Results 1  10
of
20
Playing games for security: An efficient exact algorithm for solving bayesian stackelberg games
 In Proceedings of the 7th Int. Conf. on Autonomous Agents and Multiagent Systems (AAMAS 2008
, 2008
"... See next page for additional authors Follow this and additional works at: ..."
Abstract

Cited by 76 (27 self)
 Add to MetaCart
(Show Context)
See next page for additional authors Follow this and additional works at:
An efficient heuristic approach for security against multiple adversaries
 IN AAMAS
, 2007
"... In adversarial multiagent domains, security, commonly defined as the ability to deal with intentional threats from other agents, is a critical issue. This paper focuses on domains where these threats come from unknown adversaries. These domains can be modeled as Bayesian games; much work has been do ..."
Abstract

Cited by 46 (21 self)
 Add to MetaCart
(Show Context)
In adversarial multiagent domains, security, commonly defined as the ability to deal with intentional threats from other agents, is a critical issue. This paper focuses on domains where these threats come from unknown adversaries. These domains can be modeled as Bayesian games; much work has been done on finding equilibria for such games. However, it is often the case in multiagent security domains that one agent can commit to a mixed strategy which its adversaries observe before choosing their own strategies. In this case, the agent can maximize reward by finding an optimal strategy, without requiring equilibrium. Previous work has shown this problem of optimal strategy selection to be NPhard. Therefore, we present a heuristic called ASAP, with three key advantages to address the problem. First, ASAP searches for the highestreward strategy, rather than a BayesNash equilibrium, allowing it to find feasible strategies that exploit the natural firstmover advantage of the game. Second, it provides strategies which are simple to understand, represent, and implement. Third, it operates directly on the compact, Bayesian game representation, without requiring conversion to normal form. We provide an efficient Mixed Integer Linear Program (MILP) implementation for ASAP, along with experimental results illustrating significant speedups and higher rewards over other approaches.
Support for Situation Awareness in Command and Control
 In Proc. of the Seventh Int. Conf. on Information Fusion (FUSION 2004
, 2004
"... www.nada.kth.se/~klasw Abstract – To support the work by military commanders and staffs, this paper presents a classification of different support tools, including Command Support, Decision Support, MultiSensor Fusion, and Information Fusion. Together, these tools would facilitate shared Situation ..."
Abstract

Cited by 8 (2 self)
 Add to MetaCart
(Show Context)
www.nada.kth.se/~klasw Abstract – To support the work by military commanders and staffs, this paper presents a classification of different support tools, including Command Support, Decision Support, MultiSensor Fusion, and Information Fusion. Together, these tools would facilitate shared Situation Awareness, including knowledge of friendly as well as hostile, forces on different levels of abstraction. Externalizing the situation awareness, a Common Situation Model enables the integration of these support tools but also the collaboration between individuals. The main emphasis is thus on the ontology of this model, representing the language by which the situation may be described. Consequently, a UML Class Diagram is proposed that expresses a conceptual model of the Command and Control domain. This model is generic since it could be adapted to different situations.
Robust Bayesianism: Relation to evidence theory
 J. Advances in Information Fusion
"... We are interested in understanding the relationship between Bayesian inference and evidence theory. The concept of a set of probability distributions is central both in robust Bayesian analysis and in some versions of DempsterShafer’s evidence theory. We interpret imprecise probabilities as impreci ..."
Abstract

