Throughout my college career, many of my experiences have involved work with critical healthcare data. It seems as though everyday, particularly in healthcare, there are news reports of massive new data breaches, and it seems that these hackers are always one step ahead.
However, I have just come across an article that I thought was really interesting, and may be something I devote some side work time to exploring; replacing binary authentication methods (i.e: you know or don’t know your password) to utilizing unsupervised machine learning algorithms to assure you are the correct user. By unsupervised, we basically mean that there is initially no data to go off of for prediction, and the model will morph to track your habits, biometric signals, etc.
Aetna, an american healthcare insurance company, recently just launched a feature in their consumer apps that does exactly this. In an article I read from HealthcareITNews, the security feature “takes in data from many attributes of the device (software configuration, operating system version, etc.), in addition to benign attributes of consumer behavior (for example, how a mobile device is held when texting and location of the device), and matches these attributes against a device signature and a model based on previous behavior”. I found this to be a really incredible feature, and is something I am excited to try out for myself in future work. I will be posting the link below to the full article as well, and for those who are interested in this type of work should give HealthcareITNews a follow.