Maksim Tsvetovat
George Mason University, Computational Social Science, Faculty Member
Multi-agent systems can be effective tools for reasoning about human and group behavior. Their effectiveness is enhanced when the algorithms lead the simulated agents to behave as humans behave, rather than doing what is optimal for the... more
Multi-agent systems can be effective tools for reasoning about human and group behavior. Their effectiveness is enhanced when the algorithms lead the simulated agents to behave as humans behave, rather than doing what is optimal for the task. Such systems are even more effective when the model's inputs are real data and the generated outputs are comparable to actual data files in the real world.
Abstract In this paper we adapt Axelrod's metanorms model to run on arbitrary interaction structures. We propose a hybrid analytical methodology combining computer simulations with a novel mathematical approach that abstracts some details... more
Abstract In this paper we adapt Axelrod's metanorms model to run on arbitrary interaction structures. We propose a hybrid analytical methodology combining computer simulations with a novel mathematical approach that abstracts some details of evolutionary dynamics while retaining explicit and exact representation of underlying network topology.
Abstract: Given the increasing threat of terrorism and spread of terrorist organizations, it is of vital importance to understand the properties of such organizations and to devise successful strategies for destabilizing them or... more
Abstract: Given the increasing threat of terrorism and spread of terrorist organizations, it is of vital importance to understand the properties of such organizations and to devise successful strategies for destabilizing them or decreasing their efficiency. However, intelligence information on these organizations is often incomplete, inaccurate or simply not available. This makes the study of terrorist networks and the evaluation of destabilization strategies difficult.
Abstract In an automated contracting environment, where a contractor agent must negotiate with other self-interested supplier agents in order to execute its plans, there is a tradeoff between giving the suppliers sufficient flexibility to... more
Abstract In an automated contracting environment, where a contractor agent must negotiate with other self-interested supplier agents in order to execute its plans, there is a tradeoff between giving the suppliers sufficient flexibility to incorporate the requirements of the contractor's call-for-bids into their own resource schedules, and ensuring the contractor that any bids received can be composed into a feasible plan.
Abstract: Testing large-scale dynamic network simulation packages, such as NetWatch, requires a large quantity of test data to be available for each of the experiments. The test data includes initial topologies of agents' social networks... more
Abstract: Testing large-scale dynamic network simulation packages, such as NetWatch, requires a large quantity of test data to be available for each of the experiments. The test data includes initial topologies of agents' social networks and specification of knowledge networks for each of the agents to fit an empirically derived distribution of knowledge.
Abstract In an automated contracting environment, where a" contractor" agent must negotiate with other selfinterested" supplier" agents in order to execute its plans, there is a tradeoff between giving the suppliers sufficient flexibility... more
Abstract In an automated contracting environment, where a" contractor" agent must negotiate with other selfinterested" supplier" agents in order to execute its plans, there is a tradeoff between giving the suppliers sufficient flexibility to incorporate the requirements of the contractor's call-for-bids into their own resource schedules, and ensuring that any bids received can be composed into a feasible plan.
Abstract On December 16, 2005, a New York Times article [31] revealed that in the immediate aftermath of the September 11th attacks, the US Government began a broad program of domestic signal intelligence collection. As press reports... more
Abstract On December 16, 2005, a New York Times article [31] revealed that in the immediate aftermath of the September 11th attacks, the US Government began a broad program of domestic signal intelligence collection. As press reports indicated [28], NSA implemented its new collections program based on the snowball sampling methods, which is generally used in surveying hidden populations and networks.
Abstract. Bid evaluation in a multi-agent automated contracting environment presents a challenging search problem. We introduce a multicriterion, anytime bid evaluation strategy that incorporates cost, task coverage, temporal feasibility,... more
Abstract. Bid evaluation in a multi-agent automated contracting environment presents a challenging search problem. We introduce a multicriterion, anytime bid evaluation strategy that incorporates cost, task coverage, temporal feasibility, and risk estimation into a simulated annealing framework. We report on an experimental evaluation using a set of increasingly informed search heuristics within simulated annealing.
1 In the last few years, the electronic marketplace has witnessed an exponential growth in worth and size, and projections are for this trend to intensify in coming years. Yet, the tools available to market players are very limited, thus... more
1 In the last few years, the electronic marketplace has witnessed an exponential growth in worth and size, and projections are for this trend to intensify in coming years. Yet, the tools available to market players are very limited, thus imposing restrictions on their ability to exploit market opportunities. While the Internet offers great possibilities for creation of spontaneous communities, this potential has not been explored as a means for creating economies of scale among similar-minded customers.
Abstract U-Mart International Experiment (UMIE 200x) has held by U-Mart Organizing Committee and U-Mart System Operating Committee since 2002. This is the annual series of open experiment of artificial market using U-Mart System–virtual... more
Abstract U-Mart International Experiment (UMIE 200x) has held by U-Mart Organizing Committee and U-Mart System Operating Committee since 2002. This is the annual series of open experiment of artificial market using U-Mart System–virtual futures market of stock index. Machine agents were recruited publicly and compete their performance of investment each other.
Abstract This paper seeks to expand upon what we know regarding the structure of interest group networks. To explore the potential for Social Network Analysis in the larger context of group behavior and information sharing, we make use of... more
Abstract This paper seeks to expand upon what we know regarding the structure of interest group networks. To explore the potential for Social Network Analysis in the larger context of group behavior and information sharing, we make use of existing federal contribution data to explore how non-corporate political action committees are linked to one another at the federal level via regulated campaign finance. Thus, we explore the density of networks, who the central players are and group relationships with respect to investment behavior.
Abstract. It has been observed in many instances that markets have a tendency to segment themselves into distinct sub-markets. This paper presents a multi-agent model that illustrates emergent market segmentation. The model illustrates... more
Abstract. It has been observed in many instances that markets have a tendency to segment themselves into distinct sub-markets. This paper presents a multi-agent model that illustrates emergent market segmentation. The model illustrates the way local optimization processes result in an emergent global behavior.
Abstract We describe an approach to agent-based contracting that combines the advantages of automated negotiation and evaluation of alternatives, with the realities of human decision-making and authority. The MAGNET architecture for... more
Abstract We describe an approach to agent-based contracting that combines the advantages of automated negotiation and evaluation of alternatives, with the realities of human decision-making and authority. The MAGNET architecture for multi-agent automated contracting provides a secure market-based infrastructure for complex negotiation, including an extensible ontology layer that defines the terms of discourse for inter-agent communication.
Abstract: Given the increasing threat of terrorism and spread of terrorist organizations, it is of vital importance to understand the properties of such organizations and to devise successful strategies for destabilizing them or... more
Abstract: Given the increasing threat of terrorism and spread of terrorist organizations, it is of vital importance to understand the properties of such organizations and to devise successful strategies for destabilizing them or decreasing their efficiency. However, intelligence information on these organizations is often incomplete, inaccurate or simply not available. This makes the study of terrorist networks and the evaluation of destabilization strategies difficult.
To facilitate the analysis of real and simulated data on groups, organizations and societies, tools and measures are needed that can handle relational or network data that is multi-mode, multi-link and multi-time period in which nodes and... more
To facilitate the analysis of real and simulated data on groups, organizations and societies, tools and measures are needed that can handle relational or network data that is multi-mode, multi-link and multi-time period in which nodes and edges have attributes with possible data errors and missing data. The integrated CASOS dynamic network analysis toolkit described in this paper is an interoperable set of scalable software tools.
Link mining, social network and semantic network construction are data fitting problems for which relevant information content is encoded in language and names. Norbert Wiener [7] noted that those features that distinguish human... more
Link mining, social network and semantic network construction are data fitting problems for which relevant information content is encoded in language and names. Norbert Wiener [7] noted that those features that distinguish human communication are “...(1) the delicacy and complexity of the code used, and (2) the high degree of arbitrariness of this code.”
Abstract We define a universal data interchange format to enable exchange of rich social network data and improve compatibility of analysis and visualization tools. DyNetML is an XML-derived language that provides means to express rich... more
Abstract We define a universal data interchange format to enable exchange of rich social network data and improve compatibility of analysis and visualization tools. DyNetML is an XML-derived language that provides means to express rich social network data. DyNetML also provides an extensible facility for linking anthropological, process description and other data with social networks. DyNetML has been implemented and in use by the CASOS group at Carnegie Mellon University as a data interchange format.
Abstract. In this paper, we introduce a computational model of asymmetric conflict between organizations. First, we describe a tactical simulation of organizational behavior and learning, based on Belief-Desire-Intention heuristics.... more
Abstract. In this paper, we introduce a computational model of asymmetric conflict between organizations. First, we describe a tactical simulation of organizational behavior and learning, based on Belief-Desire-Intention heuristics. Later, we introduce strategic reasoning using n-th order rationality and anchor it within the tactical simulation. The paper concludes with tests of our approach using artificial scenarios. Key words: Multiagent Simulation, Organizational Learning, Bounded Rationality, Irregular Warfare
In the field of social network analysis, we are used to competing – sometimes contradictory – theories coexisting with each other for decades. For example, let's take the idea of homophily [12], a promising theory of why and how social... more
In the field of social network analysis, we are used to competing – sometimes contradictory – theories coexisting with each other for decades. For example, let's take the idea of homophily [12], a promising theory of why and how social networks form – that states that people are more likely to form ties with others that are similar (”birds of a feather flock together”).
Abstract: The authors define a universal data interchange format to enable the exchange of rich social network data and to improve the compatibility of analysis and visualization tools. DyNetML is an XML-derived language that provides a... more
Abstract: The authors define a universal data interchange format to enable the exchange of rich social network data and to improve the compatibility of analysis and visualization tools. DyNetML is an XML-derived language that provides a means to express rich social network data. DyNetML also provides an extensible facility for linking anthropological, process description, and other data with social networks. DyNetML has been implemented and in use by the CASOS group at Carnegie Mellon University as a data interchange format.
It has been observed in many instances that markets have a tendency to segment themselves into distinct sub-markets. This paper presents a multi-agent model that illustrates emergent market segmentation. The model illustrates the way... more
It has been observed in many instances that markets have a tendency to segment themselves into distinct sub-markets. This paper presents a multi-agent model that illustrates emergent market segmentation. The model illustrates the way local optimization processes result in an emergent global improvement of social welfare.
Abstract: There is a pressing need to automatically collect data on social systems as rich network data, analyze such systems to find hidden relations and groups, prune the datasets to locate regions of interest, locate key actors,... more
Abstract: There is a pressing need to automatically collect data on social systems as rich network data, analyze such systems to find hidden relations and groups, prune the datasets to locate regions of interest, locate key actors, characterize the structure, locate points of vulnerability, and simulate change in a system as it evolves naturally or in response to strategic interventions over time or under certain impacts, including modification of data.
Abstract Kinship is an essential factor in the study of armed conflict. A better understanding of kinship mechanisms and related behavioral and social patterns can contribute to a better understanding of armed conflict. We introduce an... more
Abstract Kinship is an essential factor in the study of armed conflict. A better understanding of kinship mechanisms and related behavioral and social patterns can contribute to a better understanding of armed conflict. We introduce an empirical computational social science framework to explore basal mechanisms in the production of identity and the evolution of friend-or-foe relationships. An agent-based social simulation model is developed based on empirical data taken from an existing case-study of the Swat Valley in Pakistan.
Abstract: Most commanders, politicians and intelligence agents have at least an intuitive understanding of hierarchies and how to affect their behavior. However, covert organizations, such as terrorist organizations, have network... more
Abstract: Most commanders, politicians and intelligence agents have at least an intuitive understanding of hierarchies and how to affect their behavior. However, covert organizations, such as terrorist organizations, have network structures that are distinct from those in typical hierarchical organizations. In particular, they tend to be more cellular and distributed. This makes it difficult to apply the lessons of experience in determining how best to destabilize these groups.
Abstract Given the increasing threat of terrorism and spread of terrorist organizations, it is of vital importance to understand the properties of such organizations and to devise successful strategies for destabilizing them or decreasing... more
Abstract Given the increasing threat of terrorism and spread of terrorist organizations, it is of vital importance to understand the properties of such organizations and to devise successful strategies for destabilizing them or decreasing their efficiency. However, intelligence information on these organizations is often incomplete, inaccurate or simply not available–thus making a study of terrorist networks more difficult.
Abstract: In study of covert networks and destabilization strategies thereof, much attention has been paid to the task of locating and isolating key individuals within the organization. However, the real-world covert networks that have... more
Abstract: In study of covert networks and destabilization strategies thereof, much attention has been paid to the task of locating and isolating key individuals within the organization. However, the real-world covert networks that have been a target of various isolation strategies have shown a high level of resilience, and were able to quickly recover from the damage caused by the isolation.
Abstract: This paper is first stage of a project that project seeks to expand upon what we know about group networks. To explore the potential for network analysis in the larger context of group behavior and information sharing, we make... more
Abstract: This paper is first stage of a project that project seeks to expand upon what we know about group networks. To explore the potential for network analysis in the larger context of group behavior and information sharing, we make use of existing data. This data does not answer the important questions scholars raise, but it does allow us to investigate one aspect of interest group behavior: PAC influence. Thus it allows us to explore the density of networks on one level of behavior: investment behavior.
Abstract We raise a question of whether it is appropriate to simulate human and social systems using simple” ant-like” agents, and argue that in order to simulate emergence and evolution of social order, complexity must reside not only on... more
Abstract We raise a question of whether it is appropriate to simulate human and social systems using simple” ant-like” agents, and argue that in order to simulate emergence and evolution of social order, complexity must reside not only on system level (ie simple agents with complex relations), but also on the level of an individual agent, in the form of boundedly rational reasoning and planning capabilities.
Conflict is often viewed by scientists as a normative concept, something to be resolved, mitigated, abated and otherwise removed from the fabric of social structure. Yet, stubbornly, it remains a force that is neither going away, nor has... more
Conflict is often viewed by scientists as a normative concept, something to be resolved, mitigated, abated and otherwise removed from the fabric of social structure. Yet, stubbornly, it remains a force that is neither going away, nor has any plans to in the near future.
Abstract. The complexity of human social structures often masks the simplicity involved in their development. Social networks are a product of dynamic processes and feedback. In other words, the ties that people make affect the topology... more
Abstract. The complexity of human social structures often masks the simplicity involved in their development. Social networks are a product of dynamic processes and feedback. In other words, the ties that people make affect the topology of a network and the form of a network affects the ties that people make. Therefore, social network structure evolves in a path-dependent manner.
Computational social science and in particular agent-based social simulation continue to gain momentum in the academic community. Social network analysis enjoys even more popularity. They both have much in common. In agent-based models,... more
Computational social science and in particular agent-based social simulation continue to gain momentum in the academic community. Social network analysis enjoys even more popularity. They both have much in common. In agent-based models, individual interactions are simulated to generate social patterns of all kinds, including relationships that can then be analyzed by social network analysis. This chapter describes and discusses the role of agent-based modeling in the generative-analytical part of this symbiosis.
Abstract Snowball sampling methods are known to be a biased toward highly connected actors and consequently produce core-periphery networks when these may not necessarily be present. This leads to a biased perception of the underlying... more
Abstract Snowball sampling methods are known to be a biased toward highly connected actors and consequently produce core-periphery networks when these may not necessarily be present. This leads to a biased perception of the underlying network which can have negative policy consequences, as in the identification of terrorist networks. When snowball sampling is used, the potential overload of the information collection system is a distinct problem due to the exponential growth of the number of suspects to be monitored.
Abstract We present a system for multi-agent contract negotiation, implemented as a generalized market architecture called MAGNET. MAGNET provides support for a variety of types of transactions, from simple buying and selling of goods and... more
Abstract We present a system for multi-agent contract negotiation, implemented as a generalized market architecture called MAGNET. MAGNET provides support for a variety of types of transactions, from simple buying and selling of goods and services to complex multi-agent contract negotiations. In the latter case, MAGNET is designed to negotiate contracts based on temporal and precedence constraints, and includes facilities for dealing with time-based contingencies.
Abstract We present a system for multi-agent contract negotiation, implemented as a generalized market architecture called MAGNET. MAGNET provides support for a variety of types of transactions, from simple buying and selling of goods and... more
Abstract We present a system for multi-agent contract negotiation, implemented as a generalized market architecture called MAGNET. MAGNET provides support for a variety of types of transactions, from simple buying and selling of goods and services to complex multi-agent contract negotiations. In the latter case, MAGNET is designed to negotiate contracts based on temporal and precedence constraints, and includes facilities for dealing with time-based contingencies.
Abstract Agent-based approaches provide an invaluable tool for building decentralized, distributed architectures and tying together sets of disparate software tools and architectures. However, while the agents themselves have been gaining... more
Abstract Agent-based approaches provide an invaluable tool for building decentralized, distributed architectures and tying together sets of disparate software tools and architectures. However, while the agents themselves have been gaining complexity, and agent specification languages have been gaining expressive power, little thought has been given to the complexity of agent societies, and languages for describing such societies.
In this paper, we focus on computational modeling of adversarial activities and asymmetric warfare in a tactical setting. As a topic for simulation study, asymmetric warfare is an odd and ill-conditioned problem. Empirical data is scarce... more
In this paper, we focus on computational modeling of adversarial activities and asymmetric warfare in a tactical setting. As a topic for simulation study, asymmetric warfare is an odd and ill-conditioned problem. Empirical data is scarce or one-sided, while the subject itself presents a constantly adapting and moving target that makes it a strategic priority to conceal its inner workings from the observers.