The EnSafe project (Enhancing Environmental Protection: Anomaly Detection in Waste Transportation using Network Science) is developing AI-based solutions to tackle environmental crimes, focusing on waste transportation chains. EnSafe benefits from the active involvement of INESC TEC, which is developing technologies to detect irregular and suspicious behaviours in a sector that is vulnerable to fraud and environmental corruption.
The main goal of EnSafe is to identify anomalous or suspect behaviours in waste movements. These patterns may indicate regulatory violations, inefficient processes, safety risks, or even issues with data quality. The ability to detect these irregularities is vital to protect the environment and ensuring legal and sustainable waste management practices. But how can one identify these patterns?
“INESC TEC’s approach goes beyond traditional models, processing data not as isolated entries, but as an interconnected network – like a ‘social network’ for waste transportation. This model, combining network patterns with proven machine learning techniques, allows for the detection of unusual behaviours or activity patterns,” explained Shazia Tabassum, a researcher at INESC TEC.
In addition, the project resorts to innovative AI that learns from the structure of the data itself, identifying subtle signs of suspicious activity that often go unnoticed by human perception, resulting in a more accurate and adaptable system.
Within this network of transactions, it becomes possible to identify, for example, suspicious activities of stolen car parts, particularly catalytic converters, which made headlines in 2022. Other notable case includes waste trafficking, such as the illegal international trade and dumping of waste, where hazardous materials are deliberately misclassified as recyclable or non-hazardous. “These crimes are particularly concerning as they are considered low risk – investigations are rare and penalties are low – and highly profitable, posing serious threats to the environment. We are actively working to detect and prevent said violations as well”, stated the researcher.
However, detecting illegal activities is not always easy; non-compliance tends to be well concealed and can easily go unnoticed. “The data itself may be inconsistent, incomplete, or poorly structured, and in most cases, there is no pre-labelled dataset identifying illegal activity available to train the systems. EnSafe is tackling a complex issue in a field where there are still few effective solutions,” added Shazia Tabassum.
As part of the effort to connect technology with practical application, the EnSafe team recently conducted a technical visit to IGAMAOT – the General Inspectorate of Agriculture, Sea, Environment and Spatial. Planning -, the body responsible for environmental monitoring in Portugal. The aim of the visit was to strengthen collaboration with regulators and to gain a better understanding of detecting environmental crimes in the field.
The Institute designed a scalable and adaptable system, as it uses data aggregated annually, allowing for efficient processing without overloading the system with excessive detail.
EnSafe is funded by the Foundation for Science and Technology (FCT) and is part of the national strategy to tackle environmental crime.
The researcher mentioned in this news piece is associated with INESC TEC.
News, current topics, curiosities and so much more about INESC TEC and its community!
