprets them to:
Understand current status of issues, activities and constraints
Understand current world state, especially status of process products
An I-X Process Panel can cope with partial knowledge and can operate even where little or no pre-built knowledge of the domain or knowledge of relationships to other panels or services is available – effectively becoming a simple “to-do” list aid in that case.
Figure 3: I-X Instant Messaging Style Interface
In 2001, the Navy Warfare Development Command (NWDC) in Newport, Rhode Island, trialed I-X/I-P 2 during the "Millennium Challenge" training exercise, leading to a pivotal shift in systems development Initially, a test interface panel allowed users to send messages to their own and other panels NWDC utilized I-P 2 with an Instant Messaging tool to log communications among coalition countries and commands, showcasing the effectiveness of agent technology over secure channels It became evident that the exchanged messages often pertained to manageable entities such as issues, activities, and preferences Consequently, the test panel evolved into an Instant Messaging-style interface that facilitated structured messaging while still supporting simple text chat This transition made it easier to convey task-related information, allowing I-X Process Panels to be described as providing "augmented" instant messaging, enhancing process, activity, and task support along with progress reporting.
Since their inception, I-X Process Panels have embraced an "intelligible messaging" interface, which has evolved to enhance cooperative and human-centric applications, particularly in scientific meetings and group work (Buckingham Shum et al., 2002) This refined interface is now central to our methodology Additionally, we have integrated a Jabber communications strategy, enabling Instant Messaging through XML content, which facilitates simpler and larger-scale deployments of the I-X Process Panels.
I-Plan
The I-X Process Panels feature an AI planner called I-Plan, which offers context-sensitive solutions for managing issues, achieving objectives, performing activities, and meeting constraints effectively.
Figure 4: I-P2 Context-sensitive “Action” Menu
The panel features an "Action" column that displays the current status of each activity along with the options available for execution The use of colors signifies the readiness of each item for immediate action.
White indicates that the item is not currently ready for execution (i.e., some temporal ordering, preconditions or other constraints might not be met).
Orange indicates that the action is ready to perform and that all preconditions and constraints are met.
Green indicates that the item is currently being performed.
Red indicates a failure for which failure recovery planning steps might be initiated.
The "Actions" menu on the panel offers a context-sensitive and dynamically generated set of options tailored to the capabilities of other panels and services It utilizes the integrated planner, I-Plan, to identify and select relevant plans or Standard Operating Procedures (SOPs) that correspond to the specific item in question.
I-Plan can perform hierarchical partial-order composition of plans from a library of single level plan schemas or “Standard Operating Procedures” This library can be augmented during planning either with a simple “activity details” interface to add in specific ways to expand a given action (intended for use by users familiar with the application domain but not AI planning techniques) or with a more comprehensive graphical domain editor Grammars and lexicons for the domain are built automatically during domain editing to assist the user.
Future enhancements to I-Plan will introduce a "How do I do this?" feature in the Action menu This option will take into consideration other active items on the panel, ensuring that it addresses mutual satisfaction among open variables and various constraints.
Other I-X Tools
There are other tools in the I-X suite include messaging tools and various information viewers (e.g map, 3D VRML and PDA interfaces) and editors, along with three specific tools: I-DE, I-Q and I-Space:
I-DE (I-X Domain Editor) allows the creation, maintenance and, ultimately, the publication of Standard Operating Procedures (SOPs), generic approaches to archetypal activities.
I-Q (I-Query) is a versatile I-X agent shell designed to enable agents to interact effectively with various query services By integrating suitable mechanisms, it enhances an agent's ability to respond by incorporating new facts or constraints into the current state of the panel A common application of I-Q is in retrieving information from external sources, such as the semantic web.
Figure 5: I-Space Organizational Relationships Tool
I-Space facilitates the management of organizational relationships among various agents, with the type of relationship—such as supervisor-supervisee—affecting the nature of their interactions An agent's choices are influenced by its position within the organizational hierarchy and its understanding of the capabilities and current status of other agents Additionally, tools like the KAoS Policy Administration Tool (KPAT) enable the exchange of relationships between agents and organizations.