Cited by 7 (1 self)
 Add to MetaCart
(Show Context)
We are interested in understanding the relationship between Bayesian inference and evidence theory. The concept of a set of probability distributions is central both in robust Bayesian analysis and in some versions of DempsterShafer’s evidence theory. We interpret imprecise probabilities as imprecise posteriors obtainable from imprecise likelihoods and priors, both of which are convex sets that can be considered as evidence and represented with, e.g., DSstructures. Likelihoods and prior are in Bayesian analysis combined with Laplace’s parallel composition. The natural and simple robust combination operator makes all pairwise combinations of elements from the two sets representing prior and likelihood. Our proposed combination operator is unique, and it has interesting normative and factual properties. We compare its behavior with other proposed fusion rules, and earlier efforts to reconcile Bayesian analysis and evidence theory. The behavior of the robust rule is consistent with the behavior of Fixsen/Mahler’s modified Dempster’s (MDS) rule, but not with Dempster’s rule. The Bayesian framework is liberal in allowing all significant uncertainty concepts to be modeled and taken care of and is therefore a viable, but probably not the only, unifying structure that can be economically taught and in which alternative solutions can be modeled, compared and explained. Manuscript received April 20, 2006; released for publication April
Realization of a bridge between highlevel information need and sensor management using a common dbn
 in The 2004 IEEE International Conference on Information Reuse and Integration (IEEE IRI2004), IEEE
, 2004
"... In a decision support system for military decision makers a plan recognition process provides estimates of enemy plans. To respond to a changing and uncertain environment the plan recognition process requires timely and relevant information. We address the rarely discussed, yet crucial, issue of con ..."
Abstract

Cited by 6 (4 self)
 Add to MetaCart
(Show Context)
In a decision support system for military decision makers a plan recognition process provides estimates of enemy plans. To respond to a changing and uncertain environment the plan recognition process requires timely and relevant information. We address the rarely discussed, yet crucial, issue of connecting the information needs of plan recognition to management of sensors. We have previously presented a framework for this purpose and here we give details of an implementation and provide some results. In our implementation both plan recognition, sensor management and the functions that connect them utilize the a priori knowledge stored in a Dynamic Bayesian Network. 1
Robust Bayesianism: Imprecise and Paradoxical Reasoning
, 2004
"... We are interested in understanding the relationship between Bayesian inference and evidence theory, in particular imprecise and paradoxical reasoning. The concept of a set of probability distributions is central both in robust Bayesian analysis and in some versions of DempsterShafer theory. Most of ..."
Abstract

Cited by 6 (1 self)
 Add to MetaCart
We are interested in understanding the relationship between Bayesian inference and evidence theory, in particular imprecise and paradoxical reasoning. The concept of a set of probability distributions is central both in robust Bayesian analysis and in some versions of DempsterShafer theory. Most of the literature regards these two theories as incomparable. We interpret imprecise probabilities as imprecise posteriors obtainable from imprecise likelihoods and priors, both of which can be considered as evidence and represented with, e.g., DSstructures. The natural and simple robust combination operator makes all pairwise combinations of elements from the two sets. The DSstructures can represent one particular family of imprecise distributions, Choquet capacities. These are not closed under our combination rule, but can be made so by rounding. The proposed combination operator is unique, and has interesting normative and factual properties. We compare its behavior on Zadeh's example with other proposed fusion rules. We also show how the paradoxical reasoning method appears in the robust framework.
An Information Fusion Game Component
, 2004
"... Higher levels of the data fusion process call for prediction and awareness of the development of a situation. Since the situations handled by command and control systems develop by actions performed by opposing agents, pure probabilistic or evidential techniques are not fully sufficient tools for pr ..."
Abstract

Cited by 4 (3 self)
 Add to MetaCart
(Show Context)
Higher levels of the data fusion process call for prediction and awareness of the development of a situation. Since the situations handled by command and control systems develop by actions performed by opposing agents, pure probabilistic or evidential techniques are not fully sufficient tools for prediction. Gametheoretic tools can give an improved appreciation of the real uncertainty in this prediction task, and also be a tool in the planning process. Based on a combination of graphical inference models and game theory, we propose a decision support tool architecture for command and control situation awareness enhancements. This paper outlines a framework for command and control decisionmaking in multiagent settings. Decisionmakers represent beliefs over models incorporating other decisionmakers and the state of the environment. When combined, the decisionmakers’ equilibrium strategies of the game can be inserted into a representation of the state of the environment to achieve a joint probability distribution for the whole situation in the form of a Bayesian network representation.
GameTheoretic Reasoning and Command and Control
 In Proceedings of the 15 th MiniEURO Conferences: Managing Uncertainty in Decision Support Models
, 2004
"... ABSTRACT. Developers of tomorrow’s Command and Control centers are facing numerous problems related to the vast amount of available information obtained from various sources. On a lower level, huge amounts of uncertain reports from different sensors need to be fused into comprehensible information. ..."
Abstract

