Recent publications

October 2022

Measurably Stronger Explanation Reliability Via Model Canonization

Franz Motzkus, Sebastian Lapuschkin, Leander Weber

Network canonization has recently been introduced, restructuring a neural network model into a functionally identical equivalent to which established explanation methods can be applied optimally. In this work, we quantitatively verify the...


October 2022

History Dependent Significance Coding for Incremental Neural Network Compression

Gerhard Tech, Karsten Müller, Thomas Wiegand, Detlev Marpe, Heiko Schwarz, Heiner Kirchhoffer, Wojciech Samek, Jonathan Pfaff, Paul Haase, Daniel Becking

This paper presents an improved probability estimation scheme for the entropy coder of Incremental Neural Network Coding (INNC), which is currently under standardization in ISO/IEC MPEG. Major finding is that the probability of a significant...


October 2022

To Pretrain or Not? A Systematic Analysis of the Benefits of Pretraining in Diabetic Retinopathy

Vignesh Srinivasan, Klaus-Robert Müller, Wojciech Samek, Alexander Binder, Nils Strodthoff, Jackie Ma

In this work, we aim to understand what type of pretraining works reliably in practice and what type of pretraining dataset is best suited to achieve good performance in small target dataset size scenarios. Considering diabetic retinopathy...


October 2022

Ultrawideband Systems and Networks: Beyond C+L -Band

Takeshi Hoshida, Johannes Fischer, Tomoyuki Kato, Vittorio Curri, Wladek Forysiak, Lidia Galdino, David T. Neilson, Pierluigi Poggiolini

In the evolution of optical networks, enhancement of spectral efficiency (SE) enhancement has been the most cost-efficient and thus the main driver for capacity enhancementincrease for decades. As a result, the development of optical transport...


October 2022

Hyperspectral Demosaicing of Snapshot Camera Images Using Deep Learning

Eric Wisotzky, Peter Eisert, Anna Hilsmann, Charul Daudkhane

This work proposes a parallel neural network based demosaicing procedure for single-camera-one-shot-for hyperspectral imaging trained on a new ground truth dataset captured in a controlled environment by a hyperspectral snapshot camera with a 4×4...


September 2022

Experimental Investigation of Information Bit Scrambling for Physical-Layer Security in Coherent Fiber-Optic Systems

Carsten Schmidt-Langhorst, Colja Schubert, Robert Elschner, Robert F. H. Fischer, Robert Emmerich, Johannes Pfeiffer, Fabian Chowanek, In-Ho Baek

We experimentally demonstrate tap-proof coherent optical 640-Gb/s transmission based on encryption-less physical layer security. Information bit scrambling combined with soft-decision error-correction coding yields favorably small security gaps,...


September 2022

Advanced DSP-based Monitoring for Spatially resolved and Wavelength-dependent Amplifier Gain Estimation and Fault Location in C+L-band Systems

Matheus Ribeiro Sena, Colja Schubert, Ronald Freund, Johannes Fischer, Antonio Napoli, Vittorio Curri, Robert Emmerich, Wladek Forysiak, Mohammad Behnam Shariati, Caio Marciano Santos, Pratim Hazarika, Bruno Correia

We study the benefits of applying advanced DSP-based monitoring on multiple wavelength division multiplexing (WDM) channels allocated in the optical grid to infer wavelength-wise characteristics of a C+L-band optical line system. In that context,...


September 2022

Towards High-Capacity THz-Wireless P2MP Communication Systems for 6G

Oliver Stiewe, Colja Schubert, Ronald Freund, Robert Elschner, Stefan Weide, Andreas Maaßen

We present different concepts of a THz-wireless point-to-multipoint (P2MP) communication system based on fast spatio-temporal beam-switching in the THz domain and discuss implications for the system and the DSP design. Characterization results...


August 2022

Perfusion Assessment via Local Remote Photoplethysmography (rPPG)

Benjamin Kossack, Peter Eisert, Anna Hilsmann, Eric Wisotzky, Sebastian Schraven, Brigitta Globke

We present an approach to assess the perfusion of visible human tissue from RGB video files. We show that locally resolved rPPG-signals can be used for intraoperative perfusion analysis and visualization during skin and organ transplantation as...


August 2022

Explaint to not Forget: Defending Against Catastrophic Forgetting with XAI

Sami Ede, Wojciech Samek, Sebastian Lapuschkin, Leander Weber, Serop Baghdadlian, An Nguyen, Dario Zanca

The ability to continuously process and retain new information like we do naturally as humans is a feat that is highly sought after when training neural networks. Unfortunately, the traditional optimization algorithms often require large amounts...


August 2022

Customizing the Appearance of Sparks with Binary Metal Alloys

Philipp Memmel, Wolfgang Schade, Jannis Koch, Mingji Li, Eike Hübner, Felix Lederle, Martin Söftje

Alloys consisting of >65 at. % of a brightly emitting and low-boiling-point metal and a carrier metal allow achieving long-flying deeply colored sparks. Besides the color, branching of sparks is crucial for the visual appearance. Rare-earth...


