Determining anomalies in data using classical machine learning techniques typically requires characterizing the notion of what is “normal” or “expected” in the instance space. Upon doing so, one would then utilize this profile to identify points that do not coincide with this description of normal. However, this process tends to be costly computationally, thus limiting the size and dimensionality of the data to be used. By implementing a forest-based technique, called an isolation forest (iForest), one can directly isolate anomalies as opposed to determining what is normal, and avoid the computational costs typically associated with more traditional methods. (This talk is based on work during an internship at JHUAPL.)

Zoom invitation is sent to the seminar list. If you are not on the list, please, contact Arvind Saibaba to get the link