"Researchers Improve Security for Smart Systems"

Sensors collect and share large amounts of data to help decision-makers in an increasingly connected and smart world. Through these sensors, people receive ever-increasing amounts of data in ways that can be difficult to decipher. A group of researchers at Washington State University developed a method for statistically analyzing such complex sensor data, allowing computer algorithms that make data-based decisions to be more resilient and capable of dealing with minor errors. The research has many applications, including mobile health, smart homes, the electric power grid, and agriculture. The work, led by Jana Doppa, Huie-Rogers Endowed Chair Associate Professor of Computer Science, is said to be a significant and novel contribution to the security of Machine Learning (ML) systems. ML algorithms are increasingly being used in various applications, such as smart grid management and smart agriculture. For example, they could be used to collect data from sensors in farm fields and weather instruments to learn and predict optimal watering times. Many smart applications collect data in the form of a time series, which is a data set that tracks a sample over time and provides a series of timestamp data points. Although computers collect data and generate time series lines and charts, people are not well suited to read and comprehend them. In addition, they may overlook minor but significant changes, even those that are done maliciously. As more systems incorporate ML, the security of those systems has been an understudied issue, according to Doppa. Adversarial attacks can occur when an attacker gains access to smart sensors and then causes small perturbations in the data that are undetectable to an observer. The perturbations can cause the prediction and decision-making processes to fail. The WSU researchers added a security layer to their ML algorithm to detect potential disturbances and determine how statistically likely they are to occur, providing system resilience and preventing major failures. This article continues to discuss the WSU researchers' work to improve the security of smart systems.

WSU reports "Researchers Improve Security for Smart Systems"