Proceedings of the second international joint conference on Autonomous agents and multiagent systems - AAMAS '03, 2003
Mobile agents have the potential to substantially improve the speed and efficiency with which dis... more Mobile agents have the potential to substantially improve the speed and efficiency with which distributed and heterogeneous data is retrieved. By moving the computation to the data, retrieval times can be reduced by the elimination of unnecessary data transfer. One way to improve a mobile agent system's retrieval efficiency is to incorporate various query optimization techniques (Das et. al., 2002). These methods involve re-writing of the query execution graph so each mobile agent retrieves its requested data in an optimized order, thus minimizing total data transfer size. While these query re-writing methods can be highly effective in reducing both retrieval times and data transfer sizes, they are generally "static", in that the mobile agents retrieve data in a particular order based on an itinerary that is fixed at the time the plan is generated. We have developed a system by which the advantages of mobile agents are leveraged to optimize data retrieval by dynamically optimizing the retrieval strategy as it is carried out. This strategy equips each spawned agent with the full query execution graph and necessary code to execute the retrieval plan at any data site in the network. The spawned agents communicate and collaborate with each other to dynamically decide where to migrate, send data, and perform necessary computations. These decisions depend on retrieval factors such as network speed, data size, and the computational capabilities of the data servers involved in the retrieval. The feasibility of the approach has been demonstrated within a local area network environment using Earth Science data and we present some experimental results in this context.
Trustworthy situation assessment via belief networks
Proceedings of the Fifth International Conference on Information Fusion. FUSION 2002. (IEEE Cat.No.02EX5997)
Abstract - A probabilistic approach to truth maintenance is presented, specifically geared for au... more Abstract - A probabilistic approach to truth maintenance is presented, specifically geared for automated problem solvers based on Bayesian Belief Network (BN) technology. Nodes and links in a BN capture semantic relationships among various domain related concepts. In ...
Computational Approaches to Situation Assessment and Decision Support
Proceedings of the second international joint conference on Autonomous agents and multiagent systems, 2003
Mobile agents have the potential to substantially improve the speed and efficiency with which dis... more Mobile agents have the potential to substantially improve the speed and efficiency with which distributed and heterogeneous data is retrieved. By moving the computation to the data, retrieval times can be reduced by the elimination of unnecessary data transfer. One way to improve a mobile agent system's retrieval efficiency is to incorporate various query optimization techniques (Das et. al., 2002). These methods involve re-writing of the query execution graph so each mobile agent retrieves its requested data in an optimized order, thus minimizing total data transfer size. While these query re-writing methods can be highly effective in reducing both retrieval times and data transfer sizes, they are generally "static", in that the mobile agents retrieve data in a particular order based on an itinerary that is fixed at the time the plan is generated. We have developed a system by which the advantages of mobile agents are leveraged to optimize data retrieval by dynamically optimizing the retrieval strategy as it is carried out. This strategy equips each spawned agent with the full query execution graph and necessary code to execute the retrieval plan at any data site in the network. The spawned agents communicate and collaborate with each other to dynamically decide where to migrate, send data, and perform necessary computations. These decisions depend on retrieval factors such as network speed, data size, and the computational capabilities of the data servers involved in the retrieval. The feasibility of the approach has been demonstrated within a local area network environment using Earth Science data and we present some experimental results in this context.
We describe here an agent-based Distributed Analytical Search (DAS) tool to search and query dist... more We describe here an agent-based Distributed Analytical Search (DAS) tool to search and query distributed “big data” sources regardless of data’s location, content or format. DAS semantically analyzes natural language queries from a web-based user interface. It automatically translates the query to a set of sub-queries by deploying a combination of planning and traditional database query optimization techniques. It then generates a query plan represented in XML and guide the execution by spawning intelligent agents with various types of wrappers as needed for distributed sites. The answers returned by the agents are merged appropriately and return them to the user. We have demonstrated DAS using a variety of data sources that are distributed and heterogeneous. The tool is the prime product of our company with big enterprises as our target market.
