Keywords: Anomaly Detection, Artificial Intelligence
Anomaly detection is one of the fundamental problems in artificial intelligence.
Anomaly detection systems are broadly applicable in artificial intelligence across a wide set industries and use cases. Our innovation can be used to discover credit-card fraud, detect cyber intrusion, alert predictive maintenance of industrial equipment, medical prevention diagnosis, and for discovering attractive stock market opportunities.
We have created a novel method for detecting anomalies - defined as finding patterns that substantially deviate from those seen previously. Our innovation works by applying a random transformation to the input and predicting the transformation given the transformed input.
Our technology works automatically, in a plug-and-play fashion and without requiring a specific domain or technological knowledge and expertise.
The invention is widely applicable to unstructured data, allowing the selection of unlimited number of transformations.