Abstract: 

Anomaly Detection techniques have been around for decades, but they often remain ineffective in real-world situations because they produce so many false positives. Sifting through the noise is rarely worth the effort. This talk will present a new technique we call "novelty detection" which uses the freely available "Quine" streaming graph to score incoming event data immediately. This technique is able to use categorical data directly instead requiring the traditional one-hot encoding (or other encodings) and makes use of context to accurately score events never seen before. The end result of this process is a live stream of real-time explanations and "novelty scores" which provide a total-ordering of how unusual each observation is compared to all data seen so far. This unsupervised approach requires no training or data labeling, is very fast, and incredibly accurate—producing actionable results and avoiding the false positives.

Bio: 

Ryan Wright is the creator of Quine, and has been leading software teams focused on data infrastructure and data science for two decades. He has served as principal engineer, director of engineering, principal investigator on DARPA-funded research programs, and is currently the founder and CEO of thatDot—the company supporting Quine. Ryan particularly enjoys taking the philosophical ends of computer science—usually problems related to language, meaning, and data—and making them more practical.