I-X Message Formats
There are a number of messages that are used within the I-X Process Panels and that can be passed between panels and other services and agents.
Issues, Activities, Constraints and Annotations
Current state information (world state constraints)
Plans (composites of Issues, Activities, Constraints and Annotations)
Reports on progress or completion of nominated activities
The first 3 relate to the core underlying ontology on which I-X is based The other two message types provide status and other contextual information.
Reports and Current State
Activities and panel items can be transferred between panels or to compatible services and agents, allowing them to send back progress and completion reports to the original sender This system enables effective monitoring of activity progress, receipt of milestone updates, and verification of activity completion.
Figure 6: I-X Custom State Viewer – Map View
Current environmental data can be transmitted to panels through "world state" constraints, which may originate from sensors or be derived from analytical reporting systems.
Instant messaging systems, like Jabber, provide valuable presence or status information that indicates whether a user, panel, or agent is active and available for communication This feature allows registered users to access real-time availability data for their contacts, which are organized in "buddy lists." By offering this current state information, Jabber enhances communication efficiency within process panels.
Incoming completion reports and current state information can activate subsequent activities once specific temporal or other constraints are met For instance, when users exchange presence or location data, this information is displayed on the receiver's state panel, potentially triggering actions that depend on a user's status, such as waiting for them to come online.
I-X also allows custom state viewers to be added to augment or replace the simple tabular current state view in a normal I-P 2 panel An example of a viewer for such state information could be the BBN OpenMap™ tool (BBN, 2003) Changes to information in any viewer, or coming in via messages from outside of panels are synchronized.
The I-N-C-A framework (Issues - Nodes - Constraints - Annotations) serves as the foundational ontology for the I-X approach, facilitating the representation of processes and their products within I-X Process Panels This conceptual model enables effective communication and collaboration between human users and system components, ensuring a structured exchange of activity-oriented I-X Messages for shared tasks.
In , processes and their products are viewed as composed of a collection of "Issues" that relate to them, representing potential requirements and inquiries that arise from analysis and critique.
Nodes represent activities within a process or components of a physical product, often consisting of sub-nodes that create a hierarchical structure These nodes are interconnected through a variety of detailed constraints Additionally, annotations can be associated with the processes or products, offering rationale, valuable information, and other descriptive insights.
The X systems integration approach utilizes the I-N-C-A Model of Synthesized Artifacts, offering a straightforward abstraction that ensures a highly flexible, extendable, and comprehensible representation of processes and their products in I-X This model facilitates effective communication between human and system agents working collaboratively on shared tasks, allowing each party to take the initiative in managing different aspects at various stages.
The I-N-C-A and I-N-OVA frameworks, developed by Tate in 1996, were designed to enhance communication among various communities focused on formal planning theories, practical planning systems, and systems engineering methodologies These frameworks aimed to facilitate advancements in automatic plan manipulation, human communication regarding plans, reliable plan information acquisition, and formal reasoning about plans Over time, they have served as a foundation for numerous research initiatives, practical applications, and the development of international standards for plan and process representations.
AI on plan representations, work from the process and design communities and the standards bodies, and the part that played in this see Tate (1998).
The typography for rendering has evolved throughout the I-X research development, reflecting a clearer understanding of its components Initially, represented Issues, Nodes, Critical, and Auxiliary Constraints, with a focus on distinguishing critical constraints from auxiliary ones This distinction remains vital in managing the "C" (constraints) component within the I-X architecture Annotations have always been integral to the ontology and can be linked to all components However, as the project progressed, the significance of top-level annotations—capturing the rationale behind synthesized products or processes—has increased, particularly with the growing importance of mixed-initiative and human communication Consequently, the updated rendering now signifies Issues, Nodes, Constraints, and Annotations.
