Timezone: »

Deep One-Class Classification
Lukas Ruff · Nico Görnitz · Lucas Deecke · Shoaib Ahmed Siddiqui · Robert Vandermeulen · Alexander Binder · Emmanuel Müller · Marius Kloft

Wed Jul 11 09:15 AM -- 12:00 PM (PDT) @ Hall B #147

Despite the great advances made by deep learning in many machine learning problems, there is a relative dearth of deep learning approaches for anomaly detection. Those approaches which do exist involve networks trained to perform a task other than anomaly detection, namely generative models or compression, which are in turn adapted for use in anomaly detection; they are not trained on an anomaly detection based objective. In this paper we introduce a new anomaly detection method---Deep Support Vector Data Description---, which is trained on an anomaly detection based objective. The adaptation to the deep regime necessitates that our neural network and training procedure satisfy certain properties, which we demonstrate theoretically. We show the effectiveness of our method on MNIST and CIFAR-10 image benchmark datasets as well as on the detection of adversarial examples of GTSRB stop signs.

Author Information

Lukas Ruff (Hasso Plattner Institute)
Nico Görnitz (TU Berlin)

After an internship with the eScience Group, led by David Heckerman (Microsoft Research, Los Angeles, US) in 2014, Nico received a scholarship and is currently enrolled as a research associate in the machine learning group at the Berlin Institute of Technology (TU Berlin, Berlin, Germany) headed by Klaus-Robert Müller. Before, Nico was employed as a research associate from 2010-2014 and during 2010-2012 also affiliated with the Friedrich Miescher Laboratory of the Max Planck Society in Tübingen, where he was co-advised by Gunnar Rätsch. He received a diploma degree (MSc equivalent) in computer engineering (Technische Informatik) from the Berlin Institute of Technology with a thesis in machine learning for computer security in 2010.

Lucas Deecke (University of Edinburgh)
Shoaib Ahmed Siddiqui (German Research Center for Artificial Intelligence)
Robert Vandermeulen (TU Kaiserslautern)
Alexander Binder (Singapore University of Technology and Design)
Emmanuel Müller (Hasso Plattner Institute)
Marius Kloft (TU Kaiserslautern)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors