Laboratory for Human Terrain
Our Publications

HBML: A Representation Language for Quantitative Behavioral Modeling in the Human Terrain
N. Sandell, R. Savell, D. Twardowski, G. Cybenko
    Human and machine behavioral modeling and analysis are active areas of study in a variety of domains of current interest. Notions of what behavior means as well how behaviors can be represented, compared, shared, integrated and analyzed are not well established and vary from domain to domain, even from researcher to researcher. Our current research suggests that a common framework for the systematic analysis of behaviors of people, networks and engineered systems is both possible and much needed. The main contribution of this paper is the presentation of a framework for representing behaviors. We believe that the proposed schema and framework can support large-scale, computational systems’ behavioral modeling and analysis across a variety of domains.

Go To Article...

Process Query Systems
George Cybenko and Vincent Berk
IEEE Computer
     Sensors produce large streams of raw events while instrumenting environments such as computer systems, communications networks, physical spaces,and human organizations. Extracting meaningful and actionable information from these events, however, remains a challenge. Process query systems,a new algorithmic and software paradigm, offer a powerful and generic way to address event-processing challenges.

Go To Article...

AI and the Modern Networked Organization
George Cybenko
IEEE Intelligent Systems
     Given a group of people, a common goal, and a collaboration technology, how will they collaborate and with what effectiveness?
Go To Article...

Semantic Depth and Markup Complexity
Guofei Jiang, George Cybenko, James A. Hendler
     In order to achieve interoperability among heterogeneous systems, markup languages such as XML and DAML are being used to describe distributed systems and data. The ability to successfully interoperate based on semantic markup depends on the ability to create, use and manage shared ontologies of concepts and their interrelationships. Specifically, communicating systems in a networked environment have to achieve a certain level of semantic agreement for them to understand and process exchanged data. A challenging question is how deep the semantic agreement has to be in order to satisfy the communication needs in an environment. Additionally, what is the markup complexity resulting from pursuing that depth of semantic agreement? This paper introduces the concept of semantic depth and markup complexity and proposes models to measure the markup complexity. Furthermore, it is shown that markup complexity can be reduced by employing hierarchical ontologies after partitioning the domain into smaller sub-domains.

Go To Article...

Mining for Social Process Signatures in Intelligence Data Streams
Robert Savell, George Cybenko
     The detection and tracking of embedded subnets in an active social network can be computationally daunting due to the quantity of transactional data generated in the natural interaction of large numbers of actors comprising a network. In addition, detection of malevolent behavior may be further complicated by evasive strategies designed to camouflage the activities of malicious sub-nets. As a result, the relevant human or signal intelligence data is often quite sporadic and noisy, and robust methods for identifying and tracking clandestinely operating sub-nets via sparse evidence are critical to the success of real-world intelligence applications. In this work, we move beyond traditional static methods of social network analysis to develop a methodology for mining intelligence data for evidence of dynamic social processes operating on the network substrate. This work is informed by earlier applications of the Process Query System (PQS) to process mining applications in various physical contexts. Given a process description encoding personal and/or coordinated behavior of social entities, we parse the networks transactional stream for instances of the active process and assign process states to events and functional entities based on a projection of the evidence onto the process model— with the goal not only to define the social network, but also to identify and track the dynamic states of functionally coherent subnetworks. In an application of our methodology to the Ali Baba simulated intelligence data set, we demonstrate superior results both in partitioning and contextualizing the social network and also in tracking the dynamic states of the malevolent social entities.

Go To Article...

Adversarial Models for Opponent Intent Inferencing
Eugene Santos Jr. and Qunhua Zhao
Adversarial Reasoning
     Taking into account the characteristics and behaviors of one's adversary is essential for success in any competitive activity, such as in sports, business, or warfare. Obviously, if one's enemies are well understood, their actions can then be better anticipated and countered. To do so, the key is to capture the adversary's intentions. An intuitive approach that immedatiately comes to mind is to think what you would do if you were "in your opponent's shoes."
Go To Article...

A Cognitive Architecture for Adversary Intent Inferencing: Structure of Knowledge and Computation
Eugene Santos Jr.
     Existing target-based and objectives-based (“strategy-to-task”) approaches to mission planning do not explicitly address the adversary’s decision-making processes. Obviously, the adversary’s courses of action (COA) are influenced in a cause-and-effect manner by actions taken by friendly forces. Given the iterative/interleaved nature of actions taken by enemy and friendly forces, mission planning must clearly take adversarial decision making into account especially during concurrent mission planning and execution. Currently, adversarial behaviour with regards to cause-and-effect are difficult to account for within the framework of existing planning approaches. This paper describes a cognitive architecture for computationally modeling, predicting, and explaining adversarial behaviors and COAs and proposes an integrated framework for mission planning. Our framework fits naturally within the Effects-Based Operations (EBO) approach to mission planning.

