Analyzing Lashkar-e-Taiba

Analyzing Lashkar-e-Taiba

Lashkar-e-Taiba Script Lashkar-e-Taiba (LeT) is one of the deadliest terrorist groups in the world. Since 2004, LeT has carried out over 100 attacks worldwide, resulting in over 700 deaths [1]. In the last decade, LeT became both a proxy militia for the Pakistani Army and a political force within Pakistan. LeT is a terror group capable of carrying out complex, coordinated attacks against large targets, like the deadly 2008 attacks in Mumbai.

We have developed sophisticated algorithms to automatically learn SOMA models (Stochastic Opponent Modeling Agents) of the behavior of LeT based on an analysis of over 750 variables covering a period since the inception of LeT. These computational models have allowed us to derive significant insights into when LeT carries out attacks. In addition, we have used game-theoretic methods to understand the complex relationship between LeT, the Pakistani military, the Pakistani civilian government, the U.S., and India. These have allowed us to derive mathematically sound policies that the U.S. can adopt to rein in LeT's bad behaviors. Computational methods are needed for these analyses because of the enormous size of the data and the space of options available to policy makers.

Evaluating LeT Attacks in Jammu & Kashmir

LeT has taken various violent actions in the contested northern state of Jammu and Kashmir. We analyze this behavior using data from the Computational Modeling of Terrorism (CMOT) project, which is a cookbook for developing datasets for violent organizations. Data is collected at a granularity level of months, and is categorized as either "action" or "environmental". Action variables describe events like bombings or kidnappings, while environmental variables include such categories as idealogical goals of a group (religious rule, autonomy, etc) and the various political, geographical, economical, or cultural contexts under which the group operates, as well as many others. Overall, 770 unique variables are considered.

Then, using the SOMA system created at LCCD, we model the expected behavior of LeT—both in the past and into the future. We specifically explore situations during which LeT ramps up attacks in the Jammu and Kashmir region. We then provide some preliminary policy recommendations to analysts and decision makers in the field. This work has been accepted for publication and presentation at the European Conference on Intelligence Security Informatics (EISIC) in September 2011. The full paper can be found here.

A Game-Theoretic Perspective of LeT

LeT is not an independent entity acting in a vacuum; rather, it is supported in various ways by the Pakistani military and government and other countries, criticized (and more) by the United States and India, and in contact with multiple other actors on the world stage. Using non-zero-sum game theory, we model a situation involving five actors representing the U.S., India, LeT, the Pakistani military, and the Pakistani government. Each actor can take various actions (e.g., the U.S. can cut aid to Pakistan's government, or India can perform covert operations in Pakistan, etc). An assignment of actions to each agent is called a strategy profile. We explore these various strategy profiles using computational game theory.

To do this, we had an expert on LeT (who knew nothing about game theory) construct a payoff matrix describing every actor's response to actions performed by all other actors being modeled. From here, we exhaustively calculated all pure Nash equilibria (strategies where no agent can improve its situation by performing a different action, assuming all other agents follow the actions assigned by this strategy). We also calculated a large number of mixed Nash equilibria. We then analyzed the equilibria with desirable actions by LeT—namely, those where LeT's best action is to deterministically withhold any attacks, and eliminate its armed wing altogether with nonzero probability.

Interestingly, none of the desirable equilibria involve the U.S. increasing support to Pakistan. Furthermore, the majority of the equilibria imply that LeT's bad behavior can only be reined in if the U.S. and India concurrently engage in covert operations against LeT and/or coercive diplomacy against Pakistan. There is no need for the two to cooperate—they just need to take these actions at the same time. We describe these and a wealth of other results in a paper to be presented at OSINT in September 2011. The full paper can be found here.

Analyzing Indian Mujahideen

Similar techniques are currently being used to determine rules and policies relating to Indian Mujahideen, a terrorist group indigenous to India. Results are forthcoming.

For additional information, please contact Dr. V.S. Subrahmanian.

[1] Data from the U.S. National Counterterrorism Center's Worldwide Incident Tracking System.

Last updated: July 2011 by John Dickerson

Research and implementation of this project performed jointly by members of the University of Maryland, Carnegie Mellon University, and Northeastern University.

Project Contributors


This page will be updated as our work enters print. For information about receiving draft publications, technical reports, and conference presentations, please do not hesitate to contact team members.

The following sections may include links to restricted access material. Please do not hesitate to contact a group member for instructions regarding how to obtain a username and password.


Presentation links contingent upon publication!

The following sections may include links to restricted access material. Please do not hesitate to contact a group member for instructions regarding how to obtain a username and password.


Software links contingent upon publication!