Cited by 3 (0 self)
 Add to MetaCart
(Show Context)
ABSTRACT. Developers of tomorrow’s Command and Control centers are facing numerous problems related to the vast amount of available information obtained from various sources. On a lower level, huge amounts of uncertain reports from different sensors need to be fused into comprehensible information. On a higher level, representation and management of the aggregated information will be the main task, with the overall objective to provide reliable and comprehensible situation awareness to commanders. Hence, we consider prediction of future course of events being a necessary ingredient. Unfortunately, traditional agent modeling techniques do not capture gaming situations, i.e., situations where commanders make decisions based on other commanders’ reasoning about one’s own reasoning. To cope with this problem, we propose an architecture based on game theory for inference, coupled with traditional methods for uncertainty modeling. Applying an example, we show that our architecture could be used as a decision support tool, offering enhanced situation awareness in Command and Control. Finally, we wind up with a philosophical discussion regarding the ambiguities and the difficulties in interpreting the solution that game theory offers in the form of mixed strategy Nash equilibria. 1.
An Efficient Heuristic for Security Against Multiple Adversaries in Stackelberg
"... In adversarial multiagent domains, security, commonly defined as the ability to deal with intentional threats from other agents, is a critical issue. This paper focuses on domains where these threats come from unknown adversaries. These domains can be modeled as Bayesian games; much work has been do ..."
Abstract

Cited by 1 (1 self)
 Add to MetaCart
(Show Context)
In adversarial multiagent domains, security, commonly defined as the ability to deal with intentional threats from other agents, is a critical issue. This paper focuses on domains where these threats come from unknown adversaries. These domains can be modeled as Bayesian games; much work has been done on finding equilibria for such games. However, it is often the case in multiagent security domains that one agent can commit to a mixed strategy which its adversaries observe before choosing their own strategies. In this case, the agent can maximize reward by finding an optimal strategy, without requiring equilibrium. Previous work has shown this problem of optimal strategy selection to be NPhard. Therefore, we present a heuristic called ASAP, with three key advantages to address the problem. First, ASAP searches for the highestreward strategy, rather than a BayesNash equilibrium, allowing it to find feasible strategies that exploit the natural firstmover advantage of the game. Second, it provides strategies which are simple to understand, represent, and implement. Third, it operates directly on the compact, Bayesian game representation, without requiring conversion to normal form. We provide an efficient Mixed Integer Linear Program (MILP) implementation for ASAP, along with experimental results illustrating significant speedups and higher rewards over other approaches.
Contagion and observability in security domains
 In Proceedings of the 51st Annual Allerton Conference on Communication, Control, and Computing (Allerton), 1364–1371. IEEE
"... We examine security domains where defenders choose their security levels in the face of a possible attack by an adversary who attempts to destroy as many of them as possible. Though the attacker only selects one target, and only has a certain probability of destroying it depending on that defender’s ..."
Abstract

Cited by 1 (0 self)
 Add to MetaCart
(Show Context)
We examine security domains where defenders choose their security levels in the face of a possible attack by an adversary who attempts to destroy as many of them as possible. Though the attacker only selects one target, and only has a certain probability of destroying it depending on that defender’s security level, a successful attack may infect other defenders. By choosing a higher security level the defenders increase their probability of survival, but incur a higher cost of security. We assume that the adversary observes the security levels chosen by the defenders before selecting whom to attack. We show that under this assumption the defenders overprotect themselves, exhausting all their surplus, so optimal policy requires taxing security, as opposed to the subsidies recommended by alternative models for contagious attacks which do not take into account the attacker’s ability to observe the defenders ’ choices. 1