Go To Article...

Toward Detecting Deception in Intelligent Systems
Eugene Santos Jr. and Gregory Johnson Jr.
     Contemporary decision makers often must choose a course of action using knowledge-based diagnostic on these electronic sources such as knowledge-based diagnostics or decision support systems or through data mining techniques. As the decision maker becomes more dependent on these electronic information sources, detecting deceptive information fom these sources becomes vital to making a correct, or at least more informed decision. This applies to unintentional disinformation as well as intentional misinformation. Our ongoing research focuses on employing models of deception and deception detection algorithms for probabilitistic intelligent systems. The deception detection algorithms are used to detect, classify and correct attempts at deception. Algorithms for detecting unexpected information rely upon a prediction algorithm from the collaborative filtering domain to predict agent responses in a multi-agent system.

Go To Article...

Modeling Multiple Communities of Interest for Interactive Simulation and Gaming: The Dynamic Adversarial Gaming Algorithm Project
Eugene Santos Jr., Qunhua Zhao, Felicia Pratto, Adam R. Pearson, Bruce McQueary, Andy Breeden, Lee Krause
     Nowadays, there is an increasing demand for the military to conduct operations that are beyond traditional warfare. In these operations, analyzing and understanding those who are involved in the situation, how they are going to behave, and why they behave in certain ways is critical for success. The challenge lies in that behavior does not simply follow universal/fixed doctrines; it is significantly influenced by soft factors (i.e. cultural factors, societal norms, etc.). In addition, there is rarely just one isolated enemy; the behaviors and responses of all groups in the region, and the dynamics of the interaction among them composes an important part of the whole picture. The Dynamic Adversarial Gaming Algorithm (DAGA) project aims to provide a wargaming environment for automation of simulating dynamics of geopolitical crisis and eventually be applied to military simulation and training domain, and/or commercial gaming arena. The focus of DAGA is on modeling communities of interest (COIs), where various individuals, groups, and organizations as well as their interactions are captured. The framework should provide a context for COIs to interact with each other and influence others’ behaviors. These behaviors must incorporate soft factors by modeling cultural knowledge. We do so by representing cultural variables and their influence on behavior using probabilistic networks. In this paper, we describe our COI modeling, the development of cultural networks, the interaction architecture, and a prototype of DAGA.

Go To Article...

Impacts of User Modeling on Personalization of Information Retrieval: An Evaluation with Human Intelligence Analysts
Eugene Santos Jr. Qunhua Zhao, Hien Nguyen, and Hua Wang
     User modeling is the key element in assisting intelligence analysts to meet the challenge of gathering relevant information from the massive amounts of available data. We have developed a dynamic user model to predict the analyst’s intent and help the information retrieval application better serve the analyst’s information needs. In order to justify the effectiveness of our user modeling approach, we have conducted a user evaluation study with actual end user, three working intelligence analysts, and compared our user model enhanced information retrieval system with a commercial off-the-shelf system, the Verity Query Language. We describe our experimental setup and the specific metrics essential to evaluate user modeling for information retrieval. The results show that our user modeling approach tracked individual’s interests, adapted to their individual searching strategies, and helped retrieve more relevant documents than the Verity Query Language system.

Go To Article...


Human Terrain News
The Why and How of Human Terrain Teams
Inside Higher Ed
February 19, 2009

There are a number of ways that an HTT can keep the population and nonlethal options on the front burner. In the case of my team, we used very standard research and analysis methods to get at both primary and secondary open source data. At all times we endeavored to engage in best practices, both in terms of methodology and ethics. We essentially used four basic methods of collection: archival, process observation, participant observation, and semi-structured elite level interviews.

McFate explains Human Terrain Teams
The Dartmouth
September 26, 2008

Cultural anthropologist Montgomery McFate emphasized the importance of sociocultural knowledge in forming national security strategy, explaining a new counter-insurgency theory that relies on experts in social-science disciplines, in a lecture at the Rockefeller Center on Thursday evening.

Army Enlists Anthropology in War Zones
New York Times
October 5, 2007

The SHABAK VALLEY, Afghanistan — In this isolated Taliban stronghold in eastern Afghanistan, American paratroopers are fielding what they consider a crucial new weapon in counterinsurgency operations here: a soft-spoken civilian anthropologist named Tracy.


Pentagon Tech Push
April 3, 2007

The Pentagon announced yesterday its “new starts” for the 2007 Joint Capability Technology Demonstration program including ...Mapping the Human Terrain (MAP-HT) - Visualization of socio-cultural information

Koppel talks war, U.S. toll in Discovery Channel Special
March 10, 2007

The phrase "hearts and minds" is in disrepute. The new buzz phrase is "human terrain," but it still means winning friends among people prone to distrust all outsiders, particularly Americans. "But if that is the terrain the U.S. is trying to conquer, the war is not going well," Koppel concludes.