Specifying Deductive Databases and Integrity Constraints in Meta-logic
Workshops in Computing, 1992
A meta-logic statement over a language is characterised by a formula in typed ω-order λ-calculus.... more A meta-logic statement over a language is characterised by a formula in typed ω-order λ-calculus. The head of the λ-term corresponding to a meta-logic statement is always a constant when it is in normal form. It is shown that both deductive database statements and integrity constraints, particularly those involving aggregate operations, can be represented conveniently in meta-logic statements. A special case of Huet’s unification algorithm is considered on the domain of meta-logic terms and a proof procedure based on this unification is presented to answer database queries and to verify integrity constraints. Although, the main focus is to handle deductive databases problems, the results of this paper deal indirectly with a metalevel extension of first-order logic programming.
Proceedings of the second international conference on Autonomous agents - AGENTS '98, 1998
In this paper, we present an event based agent architecture for increased agent autonomy in dynam... more In this paper, we present an event based agent architecture for increased agent autonomy in dynamic environments. Our approach is based on an event description language called MDL that has been developed to facilitate application modeling by associating events with concepts such as objects and attributes. Events can model a wide variety of real-world scenarios -ranging from database transactions to user interactions. Our architecture employs a novel learning service for the integration of event sequences to produce facts about the environment for increased agent autonomy. This generic architccturc has been successfully instantiated in a number of research and commercial application domains of which the following two arc described in this paper: 1) A spacecraft data analysis agent that identifies recurring patterns within spacecraft telemetry data to characterize normal spacecraft operations, 2) A Macintosh desktop interface agent that finds repetitive user patterns by observing user actions in the background and automating upon approval. 1.1 Keywords Soflwarc agent, user modeling, learning, autonomy I'ClYlllS!!lOll 10 tllnlN digilnl/lmrd copies of all or pnrl of this nlaterinl I&pcrwollnl or chssroow USC is gmntcd witbout fee provided hat IIIC copies nrc IIO~ mdc or dislributed for prolit or conmlercinl advantage, cl~r: copyright IIO~~CC, 1111: title of~hc poblicntion nnd it$ date nppcar. nod notice ii ~iVcl1 W copyiug is by pen~hsio~~ ol'AC!M. Inc. l'o copy oh-wise. IO rcpublith, IO posl WI servem or IO redistribule IO lists. require prior hpociliu pemissb) nttdhr I&.
Modal logics are often criticised for their coarse grain representation of knowledge of possibili... more Modal logics are often criticised for their coarse grain representation of knowledge of possibilities about assertions. That is to say, if two assertions are possible in the current world, their further properties are indistinguishable in the modal formalism even if an agent knows that one of them is true in twice as many possible worlds as compared to the other one. Epistemic logic, that is the logic of knowledge and belief, cannot avoid this shortcomings because it inherits the syntax and semantics of modal logics. In this paper, we develop an extended formalism of modal epistemic logic which will allow an agent to represent its degrees of support about an assertion. The degrees are drawn from qualitative or quantitative dictionaries which are accumulated from agent's a priori knowledge about the application domain. A possible-world semantics of the logic is developed by using the accessibility hyperelation and the soundness and completeness results are stated. The abstract syntax and semantics are illustrated and motivated by an example from the medical domain.
We present a Network-based Truth Maintenance System (NTMS) for problem solvers based on Bayesian ... more We present a Network-based Truth Maintenance System (NTMS) for problem solvers based on Bayesian belief network (BN) technology. BN technology has been proven to be effective in various domains, e.g. assessing battlefield situations, such as the enemy's likely point of interdiction. Nodes and links in a BN capture semantic relationships among various domain related concepts. In the absence of firmer knowledge, default assumptions provide the beliefs of some nodes in a BN. Before posting incoming evidence into a BN node, a truth maintenance procedure is invoked to check for information consistency between the node's current expected state and the new observed state. In case of inconsistency, the truth maintenance procedure revises some default assumptions, by isolating those nodes causing inconsistency, via a sensitivity analysis procedure that exploits the strengths of BN causal dependency. We have applied our approach for trustworthy situation assessment in the context of...
Journal of Experimental & Theoretical Artificial Intelligence, 1997
Page 1. A exible architecture for autonomous agents SK Das IC-Parc William Penney Laboratory Impe... more Page 1. A exible architecture for autonomous agents SK Das IC-Parc William Penney Laboratory Imperial College London SW7 2AZ England [email protected] tel: +44 71 594 8424 J. Fox Advanced Computation Laboratory ...