The representation of issues highlights unresolved questions and unmet objectives, often stemming from analysis Constraints can indicate potential future limitations that may need to be incorporated into the design to address these issues Previously, the focus in I-X was on task-oriented actions related to ongoing processes However, this approach is evolving towards the gIBIS framework, which emphasizes framing issues as specific types of questions to be addressed This shift aims to enhance clarity and facilitate more effective problem-solving in design processes.
Deontic questions - What should we do?
Instrumental questions - How should we do it?
Criterial questions - What are the criteria?
Meaning or conceptual questions - What does X mean?
Factual questions - What is X? or Is X true?
Background questions - What is the background to this project?
Stakeholder questions - Who are the stakeholders of this project?
Miscellaneous questions - To act as a catch all.
The top five elements are expected to be the most prevalent in our task support environment, aligning with the Questions-Options-Criteria framework proposed by MacLean et al (1991) This framework has been previously utilized for capturing rationales in plans and plan schema libraries, as demonstrated in our earlier research (Polyak and Tate, 1998).
In 1999, mapping techniques akin to those utilized in Compendium (Selvin et al., 2001) were developed Compendium is capable of exchanging its issues, activities, and certain constraints and annotations with I-P 2, as noted by Buckingham Shum et al (2002) and Chen-Burger and Tate (2003).
The specifications outline various nodes that represent essential components of the design Each node can function as an artifact, possessing its own structure that includes sub-nodes and additional refinements linked to them.
KAoS Technology
KAoS Policy and Domain Management Services
KAoS is a pioneering initiative that utilizes the Semantic Web language OWL to represent domain structures and policies It offers services and tools for the specification, management, conflict resolution, and enforcement of policies within complex organizational contexts Originally designed for the dynamic needs of software agent applications, KAoS has been adapted to effectively serve both agent and traditional clients across various distributed computing platforms, including CORBA, Web Services, and Grid Computing.
Ontological Representation of KAoS Policies
KAoS utilizes ontology concepts encoded in OWL to develop policies Initially, it loads the KAoS Policy Ontologies (KPO), which define the concepts needed to describe a generic actor's environment and the associated policies Subsequently, KAoS incorporates additional ontologies that enhance these generic concepts with specific notions tailored to the particular controlled environment and application.
KAoS differentiates between authorizations, which are constraints that allow or prohibit specific actions, and obligations, which are constraints that mandate actions to be taken when certain triggers occur, or provide a waiver for such requirements Additionally, other policy constructs, including delegation and role-based authorization, are developed from these fundamental primitives and the four primary policy types.
The KAoS Policy Administration Tool (KPAT) simplifies policy management by automatically generating policies categorized into four classes: PositiveAuthorization, NegativeAuthorization, PositiveObligation, and NegativeObligation Each policy's property values, such as priority, guide its management information, while the policy type dictates the constraints KAoS applies to specific actions The action class, defined using OWL restrictions, tailors the policy to meet particular needs and determines its applicability based on the defined range of actors This range can be specified through various OWL constructs, including enumerations, actor classes, or contextual descriptions Additionally, XML Schema expressions can refine datatype property ranges, allowing for complex policy definitions Ultimately, KAoS policies operate without conditional rules, relying on context restrictions linked to the action class to establish policy relevance in different scenarios.
Figure 7: KAoS KPAT - OWL policy editor and administration tool
An action class categorizes actions that actors plan to execute or are in the process of executing Components like KAoS Guards create RDF descriptions of these action instances to assess policy implications KAoS utilizes the inference capabilities of Stanford University's Java Theorem Prover (JTP) to classify these instances and identify relevant policies Subsequently, KAoS evaluates the relative precedence of the identified policies and organizes them to determine the dominant authorization policy.
When the prevailing authorization is positive, KAoS gathers obligations from any activated obligation policies in a prioritized manner It then conveys the results to relevant stakeholders, typically the enforcement mechanisms tasked with preventing prohibited actions and ensuring the fulfillment of obligations.
Representing policies in OWL enhances the ability to reason about controlled environments, policy relationships, and disclosure, while also aiding in detecting and harmonizing policy conflicts This approach leverages description logic subsumption and instance classification algorithms to analyze domain structures and concepts effectively Additionally, KAoS can identify conflicting policies and harmonize them using specialized algorithms implemented in JTP.
