Eugene Santos

… on Intelligent Environments, 1998

The utilization of immersive distributed virtual en-vironments for military training is readily a... more

Dynamic user model construction with bayesian networks for intelligent information queries

The complexity of" current soft.ware applications is overwhelming users. The need exists for inte... more The complexity of" current soft.ware applications is overwhelming users. The need exists for inteUigent interface agents to address the problems of increasing taskload that is overwhelming the hunloax user. Interface agents could help alleviate user taskload by extracting and analyzing relevant information, and providing information abstractions of that informatiou, and providing timely, beneficial assistance to users. Central to providing assistance to a user is the issue of correctly determining the user's intent. The C.lavin project is to build an intelligent natural language query information management system. Clavin must mMntain a dynamic user model of the relevant concepts in the user inquiries as they relate to the information sources. The primary goal of Clavin is to antonomously react to changes in user intent as well as the information sources, by dynamically constructing the appropriate queries relative to the changes identified. In this paper, we discuss the problems and issues that arise in achieving user-intent ascription through dynamic user model construction with Bayesian networks.

Inferencing Over Incomplete Solution Spaces with Genetic Algorithms for Probabilistic Reasoning

by Brett Borghetti and Eugene Santos

We develop a new method for handling incomplete knowledge when inferencing over Bayesian Knowledg... more We develop a new method for handling incomplete knowledge when inferencing over Bayesian Knowledge Bases(BKB) using genetic algorithms(GA) . The fitness function for a genetic algorithm requires that we give a score to each solution it generates. When a solution is complete, we can use the joint probability of our generated solution, but when incompleteness occurs, the joint probability is undefined. In this paper, we characterize incompleteness in BKBs and present an easily computable method for scoring a solution which contains incompleteness. 1 Introduction A Bayesian Knowledge Base (BKB) is a knowledge representation for information containing uncertainty. BKBs use random variables(RVs) and probabilities to represent knowledge. Unlike other approaches such as Bayesian Networks[4], BKBs are designed to handle incompleteness in our knowledge 2 . One of the inferencing tasks performed on BKBs is belief revision. In belief revision, we attempt to find the most likely state for every...

Intra-Organizational Logistics Management Through Multi-Agent Systems

A Performance Study on a Multiagent E-Scheduling and Coordination Framework for Maintenance Networks

IEEE Transactions on Systems, Man and Cybernetics, Part C (Applications and Reviews), 2000

Many maintenance networks as well as supply networks and virtual enterprises consist of multiple ... more Many maintenance networks as well as supply networks and virtual enterprises consist of multiple organizations. A common problem arising in these different domains is multiorganization scheduling and coordination. The traditional centralized methods are not appropriate because of the existence of private information and decision-making authorities at different organizations. Although many distributed mechanisms have been presented for supply networks and virtual enterprises, they may not be effective for maintenance networks because of the difficulties of scheduling the tightly related maintenance operations and handling the massive uncertainties involved. These difficulties as well as the heterogeneity of the distributed environment make it challenging to develop an efficient framework for maintenance networks that can obtain a high-quality solution under different conditions, schedule in a timely manner, solve large-scale problems, and so on. In this paper, a price-based multiagent scheduling and coordination framework for maintenance networks is explored, and a systematic experimental study is carried out to evaluate the effects of different factors on its performance. The results show that the framework is able to overcome these difficulties and could be a step toward the next generation of e-scheduling for maintenance networks.

A Massively Parallel Self-Tuning Context-Free Parser

Analyzing the Impact of Knowledge on Algorithm Performance in Discrete Optimization

For discrete optimization, it is important to understand the relationship between knowledge about... more For discrete optimization, it is important to understand the relationship between knowledge about the characteristics of problems and algorithm performance. Such an understanding will provide us useful hints and better prediction of algorithm behavior when we choose or design algorithms. In this paper, we seek to formally model and analyze the impact of such knowledge on algorithm performance/behavior. We propose

Large-Scale Multi-Agent Distributed Mission Planning and Execution in Complex Dynamic Environments

... 22. Santos, Eugene, Jr., Zhang, Feng, and Luh, Peter В., "Multi-Agent Logistics Manageme... more

The Parsing and Learning System(PALS) is a massively parallel self-tuning context-free parser. It... more The Parsing and Learning System(PALS) is a massively parallel self-tuning context-free parser. It is capable of parsing sentences of unbounded length mainly due to its parse-tree representation scheme. The system is capable of improving its parsing performance through the presentation of training examples.

