Fake account detection is a constant battle for most online companies. As the number of OSNs (On-line Social Networks) continues to expand, so too does the opportunity for fraudsters to exploit. This means businesses must have a persistently reliable way to identify the offending device.
Machine learning algorithms can be a boon when it comes to identifying suspicious activity. These algorithms are able to accurately detect fake accounts and their associated behaviors. Some of the most impressive ones have been developed by Facebook, a company that uses machine learning to combat spam, scams, and phishing.
One of the most sophisticated systems used by Facebook to fight bad actors is called the TIES (Technical Integrity Evaluation System). The system employs a complex machine learning model and features a number of layers of analysis. For instance, it has learned that an image with textual content attracts more attention than a similar image without it.
Another useful tool for detecting fake accounts is Facebook’s Deep Entity Classification (DEC) system. It is capable of detecting bogus accounts before they are created and can identify many of the most egregious cases of abuse.
A related technology, the SybilEdge algorithm, is able to spot fake accounts that evade anti-abuse filters and attempt to launch attacks. It compares a user’s extended social network to that of their contacts, thereby allowing it to find and disable accounts that fake account detection might look legitimate, but actually aren’t.
The aforementioned TIES system was only one of the more than a dozen AI-based tools that Facebook has been able to use to detect fraudulent accounts. Other models include Support Vector Machines (SVM), a nifty little number crunching heuristic, and a simple genetic algorithm.
For the most part, these machine-learning algorithms are only able to catch the most flagrant of bad actors. However, Facebook is also investing in several other layers of analysis that allow it to spot and prevent many of the lesser-known nefarious activities that plague its platform.
In the simplest terms, a fake account is one that is created by a spammer or a fraudster. Despite Facebook’s efforts to prevent the creation of phony accounts, these entities continue to exist. Their presence is a danger to everyone on the site, as they pose a threat to user privacy and security, and are known to spread false information.
There are other ways to detect suspicious activity, such as scanning the activity stream for signs of malicious behavior. Additionally, you can report suspicious activity to Facebook or contact your local police department. To ensure that you stay safe on the web, make sure to avoid creating accounts in the first place.
If you need help detecting a scam, you can consult Darktrace, a company that specializes in machine learning approaches to cybersecurity. They can provide you with a detailed profile of a suspected fake account and help you identify problematic behaviors. You can also report spam or malicious messages to Facebook by completing the required forms.