Unsupervised detection of physical contacts on submarine cables

Submarine cable infrastructure forms the backbone of global communications, carrying over 95% of internet traffic. However, its exposure to human activities such as trawling or vessel anchoring poses a constant threat. Traditionally, detecting physical contacts on these cables has relied on labeled systems requiring costly and reactive monitoring. In this context, unsupervised approaches emerge as a disruptive alternative: they allow anomaly identification without the need for labeled data, drastically reducing response time and false positives. A recent example demonstrates how a deep learning-based detector, trained without event labels, managed to classify all confirmed trawler contacts among the top positions of a database with over one hundred thousand records, also corroborating other contact events with external evidence. This type of innovation not only optimizes the surveillance of critical assets but also opens the door to autonomous early warning systems. Developing these solutions requires combining knowledge of neural networks, time series processing, and deployment on cloud infrastructure. At Q2BSTUDIO, we address these challenges through custom applications that integrate artificial intelligence for predictive infrastructure monitoring. Our team designs AI for businesses capable of learning complex patterns without human intervention, optimizing anomaly detection in submarine or industrial environments. The scalability of these systems largely depends on AWS and Azure cloud services, which provide the computing power needed to process massive volumes of data in real time. Additionally, cybersecurity plays a fundamental role: protecting both transmitted data and the detection algorithms themselves from potential tampering. In this regard, we integrate cybersecurity solutions that ensure the integrity of deployed systems. For result visualization and analysis, business intelligence services such as Power BI allow operators to intuitively interpret alert signals, while AI agents automate immediate responses to detected contacts. All of this is materialized through custom software that adapts to each client's specific needs, from cable operators to offshore energy companies. The combination of these capabilities—cloud processing, unsupervised algorithms, and advanced visualization—is transforming how submerged assets are protected, shifting from a reactive strategy to a truly predictive one.