K
Klaus-Robert Müller
Researcher at Technical University of Berlin
Publications - 799
Citations - 98394
Klaus-Robert Müller is an academic researcher from Technical University of Berlin. The author has contributed to research in topics: Artificial neural network & Computer science. The author has an hindex of 129, co-authored 764 publications receiving 79391 citations. Previous affiliations of Klaus-Robert Müller include Korea University & University of Tokyo.
Papers
More filters
Proceedings Article
Explanations can be manipulated and geometry is to blame
Ann-Kathrin Dombrowski,Maximilian Alber,Christopher J. Anders,Marcel Ackermann,Klaus-Robert Müller,Pan Kessel +5 more
TL;DR: In this paper, the authors show that explanations can be manipulated arbitrarily by applying visually hardly perceptible perturbations to the input that keep the network's output approximately constant, which is disconcerting for both trust and interpretability.
Proceedings ArticleDOI
Evaluating Recurrent Neural Network Explanations
TL;DR: Using the method that performed best in the authors' experiments, it is shown how specific linguistic phenomena such as the negation in sentiment analysis reflect in terms of relevance patterns, and how the relevance visualization can help to understand the misclassification of individual samples.
Journal ArticleDOI
On the information and representation of non-Euclidean pairwise data
TL;DR: This work shows by systematic modeling of non-Euclidean pairwise data that there exists metric violations which can carry valuable problem specific information and Euclidean and non-metric data can be unified on the level of structural information contained in the data.
Proceedings Article
Optimizing spatio-temporal filters for improving Brain-Computer Interfacing
Guido Dornhege,Benjamin Blankertz,Matthias Krauledat,F. Losch,Gabriel Curio,Klaus-Robert Müller +5 more
TL;DR: This work presents a novel technique that allows the simultaneous optimization of a spatial and a spectral filter enhancing discriminability of multi-channel EEG single-trials and demonstrates the superiority of the proposed algorithm.
Book ChapterDOI
New Methods for Splice Site Recognition
TL;DR: In this paper, the authors pose splice site recognition as a classification problem with the classifier learnt from a labeled data set consisting of only local information around the potential splice sites.