image: Legitimate, and counterfeit, HP ink cartridge packaging. As an HP Fellow, Professor Steve Simske previously developed a scanning technology that can quickly identify falsified products like the one on the right. A new NSF grant will support further efforts around disrupting illicit supply chains. view more
Credit: Bill Cotton/CSU Photography
Underground networks that traffic in everything from counterfeit watches to human beings are a staggering global problem. These networks represent an estimated $2 trillion or more in illegal activity that undermines governments and victimizes marginalized people.
Colorado State University professor Steve Simske is among those trying to diffuse the threat of illicit supply chains, by understanding how they work and how to stop them.
Simske and collaborators from South Dakota School of Mines and Technology are among nine recipients of a new National Science Foundation grant supporting early-stage scientific research for studying and dismantling illicit supply networks. The team was awarded $300,000 over two years to conduct proof-of-concept research, through an NSF program called EAGER (Early-concept Grants for Exploratory Research).
A longtime HP Fellow and research director at HP Labs, Simske joined the faculty of CSU in January as a professor in mechanical engineering and systems engineering. For his NSF project, he will build upon previous work he did at HP, which helped reveal hundreds of millions of dollars in counterfeit HP products.
"Ultimately, what we are trying to do is deconvolute what's being done through licit actions, to identify illicit purposes," Simske said. "It's a hard problem to solve, and we will address it through technology."
The new awards, according to an NSF news release, support research that combines engineering with computer, physical, and social sciences to "address a danger that poses significant consequences for national and international security."
"Nimble and technologically sophisticated networks traffic in contraband that includes people, illegal weapons, drugs, looted antiquities, and exotic animal products. Unencumbered by national boundaries, they funnel illicit profits to criminal organizations, and fuel transnational and terrorist organizations," according to the agency.
One major tenet of Simske's proposal will be to improve a counterfeit image-detection technology he pioneered at HP. Between 2008 and 2012, Simske and colleagues developed an image-scanning system that uses a small set of pre-classified images to teach a computer to differentiate between real and fake sources. HP used this color-mapping, 3D barcode scanning technology to help the company identify when packaging for ink cartridges and other products had been illegally copied.
Simske and colleagues plan to leverage this existing work into additional customization and analytic capabilities. Their ultimate goal is to steer an illicit network into a situation in which they must reveal themselves.
They will accomplish this goal by determining statistically relevant deviations from legitimate supply chains, which illicit trades depend on for efficiency and for simulating legitimacy. The researchers will explore and compare physical- and cyber-forensic processes; for example, they will use machine learning and analysis of public web sales and product pricing, coupled with proprietary information like a company's projected regional sales, to empower software to reveal weaknesses in illicit supply chains.
Among the objectives of the NSF proposal is to develop:
A network "scraper" that automatically collects product pricing information, then determines appropriate analytics for identifying suspicious sales activity
A simple means to anonymize sales and distribution data
Algorithms and automatic clustering, labeling and classification technologies
Software for modeling the operational and spatial dynamics of illicit networks
Algorithms for determining the means through which illicit networks offer incentives to their users
Advanced analytics and meta-analytics to discover central factors in illicit networks.
###