no code implementations • 20 Dec 2022 • Steffen Illium, Robert Müller, Andreas Sedlmeier, Claudia-Linnhoff Popien
We apply the vision transformer, a deep machine learning model build around the attention mechanism, on mel-spectrogram representations of raw audio recordings.
no code implementations • 10 Aug 2022 • Fabian Ritz, Thomy Phan, Andreas Sedlmeier, Philipp Altmann, Jan Wieghardt, Reiner Schmid, Horst Sauer, Cornel Klein, Claudia Linnhoff-Popien, Thomas Gabor
We define a comprehensive SD process model for ML that encompasses most tasks and artifacts described in the literature in a consistent way.
no code implementations • 14 Dec 2021 • Andreas Sedlmeier, Michael Kölle, Robert Müller, Leo Baudrexel, Claudia Linnhoff-Popien
In this work, we analyze existing and propose new metrics for the detection and quantification of multimodal uncertainty in RL based World Models.
no code implementations • 14 Dec 2020 • Fabian Ritz, Thomy Phan, Robert Müller, Thomas Gabor, Andreas Sedlmeier, Marc Zeller, Jan Wieghardt, Reiner Schmid, Horst Sauer, Cornel Klein, Claudia Linnhoff-Popien
A characteristic of reinforcement learning is the ability to develop unforeseen strategies when solving problems.
Multi-agent Reinforcement Learning reinforcement-learning +1
no code implementations • 11 Aug 2020 • Steffen Illium, Robert Müller, Andreas Sedlmeier, Claudia Linnhoff-Popien
In many fields of research, labeled datasets are hard to acquire.
no code implementations • 25 May 2020 • Andreas Sedlmeier, Robert Müller, Steffen Illium, Claudia Linnhoff-Popien
One critical prerequisite for the deployment of reinforcement learning systems in the real world is the ability to reliably detect situations on which the agent was not trained.
no code implementations • 11 Apr 2020 • Sebastian Feld, Andreas Sedlmeier, Markus Friedrich, Jan Franz, Lenz Belzner
Agents of LBS, such as mobile robots or non-player characters in computer games, may use the context surprise to focus more on important regions of a map for a better use or understanding of the floor plan.
no code implementations • 11 Apr 2020 • Sebastian Feld, Steffen Illium, Andreas Sedlmeier, Lenz Belzner
In the near future, more and more machines will perform tasks in the vicinity of human spaces or support them directly in their spatially bound activities.
no code implementations • 31 Dec 2019 • Andreas Sedlmeier, Thomas Gabor, Thomy Phan, Lenz Belzner, Claudia Linnhoff-Popien
We further present a first viable solution for calculating a dynamic classification threshold, based on the uncertainty distribution of the training data.
no code implementations • 8 Jan 2019 • Andreas Sedlmeier, Thomas Gabor, Thomy Phan, Lenz Belzner, Claudia Linnhoff-Popien
Although prior work has shown that dropout-based variational inference techniques and bootstrap-based approaches can be used to model epistemic uncertainty, the suitability for detecting OOD samples in deep reinforcement learning remains an open question.