On the Generation of Alternative Explanations with Implications for Belief Revision

The Multi-Agent Distributed Goal Satisfaction System

In general, the best explanation for a given observation makes no promises on how good it is with... more In general, the best explanation for a given observation makes no promises on how good it is with respect to other alternative explanations. A major de ciency of message-passing schemes for belief revision in Bayesian networks is their inability to generate alternatives beyond the second best. In this paper, we present a general approach based on linear constraint systems that naturally generates alternative explanations in an orderly and highly e cient manner. This approach is then applied to cost-based abduction problems as well as belief revision in Bayesian networks.

Dynamic MAP Calculations for Abduction

We present a dynamic algorithm for MAP calculations. The algorithm is based upon Santos's techniq... more We present a dynamic algorithm for MAP calculations. The algorithm is based upon Santos's technique (Santos 1991b) of transforming minimal-cost-proof problems into linearprogramming problems. The algorithm is dynamic in the sense that it is able to use the results from an earlier, near by, problem to lessen its search time. Results are presented which clearly suggest that this is a powerful technique for dynamic abduction problems.

Cost-Based Abduction, Constraint Systems and Alternative Generation

Cost-based abduction attempts to prove an observation by assuming some set of hypotheses. Hypothe... more Cost-based abduction attempts to prove an observation by assuming some set of hypotheses. Hypotheses have associated cost and the best proof is the one which assumes the least costly set. Previous approaches have formalized minimum cost-based abduction as a heuristic search problem. However, e cient admissible heuristics seem to be di cult to nd. In this paper, we present an approach using linear constraints to represent causal relationships providing a new framework for modeling cost-based abduction. Moreover, we are able to utilize the highly e cient optimization tools of operations research, thus yielding a computationally e cient method for solving cost-based abduction problems. Experimentation suggests that our approach is superior to existing search techniques in that our approach exhibits a slower run-time growth rate. This formulation also permits the natural generation of alternative explanations.

AAAI-2002 Fall Symposium Series

In Proc. of the 11th Conf on Computer Generated Forces and Behavioral Representation, May, 2002, Orlando FL

Detecting Deception in Intelligent Systems I: Activation of Deception Detection Tactics

Military and domestic security analysts and planners are facing threats whose asymmetric nature w... more Military and domestic security analysts and planners are facing threats whose asymmetric nature will sharply increase the challenges of establishing an adversary's intent. This complex environment will severely limit the capabilities of the classic doctrinal approach to diagnose adversary activity. Instead, a more dynamic approach is required - adversary decision modeling (ADM) - that, while a critical capability, poses a range of daunting technological challenges. We are developing methodologies and tools that represent a tractable approach to ADM using intelligent software-based analysis of adversarial intent. In this paper we present work being performed by our team (University of Connecticut, Lockheed Martin Advanced Technology Laboratories, and the Air Force Research Laboratory Human Effectiveness Directorate) toward a preliminary composite theory of adversary intent and its descriptive models, to provide a coherent conceptual foundation for addressing adversary decision pr...

On a framework for the prediction and explanation of changing opinions

2009 IEEE International Conference on Systems, Man and Cybernetics, 2009

... 37, no. 6, pp. 11131130, 2007. [5] T. Shipman, Obama supporter warns Clinton over 'rac... more

Lecture Notes in Computer Science, 2004

Independence Semantics for BKBs

<title>Overview of the Distributed Mission Training Integrated Threat Environment project</title>

Modeling and Simulating Sensory Response for Real and Virtual Environments, 1998

Bayesian Knowledge Bases (BKB) are a rule-based probabilistic model that extend Bayes Networks (B... more Bayesian Knowledge Bases (BKB) are a rule-based probabilistic model that extend Bayes Networks (BN), by allowing context-sensitive independence and cycles in the directed graph. BKBs have probabilistic semantics, but lack independence semantics, i.e., a graphbased scheme determining what independence statements are sanctioned by the model. Such a semantics is provided through generalized dseparation, by constructing an equivalent BN. While useful for showing correctness, the construction is not practical for decision algorithms due to exponential size. Some results for special cases, where independence can be determined from polynomial-time tests on the BKB graph, are presented.

Parallel complexity for solving tridiagonal linear systems with multiple right-hand sides on 2-D torus interconnection networks

ABSTRACT

Active User Interfaces

The current state of user interfaces for large information spaces imposes an unmanageable cogniti... more The current state of user interfaces for large information spaces imposes an unmanageable cognitive burden upon the user. Determining how to get the right information into the right form with the right tool at the right time has become a monumental task. While advances in graphical user interfaces can partially address this problem, the basic problem of information overload{and under load{can not be solely addressed through development of a better direct manipulation interface to the information space. We survey the state of the art in two research elds, interface agents and user modeling, that address the problem of information overload and under load. First, we take a historical look at how the elds of human-computer interaction and arti cial intelligence have viewed interface agent research. Interface agents address the problem of increasing task load by serving as either an assistant or associate, extracting and analyzing relevant information, providing information abstractions of that information, and providing timely, bene cial assistance to users. Interface agents communicate with the user through the existing user interface and also adapt to user needs and behaviors. User modeling is concerned with how to represent users' knowledge and interaction within a system to adapt the system to the needs of users. The inclusion of a user model within the overall system architecture allows the system to adapt its response to the preferences, biases, expertise level, goals and needs. We provide a comprehensive look at user modeling to include the elicitation, speci cation and design, veri cation and validation, and utilization of user models. Even though interface agents have made some progress towards addressing the data overload and under load problem, there is active debate about whether agents or direct manipulation interfaces provide better access to information and services. Our paper culminates with a discussion on active user interfaces. Active user interfaces are the next logical step towards addressing the information overload and under load problem. Whereas interface agents are designed around the assistant and/or associate metaphor, active user interfaces serve as an actuator to the human-machine interface. The user of an active user interface is fully aware of any actions, whether explicit (authorized consent) or implicit (implied consent), taken by the active user interface and has a complete, intuitive understanding of such actions. We provide examples of rst-generation active user interfaces within the domain of an natural language query information management system and decision support system and show how the interface is capable of providing an intuitive, meaningful, and useful \actuation" interface for a heterogeneous collection of information sources and tools. The primary goal of active user interfaces is to autonomously react to changes in user intent as well as the information sources, providing a way for users to interact naturally with the information. As a direct result, the user's task load can be reduced.

Intelligent interface agents for intelligent environments

… on Intelligent Environments, 1998

The utilization of immersive distributed virtual en-vironments for military training is readily a... more

Dynamic user model construction with bayesian networks for intelligent information queries

Inferencing Over Incomplete Solution Spaces with Genetic Algorithms for Probabilistic Reasoning

by Brett Borghetti and Eugene Santos

Intra-Organizational Logistics Management Through Multi-Agent Systems

A Performance Study on a Multiagent E-Scheduling and Coordination Framework for Maintenance Networks

IEEE Transactions on Systems, Man and Cybernetics, Part C (Applications and Reviews), 2000

A Massively Parallel Self-Tuning Context-Free Parser

Analyzing the Impact of Knowledge on Algorithm Performance in Discrete Optimization

Large-Scale Multi-Agent Distributed Mission Planning and Execution in Complex Dynamic Environments

... 22. Santos, Eugene, Jr., Zhang, Feng, and Luh, Peter В., "Multi-Agent Logistics Manageme... more

On the Generation of Alternative Explanations with Implications for Belief Revision

The Multi-Agent Distributed Goal Satisfaction System

Dynamic MAP Calculations for Abduction

Cost-Based Abduction, Constraint Systems and Alternative Generation

AAAI-2002 Fall Symposium Series

In Proc. of the 11th Conf on Computer Generated Forces and Behavioral Representation, May, 2002, Orlando FL

Detecting Deception in Intelligent Systems I: Activation of Deception Detection Tactics

On a framework for the prediction and explanation of changing opinions

2009 IEEE International Conference on Systems, Man and Cybernetics, 2009

... 37, no. 6, pp. 11131130, 2007. [5] T. Shipman, Obama supporter warns Clinton over 'rac... more

Lecture Notes in Computer Science, 2004

Independence Semantics for BKBs

<title>Overview of the Distributed Mission Training Integrated Threat Environment project</title>

Modeling and Simulating Sensory Response for Real and Virtual Environments, 1998

Parallel complexity for solving tridiagonal linear systems with multiple right-hand sides on 2-D torus interconnection networks

ABSTRACT

Active User Interfaces