Thursday 20 March 2025
Smartphones have become an integral part of our daily lives, and it’s no surprise that they’re also a common tool for intimate partner abusers. They can use these devices to monitor their victims, track their movements, and even hack into their accounts. But what if there was a way to detect when someone is being abused using only the data from their phone? That’s exactly what a team of researchers has been working on.
The researchers developed an algorithm that uses machine learning to analyze the behavior patterns of smartphone users. They used data from 27 participants who were asked to complete various tasks, such as sending emails, uploading photos, and browsing the internet. The algorithm was trained to identify unusual patterns in the user’s behavior, which could indicate that they’re being abused.
The team found that their algorithm was able to detect abuse with an accuracy rate of over 98%. They also discovered that certain types of phone usage, such as frequent logging out and logging back in, were common among victims of intimate partner violence. This information could be used to identify potential cases of abuse earlier on, allowing for more effective intervention.
But how does it work? The algorithm is based on a two-stage architecture that processes multimodal signals from the smartphone. It uses sensors such as GPS, accelerometers, and gyroscopes to monitor the user’s physical activity, as well as data from apps like email, social media, and messaging services. This information is then used to train a machine learning model that can identify patterns of behavior that are indicative of abuse.
The researchers also developed a system called AID, which stands for Automated Intimate Partner Infiltration Detection. It’s a smartphone app that uses the algorithm to monitor the user’s behavior in real-time and alert them if it detects any suspicious activity. The app is designed to be easy to use and intuitive, with a simple interface that lets users track their phone usage and receive alerts when something unusual happens.
The potential benefits of AID are huge. It could help victims of intimate partner violence get the support they need earlier on, and potentially even prevent abuse from happening in the first place. It could also provide law enforcement agencies with valuable data to help them investigate cases of domestic violence.
Of course, there are some limitations to the system. For example, it’s only effective if the victim has a smartphone and is willing to install the app.
Cite this article: “Detecting Intimate Partner Abuse through Smartphone Data Analysis”, The Science Archive, 2025.
Smartphones, Intimate Partner Abuse, Machine Learning, Algorithm, Behavior Patterns, Gps, Accelerometers, Gyroscopes, Social Media, Messaging Services