Important KAoS Features
We highlight a few important features of KAoS below:
KAoS policy representation utilizes a homogeneous approach, as all elements are encoded in OWL This allows any third-party tool or environment that supports OWL to conduct specialized analyses of the complete knowledge base independently of KAoS Consequently, this facilitates seamless integration with a growing array of advanced OWL tools and language enhancements in the future.
Maturity Over the past few years, KAoS has been used in a wide range of applications and operating environments.
KAoS offers a comprehensive solution for access control and authorization by supporting both authorization and obligation policies It features a robust policy management infrastructure that includes advanced user interfaces for policy specification and analysis, along with a generic mechanism for policy disclosure Additionally, ongoing developments aim to enhance policy enforcement automation, including the automatic generation of code for enforcers.
The pluggability of the system allows for the seamless integration of platform-specific and application-specific ontologies over core concepts Additionally, the policy enforcement components can be easily adapted to diverse computing environments, including traditional distributed platforms like Web Services, Grid Computing, and CORBA, as well as various software and robotic agent frameworks such as Nomads, Brahms, SFX, CoABS Grid, and Cougaar.
We have significantly enhanced the scalability and performance of our policy disclosure methods, achieving an average response time of under 1 ms for queries from enforcers This efficiency stems from our use of advanced description logic subsumption and classification techniques Additionally, multiple enforcers can execute queries concurrently, allowing KAoS to leverage multiprocessor machines effectively Our rigorous evaluations in the DARPA UltraLog program demonstrate satisfactory performance even in large environments with over a thousand agents and numerous policies Dynamic policy updates can be processed, resolved, and disseminated within seconds, and ongoing improvements in reasoning methods and computer hardware will further boost this performance.
Beyond Description Logic for Policy Representation
Until recently, KAoS relied solely on OWL-DL (originally DAML) for describing policy-governed entities and their actions, benefiting from the semantic richness of OWL that surpasses traditional policy languages This allowed for greater expressivity in policy specification; however, limitations arose when defining policies where one aspect of an action's context depended on another For instance, in loop communication, the source and destination must be identical More complex scenarios, such as restricting the return of calculation results to only those parties that provided the data or were authorized by the data providers, further illustrate the limitations of OWL-DL These intricate action descriptions extend beyond the expressive capabilities of OWL-DL.
Role-value maps have been extensively studied as a crucial missing aspect of representational semantics, highlighting the equality or containment of values derived from two chains of instance properties (McIlraith et al., 2001) The Semantic Web Rule Language (SWRL) supports role-value-map semantics, yet its complex syntax poses challenges Consequently, an OWL-based representation that captures this semantics could be beneficial across various applications For example, OWL-S developers recognized the need for similar dataflow semantics and created a formulation (process:sameValues) to represent these chains, although it was limited to single-chain elements (McIlraith et al., 2001).
KAoS has been enhanced with mechanisms to incorporate role-value-map semantics into defined policy actions via the KAoS Policy Administration Tool Currently, the syntax for this semantics is based on the SWRL OWL ontology, with the code for generating this syntax housed in a specialized Java class for future modifications Our classification algorithm utilizes this information to classify action instances by verifying compliance with the OWL-DL portion of the action class and checking the relevant role-value-map constraints For instance, when determining if an intercepted communication is a loop communication, KAoS assesses whether the source of the current communication matches one of the values associated with the communication's destination.
We are collaborating with Stanford to enhance JTP by integrating subsumption reasoning within role-value-map semantics, enabling more sophisticated policy analyses.
Generic Semantic Web Service Policy Enforcer
KAoS offers customizable enforcers designed for the Semantic Web Services environment, enabling the interception of SOAP messages and filtering of results in accordance with coalition policies Our SOAP-enabled enforcer can comprehend various Semantic Web Service requests, allowing it to enforce relevant authorization policies effectively Furthermore, it includes a mechanism for executing obligation policies through additional Web Service invocations.
