University of Maryland Scientists Develop New Methods to Combat Pakistani Terrorist Group Lashkar-e-Taiba

UMD Press Release (Links Added by SWJ) - College Park, Maryland - Effectively reducing the likelihood and intensity of attacks by the Pakistani terrorist group Lashkar-e-Taiba (LeT), the perpetrators of the November 2008 assault on Mumbai, India, requires a cocktail of actions including fostering dissent within LeT, hampering the organization’s ability to conduct communication campaigns or provide social services, and disrupting the links between LeT and other Islamist terror groups, says a new study completed by an interdisciplinary research team at the University of Maryland’s Lab for Computational Cultural Dynamics.

Presented at an international symposium on Lashkar-e-Taiba held in Washington on Sep 10, 2012, the study also confirms traditional wisdom that pressuring Pakistan to rein in its terrorist proxies and disrupt LeT terrorist training camps is also necessary to reduce the scope of LeT attacks. The UMD research further showed that traditional counter-terror and law-enforcement tools such as arrests, raids, and targeting the group’s field commanders have only had a limited impact in reducing the likelihood of Lashkar-e-Taiba attacks. The study also is being published as a book: “Computational Analysis of Terrorist Groups: Lashkar-e-Taiba” (Springer, released Sep. 10, 2012).

Besides killing hundreds of civilians, LeT threatens the stability of South Asia because its attacks heighten tensions between nuclear-armed rivals India and Pakistan. Further, since the assault on Mumbai, information on Lashkar-e-Taiba suggests that it has increasingly turned its attention towards attacking the West not only in Afghanistan but also in Europe and Australia.

“Our study of LeT is different,” explains V.S. Subrahmanian, lead author of the study, and director of the University of Maryland’s Laboratory for Computational Cultural Dynamics, “It is the first in-depth analysis of a terror group that uses sophisticated data mining algorithms to learn temporal probabilistic rules as well as new algorithms to automatically suggest set policies which are sets of actions that should and should not be taken in order to elicit a desired behavior. Companies like Google and Amazon use these kinds of analytic methodologies to model the behaviors of customers every day. Decision-makers dealing with deadly threats to national security should have the same kinds of tools available.”

The findings on how best to deal with LeT are based on systematically gathered monthly data on 770 variables over a period of over 20 years. The UMD researchers mined this data for temporal probabilistic rules that not only identify conditions under which different types of terror strikes are carried out by LeT or its affiliates, but also the time delay with which these actions occur..  Given these rules about the likelihood of LeT actions, a new Policy Computation Algorithm identifies sets of actions that reduce the likelihood of LeT attacks.

A typical rule states that two months after 5-24 Lashkar-e-Taiba operatives were arrested and Lashkar-e-Taiba operatives were on trial in either India or Pakistan, there was an 88 percent probability of Lashkar-e-Taiba engaging in clashes with local security forces in which Lashkar-e-Taiba operatives are killed. The software generated hundreds of such rules about a vast range of Lashkar-e-Taiba attacks including their targeting of civilians, professional security forces, transportation centers, security installations, and symbolic/tourist locations.

Overall the rules showed that support from Pakistan’s government for LeT (which over the years has included financial, military, as well as operational support) is strongly correlated with almost every type of Lashkar-e-Taiba violence, while when Lashkar-e-Taiba is suffering from internal dissension they are less likely to carry out every type of attack. Other rules had more nuanced implications. For example, the killing of Lashkar-e-Taiba commanders is often followed by the increased likelihood of attacks on civilians but fewer cases of other attacks.

The study found that policy-makers must continue to identify methods of pressuring Pakistan to stop supporting Lashkar-e-Taiba as its proxy against India. But, since this has proved difficult to do in practice, counter-terror agencies should also seek methods of sowing internal dissent within Lashkar-e-Taiba, a strategy that has been used successfully against other terrorist organizations. Moreover, they should consider disrupting LeT’s successful news campaigns which allow them to explain their terror campaigns, and their social services’ campaigns that enable them to garner new recruits.

Former Director of Central Intelligence, R. James Woolsey, who contributed the new book’s Foreword, observes that this work is “a rich beginning to the potential utilization of such probabilistic modeling as a broadly used tool for combating many types of terrorism.” S. Ramadorai, the Chairman of Mumbai’s Stock Exchange, and an adviser to India’s Prime Minister observed that “Analytical tools from diverse disciplines converge [in the study] to provide the reader with an in-depth understanding of the secret world of terrorism.”

The Laboratory for Computational Cultural Dynamics (LCCD), which is part of the University of Maryland Institute for Advanced Computer Studies, develops computational tools to support decision-makers. The authors of the study, which was carried out at the University of Maryland, are V.S. Subrahmanian, Aaron Mannes, Amy Sliva, Jana Shakarian, and John Dickerson.

0
Your rating: None

Comments

LOL! I enjoyed the article. However I think it authors don't have it quite right. Yes it can figure the probabilities, probably courses of action etc. And though it makes places like Google a lot of money (analysis helps in any kinds of gambling, like business) it think will fail when it comes to a catastrophic event, major terrorist attack or un expected invasion etc.

It sounds as though it might be a somewhat useful tool... Like a huge CARVER matrix.

The problem as I see it will be that it will become when it fails to predict and event, it, rather than political/intelligence/military leadership will get the blame. Kinda like the movie War Games.

Good intel, with excellent analysis, provided to solid politicians and military leaders, isn't going to stop, fix or solve everything. You can't predict the future. But you can blame the machine.

Though I am second to none in my skepticism of high tech tron slinging, this article impressed me. The problem with it is the language it uses. What the heck does "temporal probabilistic rules" mean? They are talking about figuring out what people are likely to do in the future and learning how to affect their behavior, based upon what they've done in the past. The difference is the computer tech allows you to have much more good knowledge about what they have done in the past, if you have the right kind of programs to sift through that behavior.

Google and Amazon have those programs. And they work. All you have to do is look at how much money they make. It appears that the U of MD had the inspiration to apply those programs and methods in a different area. It is a great idea. I just wish they would explain clearly without the jargon. Nobody will know they have a good idea because most will stop reading right after they get hit with "sophisticated data mining algorithms to learn temporal probabilistic rules".