{"id":2328,"date":"2024-10-11T09:16:47","date_gmt":"2024-10-11T09:16:47","guid":{"rendered":"https:\/\/dataninja.nrw\/?page_id=2328"},"modified":"2025-02-10T08:14:48","modified_gmt":"2025-02-10T08:14:48","slug":"gaia-gaussian-processes-for-automatic-and-interpretable-anomaly-detection","status":"publish","type":"page","link":"https:\/\/dataninja.nrw\/?page_id=2328","title":{"rendered":"GAIA: Gaussian Processes for Automatic and Interpretable Anomaly-Detection"},"content":{"rendered":"
\n\n\n

The faces behind project GAIA<\/h3>\n\n\n\n
\n
\n
\"\"<\/figure><\/div>\n\n\n

Andreas Besginow<\/h4>\n\n\n\n

THOWL<\/p>\n<\/div>\n\n\n\n

\n
\"\"<\/figure><\/div>\n\n\n

Jan David H\u00fcwel<\/h4>\n\n\n\n

FernUniversit\u00e4t in Hagen<\/p>\n<\/div>\n<\/div>\n\n\n<\/div>\n\n\n

<\/div>\n\n\n\n

Research for Trustworthy Predictive Models<\/h1>\n\n\n\n
\n
\n
\"\"<\/figure>\n<\/div>\n\n\n\n
\n

Imagine you’re measuring the temperature on your patio every day and writing it down. These are what we call \u201cmeasurement data.\u201d Some days are warmer, some are colder, but overall, it gets warm in the summer and cold in the winter. This overall pattern is called a \u201ctrend.\u201d In our work, we focus on finding or learning these trends in the measurement data. We do this using something called Gaussian Process Models. A Gaussian Process Model is great at \u201clearning\u201d these trends. Even if there are a few unusually warm days in winter, the model still understands that it\u2019s winter. Through learning, Gaussian Processes can also make predictions for the future, such as what the temperatures might be in the coming years.<\/p>\n<\/div>\n<\/div>\n\n\n\n

Our research aims to improve these predictions and use them to detect changes and anomalies in the data. For example, we\u2019ve been able to make Gaussian Processes learn new patterns faster. Additionally, we\u2019ve taught the models some basic physical laws! Right now, we\u2019re working on figuring out when a model is \u201cgood,\u201d so we can compare different models. After all, we don\u2019t want our temperature model to predict several warm days every winter just because it happened once.<\/p>\n\n\n\n

\n
\n

Finally, we also want a Gaussian Process to notice when new data looks different from what it has seen before and explain what makes it different. For instance, if we put up an umbrella on our patio and it doesn’t get as warm during the day anymore, the model should say, \u201cWarning: it\u2019s not getting as warm during the day as before. Something has changed.\u201d<\/p>\n\n\n\n

Overall, with our project, we aim to help understand complex processes better and explain changes in those processes. By providing clear explanations, we can ensure that we can trust the predictions made by our models.<\/p>\n<\/div>\n\n\n\n

\n
\"\"<\/figure>\n<\/div>\n<\/div>\n\n\n\n

Cooperation<\/h3>\n\n\n\n
\n
\n \"Fernuni<\/a>\n <\/div>\n
<\/div>\n
\n

Data Management and Analytics Group<\/p><\/b><\/a>\n

Prof. Dr. Christian Beecks<\/p><\/a>\n

PhD student: Jan David H\u00fcwel<\/a><\/p>\n <\/div>\n<\/div>\n

\n
\n \"\"<\/a>\n <\/div>\n
<\/div>\n
\n

Department of Electrical Engineering and Computer Science<\/p><\/b><\/a>\n

Prof. Dr. Markus Lange-Hegermann<\/p><\/a>\n

PhD student: Andreas Besginow<\/a><\/p>\n <\/div>\n<\/div>\n\n\n\n

Project Publications<\/h3>\n\n\n\n