For instance, an obligation policy may require the recording of certain kinds of service transactions through a remote logging service.
CoSAR-TS Scenario
Binni Scenario
The Binni domain serves as a platform for multinational research in Command and Control, building on the findings from the Coalition Agents eXperiment (CoAX) This experiment involved 20 organizations from four countries and showcased the application of intelligent agent technology within a coalition framework.
Figure 8: Map of Binni Region of the Red Sea
CoSAR-TS Scenario
The narrative unfolds with a downed airman in the Red Sea, positioned between the fictional nations of Binni and Arabello With precise location knowledge and no local threats, the airman communicates his injuries through suit sensors The investigation explores three rescue options: a US ship-borne helicopter, a helicopter from Gao in Binni, or a patrol boat off Arabello’s coast Arabello’s hospital emerges as the best choice for specialized burn treatment, yet the selection of a rescue resource is limited by policy, preventing Gaoan helicopters from entering Arabello's airspace.
Figure 9: CoSAR-TS demo elements
The project involves various agents, including the Coalition SAR Coordinator, US SAR Officer, hospital information provider, SAR resource provider, and a Notification Agent, each fulfilling specific roles Both the Coalition SAR Coordinator and US SAR Officer utilize an I-X process panel for collaborative messaging and executing standard operating procedures A query is conducted on BBN Technologies' semantic web service to access OWL-encoded data regarding medical infrastructure in different countries The CMU Semantic Matchmaker is employed to identify appropriate SAR resources, which are described based on an ontology developed during the DARPA SONAT experiment, detailing their operational areas and countries of origin Although the current number of registered services in the Matchmaker is limited, the most time-consuming aspect is the loading and preprocessing of the ontologies for service descriptions and queries, adhering to KAoS policies established by authorized personnel via the IHMC KAoS Policy Administration Tool (KPAT) Lastly, the CMU Notification Agent leverages its knowledge of recipients to effectively notify hospital administrators or pilots.
I-K-C
I-X new capabilities supporting I-K-C
The latest enhancements to the I-DE Process editor and I-Plan planning elements enable users to create composed workflows prior to execution This development facilitates the import of OWL-S described services, enriching the existing process library The generated plans can be exported to other enactment systems, utilized for in-depth analysis such as policy compliance checks in KAoS, and assist in the dynamic reactive execution of plans in I-P2 or other tools.
Figure 11: I-Plan Web Service – Search & Rescue example
KAoS new capabilities supporting I-K-C
6.2.1 Mapping the OWL-S Representation of Process to the KAoS Concept of ActionThe OWL-S concept of Process maps semantically to the KAoS concept of Action1.Unfortunately, OWL-S made a dramatic change in representing workflow processes in the transitioning from the earlier ontology called DAML-S In DAML-S, processes were represented as classes whose instances were process executions and whose input and output parameters were defined as properties of those classes Parameter restrictions were represented as range constraints on those parameter properties In contrast, OWL-S represents processes as instances, and parameters are defined as instances of the class Parameter or its subclasses Input and Output, with their corresponding parameter restrictions defined by the value of the process:parameterType property for each parameter This significant change does not allow for a straightforward mapping between OWL-S and KAoS concepts using owl:equivalentClass and owl:equivalentProperty as it had been previously possible in the case of DAML-S OWL-S will define process executions as instances of a ProcessInstance class that refers to its process type. This approach is similar to that taken in the Process Specification Language (PSL; Schlenoff et al., 2000).
To utilize KAoS reasoning capabilities, it is essential to create an OWL class derived from the OWL-S process definition instance This involves modifying the process:parameterType to reflect the necessary restrictions We employ the OWL-S API to load OWL-S process workflows, identify all processes within a workflow, and retrieve detailed definitions Subsequently, we utilize Jena to construct the corresponding OWL class, which serves as a subclass of the KAoS Action class.
