COPKIT is developing data-driven policing technologies to support Law Enforcement Agencies in preventing, investigating and mitigating crime and terrorism.
In this blog, Gendarmerie Nationale provide an overview of the Secure Testing Lab platform which provides a safe environment for testing and evaluating the COPKIT tools using law enforcement data.
One of the main challenges when developing purpose-specific software for any field is to have access to quality and representative data to test new theories and assess the efficiency of the applied methods.
For law enforcement systems it is not different, constraints being even bigger since the access to real police data is subject to a series of deontological and legal constraints. Consequently, Law Enforcement Agencies (LEAs) can hardly share their real databases with researchers to evaluate the impact of their theories and finetune them accordingly.
LEAs also have a hard time when required to compare different technology vendors’ solutions in a fair and impartial way and often end up relying on the vendor’s statements declaring their software demonstrations are based on synthetic data,that does not always reflect the local authorities’ reality in terms of type and statistics on data.
The H2020 funded COPKIT project is making a strong contribution in this direction by designing and implementing a Secure Test Lab (STL) enabling experimenting with innovative algorithms on real data without the risks of information leakage and privacy breaches. Furthermore, COPKIT adopts an integrated data protection and ethical impact assessment (E+PIA) approach to achieve responsible use and innovation of all the project’s technologies.
COPKIT brings together industry, academia, and LEAs to develop a toolkit supporting LEAs in their fight against organised crime and terrorist groups following an intelligence-led Early Warning/Early Action methodology. One of its main focuses is the development of a toolkit supporting the application of the methodology for knowledge production and use that can be tested by LEAs on their premises. To facilitate the uptake of these technologies, the project is developing innovative ways to ensure their evaluation and application in different environments.
Secure Test Lab
One of these contributions is the Secure Test Lab, a test infrastructure allowing the performance evaluation of the developed tools using real data on LEA premises in a safe way. The STL is designed to run on a 100% isolated machine, with no Internet access and no possibility for data leaks.
Main experimentation tracking interface showing the data for one live experimentation
The Secure Test Lab allows researchers to collect performance indicators of their developed tools when applied to real data. It enables benchmarking of vendors’ tools at LEA premises in a way that respects security and privacy constraints making proof of concept with vendors much less expensive and the uptake of innovative tools faster and more suitable for LEAs.
The Secure Test Lab continuously collects and displays different non-functional parameters linked to the experimentation execution along with security incident data access logs to check any suspicious behaviour of the experimentation. This gives LEAs a good idea of the performance of the experimentation and the means to verify requirements in terms of memory, CPU, IO (Input/Output) and network usage in real-time. It also enables the fair comparison of non-functional parameters between different solutions, as these parameters are stored, and old experimentations can be compared.
Other experimentation testbeds exist but are linked to specific partnerships between a given LEA and a specific company or research institute. To the best of our knowledge, this is the first reusable experiment structure that can use real LEA data open to the research community enabling to have collaborative projects where interim tests performed by LEAs help technology providers improve their methods.
Within the context of COPKIT, the framework will be made available to LEAs of the project to evaluate the impact and results of the developed solutions. Their feedback and assessment of the proposed methods will help technical partners of the project refine their techniques.
For more information and updates, follow us on Twitter, LinkedIn and Facebook and feel free to contact our team at firstname.lastname@example.org.