Noblis Participates in Global Events to Stop Child Exploitation

Noblis staff members participate in international working groups to develop standardized and open source machine learning, data science, and analytics capabilities to help stop child exploitation

Did you know that the National Center for Missing & Exploited Children (NCMEC) reviews 25 million images containing child sexual abuse imagery (CSAM) every year?

As the volume of seized images and videos of child exploitation increases, the need for tools to analyze that data also increases. Noblis experts are dedicated to helping law enforcement solve these challenges, and over the year, have participated in international working groups to support the development of key capabilities that can help stop child exploitation.

In July, a Noblis team – Jordan Cheney, Charles Otto, and Eric Kukula participated in a Hackathon on behalf of their client, Department of Homeland Security (DHS) Science and Technology Directorate, First Responders Group in Vancouver, Canada. This hackathon, called the Protecting Innocence Hackathon, brought together over 60 participants from 5 countries to work together to build technology and global policy that can stop online child exploitation.

While in Vancouver, the Noblis team worked on two projects alongside DHS and other participants. The first project sought to develop a prototype triage analytic that classifies imagery containing child exploitation, while the objective of the second was to develop a standard CSAM scientific annotation schema for law enforcement to develop, train, and evaluate future machine learning algorithms and analytics.

“The amount of data law enforcement seizes during child exploitation investigations from physical media storage devices and online from chat rooms, newsgroups, and peer-to-peer networks has seen year-over-year growth of about 20%” said Dr. Eric Kukula, Director of Identity Intelligence at Noblis. “It’s extremely important that we invest and focus our technical energy to develop tools for investigators to more easily find the needles in the haystack and  generate investigative leads.”

On November 14-17, Dr. Kukula will co-present with law enforcement the draft standard CSAM annotation schema developed at the Hackathon during the 35thMeeting of the INTERPOL Specialists Group on Crimes Against Children (ISGCAC) in Lyon, France. IGSCAC is an international, multidisciplinary group dedicated to preventing the abuse of children through information sharing of emerging trends in evidence-based practice, public policy, technology, and law enforcement strategies.

“On average, there’s about 2 million pieces of media within a seizure,” said Kukula. “Unfortunately that number continues to grow and law enforcement does not have the time or workforce to manually analyze all of it. The standardized CSAM scientific annotation schema is one forward leaning approach where law enforcement can benefit from emerging technologies, such as machine learning, to train investigatory analytics and algorithms to automatically detect and categorize CSAM media, enabling investigators to focus on relevant information to rescue victims and apprehend these heinous offenders before they abuse again.”

To read more information about Interpol’s crimes against children mission, click here. To learn more about Noblis’ identity intelligence work, click here.