6.2.2 KAoS Capabilities for Analyzing Action Classes
After KAoS extracts an action from the workflow and converts it into an action class, it assesses the action's compliance with existing policies Unlike the process of checking authorization and obligations for action instances, workflow policy compliance focuses on action classes, necessitating the use of subsumption reasoning to identify relationships between the current action class and those linked to policies This reasoning is also employed in policy analyses, such as policy deconfliction, which involves uncovering relationships like subsumption or disjointness among action classes related to policies.
Analyses typically yield deterministic conclusions, indicating whether a specific process will be authorized or forbidden, or if it will create an obligation These results remain deterministic when the action class under investigation is a subclass of a single policy action class or a combination of policy action classes that represent either authorization or obligation policies.
In some cases, analyses can yield nondeterministic outcomes, indicating that a process instance may or may not be authorized or may generate obligations This situation arises when the action class related to the process is neither completely included nor entirely separate from a specific policy action class that defines authorization or obligation policies KAoS can create a representation of the action class by calculating the difference between the current action class and the relevant policy action class The algorithm used for this process is the same as the one previously outlined for policy harmonization (Bradshaw et al., 2003), although we are still developing methods to translate this new class into an OWL-S process instance representation.
We have created an initial version of supplementary KAoS ontology components that facilitate workflow annotation based on the outcomes of our policy analyses By utilizing the OWL-S API, we incorporated the necessary markup into the original OWL-S workflow, which was then returned from KAoS to the I- system.
Conclusions
The CoSAR-TS project enhanced AIAI's intelligent planning technology and IHMC's KAoS services with advanced functionalities, ensuring full compliance with OWL standards This collaboration has significantly strengthened the partnership between AIAI and IHMC, with a shared objective of working together on future projects and developing tools that integrate both technologies.
The project enhanced our understanding of Semantic Web technology by applying it to realistic military scenarios It also acted as a testing ground for technologies created by other participants in the DAML program Additionally, our involvement in OWL and OWL-S committees ensured that the valuable lessons learned during the project were integrated into the latest versions of the standards.
In their 2001 paper presented at the 6th International Command and Control Research and Technology Symposium at the US Naval Academy, Allsopp et al explore the role of software agents in enhancing coherent coalition operations The authors discuss how these agents can facilitate communication and coordination among diverse teams, ultimately improving operational effectiveness in military contexts Their research highlights the potential of software agents to streamline processes and support decision-making in complex coalition environments.
Allsopp, D.N., Beautement, P., Bradshaw, J.M., Durfee, E.H., Kirton, M., Knoblock, C.A., Suri,
In their 2002 study, N., Tate, A., and Thompson, C.W explored the dynamics of multi-agent cooperation within international coalitions Their findings were presented at the Second International Conference on Knowledge Systems for Coalition Operations (KSCO-2002) held in Toulouse, France This research emphasizes the importance of collaborative strategies among coalition agents to enhance operational effectiveness in complex international settings.
Allsopp, D., Beautement, P., Kirton, M., Tate, A., Bradshaw, J.M., Suri, N and Burstein, M.
(2003) The Coalition Agents Experiment: Network-Enabled Coalition Operations, Special Issue on Network-enabled Capabilities, Journal of Defence Science, Vol 8, No 3, pp 130-141, September 2003.
BBN (2003) OpenMap™ Open Systems mapping Technology - http://openmap.bbn.com
Bradshaw, J M., Beautement, M Breedy, L Bunch, S Drakunov, P Feltovich, P., Raj, A., Johnson, M., Kulkarni, S., Suri, N & A Uszok (2004) Making agents acceptable to people In
N Zhong & J Liu (Ed.), Intelligent Technologies for Information Analysis: Advances in Agents, Data Mining, and Statistical Learning (pp 361-400) Berlin: Springer Verlag.
In their 2003 paper presented at the Autonomous Agents and Multi-Agent Systems Conference, Bradshaw et al explore the representation and reasoning mechanisms for DAML-based policy and domain services within the KAoS and Nomads frameworks The authors emphasize the significance of these services in enhancing multi-agent systems, showcasing innovative approaches to policy representation that facilitate effective reasoning and decision-making Their research contributes to the development of more robust and adaptable agent systems, reflecting the evolving demands of autonomous technologies.