Mobile Agents for Distributed and Heterogeneous Information Retrieval
Information Retrieval, 2005
Abstract. The heterogeneous, distributed and voluminous nature of many government and corporate d... more Abstract. The heterogeneous, distributed and voluminous nature of many government and corporate data sources impose severe constraints on meeting the diverse requirements of users who analyze the data. Additionally, communication bandwidth limitations, time constraints, and multiple ...
A path finding method for constraint checking in deductive databases
Data & Knowledge Engineering, 1989
A method is presented for checking integrity constraints in a deductive database in which verific... more A method is presented for checking integrity constraints in a deductive database in which verification of the integrity constraints in the updated database is reduced to the process of constructing paths from update literals to the heads of integrity constraints. In constructing such paths, the ...
Safe and Sound: Artificial Intelligence in Hazardous Applications
Artificial Intelligence in Medicine, 2003
Computer science and artificial intelligence are increasingly used in the hazardous and uncertain... more Computer science and artificial intelligence are increasingly used in the hazardous and uncertain realms of medical decision making, where small faults or errors can spell human catastrophe. This book describes, from both practical and theoretical perspectives, an AI ...
2005 7th International Conference on Information Fusion
We present and demonstrate a particle filtering approach to data fusion and situation assessment ... more We present and demonstrate a particle filtering approach to data fusion and situation assessment for military operations in urban environments. Our approach views such an environment as a physical system whose state vector is composed of a large number of both discrete and continuous variables representing properties of tracked entities. Inference on such vector-based models exploits both causal dependencies among variables in the state vector via its dynamic Bayesian belief network representation and vector decomposition into weakly interacting subcomponents. To effectively leverage the decomposition, instead of straightforward particle filtering, the proposed algorithm maintains factored particles over clusters of state variables, thus resulting in smaller variance. The algorithm samples discrete modes and approximates the continuous variables by a multi-normal distribution updated at each time step by an unscented Kalman filter. The approach is demonstrated using a Marine Corps operational scenario involving a potential ambush on city streets.
2005 7th International Conference on Information Fusion
We present a generic argumentation-based framework for making decisions under uncertainty by fusi... more We present a generic argumentation-based framework for making decisions under uncertainty by fusing knowledge from multiple sources. In this framework, arguments for decision options are expressed in a high-level knowledge representation language incorporating subjective probabilities from decision makers representing the argument strengths. To aggregate a set of such probabilistic arguments for and against the decision options, we apply Dempster-Shafer theory to compute degrees of belief for decision options. Evidence converted from the underlying knowledge base is used to compute degrees of belief, and hence rankings, among the decision options. Decision-making based on such degrees of belief is therefore based on a strong mathematical foundation. The proposed decision making framework has been successfully applied in a variety of domains ranging from theater missile defense and army fire support to medical diagnosis.
Many dynamic systems involve a number of entities that are largely independent of each other but ... more Many dynamic systems involve a number of entities that are largely independent of each other but interact with each other via a subset of state variables. We present global/local dynamic models (GLDMs) to capture these kinds of systems. In a GLDM, the state of an entity is decomposed into a globally influenced state that depends on other entities, and a locally influenced state that depends only on the entity itself. We present an inference algorithm for GLDMs called global/local particle filtering, that introduces the principle of reasoning globally about global dynamics and locally about local dynamics. We have applied GLDMs to an asymmetric urban warfare environment, in which enemy units form teams to attack important targets, and the task is to detect such teams as they form. Experimental results for this application show that global/local particle filtering outperforms ordinary particle filtering and factored particle filtering.
Proceedings of the first international conference on Autonomous agents - AGENTS '97, 1997
A generic architecture for autonomous agents is presented. In common with other current proposals... more A generic architecture for autonomous agents is presented. In common with other current proposals the agent is capable of reacting to and reasoning about events which occur in its environment, executing actions and plans in order to achieve goals in the environment. In addition the agent can make decisions under uncertainty, including decisions about competing beliefs and alternative actions. The framework is grounded in a nonclassical decision model; this supports many more capabilities than classical decision theory but under restricted conditions is compatible with it. The model is embodied in a well-dened knowledge representation language, R 2 L, which explicitly supports the central concepts of decisions and plans, and associated constructs of goals, arguments and commitments. The language provides a sound basis for building knowledge based agents for practical applications including safety-critical ones. This is illustrated with examples of medical applications.
Uploads
Papers by Subrata Das