In their 2002 paper, Buckingham Shum et al present CoAKTinG, which focuses on Collaborative Advanced Knowledge Technologies within grid computing This work was featured at the Second Workshop on Advanced Collaborative Environments during the Eleventh IEEE International Symposium on High Performance Distributed Computing, held in Edinburgh, Scotland, from July 24 to 26.
Conklin, J and Begeman, M L (1988) gIBIS: A hypertext tool for exploratory policy discussion ACM Transactions on Office Information Systems, 4(6), 303-331.
In the 2003 work "Dialog Mapping: Reflections on an Industrial Strength Case Study," Conklin explores the application of dialog mapping within collaborative and educational contexts Featured in the book "Visualizing Argumentation," edited by Kirschner, Buckingham Shum, and Carr, this case study highlights the effectiveness of visual tools in enhancing group sense-making The publication, available through Springer-Verlag, underscores the significance of structured argumentation in improving communication and decision-making processes.
Chen-Burger, Y and Tate, A (2003) Concept Mapping Between Compendium and I-X, Informatics Report Series, University of Edinburgh, EDI-INF-RR-0166, May 2003.
Jabber.org is a pioneering messaging service that utilizes the eXtensible Messaging and Presence Protocol (XMPP) and has been providing free services since 1999 Users can log in using compatible messaging applications listed on the official XMPP website The service is supported by a dedicated infrastructure, volunteers, and open-source contributions from the Prosody team For any login issues, users are advised to check the status page and service notices It is important to review the service policy, and for assistance, users can refer to the FAQ section Additionally, the website archives historical data, including the JDev archives.
McIlraith, S A., Son, T C., & Zeng, H (2001) Semantic Web Services IEEE Intelligent Systems, 46-53.
Khambhampati, S and Srivastava, B (1996) Unifying Classical Planning Approaches, Arizona State University ASU CSE TR 96-006, July 1996.
MacLean, A., Young, R., Bellotti, V and Moran, T (1991) Design space analysis: Bridging from theory to practice via design rationale In Proceedings of Esprit '91, pages 720-730, Brussels, November 1991.
Polyak and Tate (1999) discuss a unified process ontology designed for process-centered organizations in their paper published in Knowledge-Based Systems An earlier version of this research was presented by Polyak as a University of Edinburgh Department of Artificial Intelligence Research paper in 1998 For further details, the document can be accessed at the provided link.
In his 1999 work, "A Coalition Force Scenario ‘Binni — Gateway to the Golden Bowl of Africa’," R.A Rathmell explores strategic planning for coalition forces Presented at the International Workshop on Knowledge-Based Planning in Edinburgh, this paper discusses the geopolitical significance of Binni as a crucial access point in Africa Rathmell's insights contribute to understanding collaborative military strategies and their implications for regional stability The proceedings, edited by A Tate, highlight the importance of knowledge-based approaches in modern warfare.
Schlenoff, C., Gruninger M., Tissot, F., Valois, J., Lubell, J., Lee, J (2000) The Process Specification Language (PSL): Overview and Version 1.0 Specification," NISTIR 6459, National Institute of Standards and Technology, Gaithersburg, MD.
Selvin, A.M (1999) Supporting Collaborative Analysis and Design with Hypertext Functionality, Journal of Digital information, Volume 1 Issue 4.
Selvin, A.M, Buckingham Shum, S.J., Sierhuis, M., Conklin, J., Zimmermann, B., Palus, C., Drath, W., Horth, D., Domingue, J., Motta, E and Li, G (2001) Compendium: Making Meetings into Knowledge Events, Knowledge Technologies 2001, Austin TX, USA, March 4-7, 2001.
Tate, A (1996) The Constraint Model of Plans, Proceedings of the Third International Conference on Artificial Intelligence Planning Systems, (ed Drabble, B.), pp 221-
228, Edinburgh, UK, May 1996, AAAI Press.
Tate, A (1998) Roots of SPAR, in Special Issue on Ontologies, Knowledge Engineering Review, Vol 13(1), March, 1998, Cambridge University Press.
In his 2000 paper presented at the AAAI-2000 Workshop in Austin, Texas, A Tate discusses the frameworks of and , focusing on how to represent plans and synthesized artifacts through a set of constraints This work addresses critical representational issues in real-world planning systems, contributing to the field of artificial intelligence.
In his 2000 paper titled "Intelligible AI Planning," presented at the Twentieth British Computer Society Special Group on Expert Systems International Conference in Cambridge, UK, Tate discusses the significance of developing artificial intelligence systems that are not only effective but also comprehensible to users The focus is on enhancing the intelligibility of AI planning processes, ensuring that stakeholders can understand and trust the decisions made by these systems This work contributes to the broader field of knowledge-based systems and applied artificial intelligence, emphasizing the importance of transparency in AI technologies.
Tate, A., Dalton, J and Levine, J (2000a) “O-Plan: a Web-based AI Planning Agent”, AAAI-
2000 Intelligent Systems Demonstrator, in Proceedings of the National Conference of the American Association of Artificial Intelligence (AAAI-2000), Austin, Texas, USA, August 2000.
Tate, A., Levine, J., Dalton, J and Nixon, A (2001) “Task Achieving Agents on the World Wide Web”, in “Creating the Semantic Web”, Fensel, D., Hendler, J., Liebermann, H and Wahlster, W. (eds.), MIT Press, 2001.
Uszok, A., Bradshaw, J M., Jeffers, R., Suri, N., Hayes, P., Breedy, M R., Bunch, L., Johnson, M., Kulkarni, S., & Lott, J (2003) KAoS policy and domain services: Toward a description- logic approach to policy representation, deconfliction, and enforcement Proceedings of Policy
In the paper presented at the 6th International Conference on Information Fusion (Fusion 2003) in Cairns, Australia, Wark et al (2003) discuss the implementation of dynamic agent systems within the CoAX Binni 2002 experiment The study focuses on the fusion of information through distributed cooperative agents, highlighting the effectiveness of dynamic systems in enhancing collaborative decision-making processes.
All available via http://www.aiai.ed.ac.uk/project/ix/documents/ or http://i-x.info/documents/
Joint AIAI and IHMC Publications
In their 2003 paper presented at the 6th International Conference on Information Fusion in Cairns, Australia, Wark et al explore the implementation of dynamic agent systems within the CoAX Binni 2002 experiment This research focuses on the innovative use of distributed cooperative agents to enhance information fusion processes.
Allsopp, D., Beautement, P., Kirton, M., Tate, A., Bradshaw, J.M., Suri, N and Burstein, M.
(2003) The Coalition Agents Experiment: Network-Enabled Coalition Operations Special Issue on Network-enabled Capabilities, Journal of Defense Science 8:3, September, 2003.
Uszok, A., Bradshaw, J.M., Jeffers, R., Johnson, M., Tate, A., Dalton, J and Aitken, S (2004) Policy and Contract Management for Semantic Web Services, AAAI Spring Symposium, Stanford University, California, USA, March 2004.
In their 2004 paper presented at the Nineteenth National Conference of the American Association of Artificial Intelligence (AAAI), Tate et al explore the development of intelligent agents designed to enhance coalition search and rescue operations The authors emphasize the significance of these agents in providing task support, ultimately improving the efficiency and effectiveness of rescue missions Their research highlights the potential of intelligent systems in real-world applications, showcasing innovative solutions for complex challenges faced during search and rescue efforts.
2004), San Jose, California, USA, July 2004.
Uszok, A., Bradshaw, J.M., Jeffers, R., Johnson, M., Tate, A., Dalton, J and Aitken, S (2004) KAoS Policy Management for Semantic Web Services, IEEE Intelligent Systems, pp 32-41, July/August 2004.