August 2022

Towards the Interpretability of Deep Learning Models for Human Neuroimaging

Simon M. Hofmann, Klaus-Robert Müller, Wojciech Samek, Arno Villringer, Sebastian Lapuschkin, Frauke Beyer, Markus Loeffler, A. Veronica Witte

Brain-age (BA) estimates based on deep learning are increasingly used as neuroimaging biomarker for brain health; however, the underlying neural features have remained unclear. We combined ensembles of convolutional neural networks with...


July 2022

Characterization of Dispersion-Tailored Silicon Strip Waveguide for Wideband Wavelength Conversion

Hidenobu Muranaka, Colja Schubert, Carsten Schmidt-Langhorst, Tomoyuki Kato, Isaac Sackey, Takeshi Hoshida, Gregor Ronniger, Shun Okada, Tokuharu Kimura, Yu Tanaka, Tsuyoshi Yamamoto

In view of application to wideband wavelength conversion, an SOI waveguide was fabricated and characterized. Conversion of C- band WDM test signals into S- and L- bands in a single waveguide is demonstrated.


July 2022

DSP-Based Link Tomography for Amplifier Gain Estimation and Anomaly Detection in C+L-Band Systems

Matheus Ribeiro Sena, Ronald Freund, Robert Emmerich, Johannes K. Fischer, Mohammad Behnam Shariati, Caio Marciano Santos

In this work, we propose a spatially-resolved and wavelength-dependent DSP-based monitoring scheme to accurately estimate the spectral gain profile of C+L-band in-line Erbium-doped fiber amplifiers deployed in a 280-km single mode fiber link.


July 2022

Bayesian Optimization for Nonlinear System Identification and Pre-distortion in Cognitive Transmitters

Matheus Ribeiro Sena, Ronald Freund, Johannes Fischer, Robert Emmerich, Mustafa Sezer Erkilinc, Mohammad Behnam Shariati, Thomas Dippon

We present a digital signal processing (DSP) scheme that performs hyperparameter tuning (HT) via Bayesian optimization (BO) to autonomously optimize memory tap distribution of Volterra series and adapt parameters used in the synthetization of a...


July 2022

Toward Explainable AI for Regression Models

Simon Letzgus, Klaus-Robert Müller, Wojciech Samek, Grégoire Montavon, Patrick Wagner, Jonas Lederer

While such Explainable AI (XAI) techniques have reached significant popularity for classifiers, so far little attention has been devoted to XAI for regression models (XAIR). In this review, we clarify the fundamental conceptual differences of XAI...


June 2022

Multiparametric MRI for characterization of the basal ganglia and the midbrain

Till M. Schneider, Jackie Ma, Patrick Wagner, Nicolas Behl, Armin Michael Nagel, Mark E. Ladd, Sabine Heiland, Martin Bendszus, Sina Straub

In this joint work with the University of Heidelberg, German Cancer Research Center, University Hospital of Erlangen we showed that multimodal quantitative MR enabled excellent differentiation of a wide spectrum of subcortical nuclei with...


June 2022

Communication-Efficient Federated Distillation via Soft-Label Quantization and Delta Coding

Felix Sattler, Wojciech Samek, Arturo Marban, Roman Rischke

Communication constraints prevent the wide-spread adoption of Federated Learning systems. In this work, we investigate Federated Distillation (FD) from the perspective of communication efficiency by analyzing the effects of active...


June 2022

Adaptive Differential Filters for Fast and Communication-Efficient Federated Learning

Daniel Becking, Karsten Müller, Heiko Schwarz, Heiner Kirchhoffer, Gerhard Tech, Wojciech Samek, Paul Haase

Federated learning (FL) scenarios inherently generate a large communication overhead by frequently transmitting neural network updates between clients and server. In this work, we propose a new scaling method operating at the granularity of...


June 2022

Differentially Private One-Shot Federated Distillation

Haley Hoech, Karsten Müller, Wojciech Samek, Roman Rischke

Federated learning suffers in the case of "non-iid" local datasets, i.e., when the distributions of the clients’ data are heterogeneous. One promising approach to this challenge is the recently proposed method FedAUX, an augmentation of federated...


June 2022

Causes of Outcome Learning: A causal inference-inspired machine learning approach to disentangling common combinations of potential causes of a health outcome

Andreas Rieckmann, Wojciech Samek, Sebastian Lapuschkin, Leila Arras, Piotr Dworzynski, Onyebuchi A. Arah, Naja H. Rod, Claus T. Ekstrom

Nearly all diseases are caused by different combinations of exposures. We present the Causes of Outcome Learning approach (CoOL), which seeks to discover combinations of exposures that lead to an increased risk of a specific outcome in parts of...


June 2022

A Benchmark Dataset for the Ground Truth Evaluation of Neural Network Explanations

Leila Arras, Wojciech Samek, Ahmed Osman

Recently, the field of explainable AI (XAI) has developed methods that provide such explanations for already trained neural networks. So far XAI methods along with their heatmaps were mainly validated qualitatively via human-based assessment, or...


May 2022

Improve the Deep Learning Models in Forestry Based on Explanations and Expertise

Ximeng Cheng, Ali Doosthosseini, Julian Kunkel

This research improves deep learning models based on explanations and expertise. The way is to set the annotation matrix for each training sample. Three image classification tasks in forestry verify the method.


May 2022

Coherent Wireless Link at 300 GHz with 160 Gbit/s Enabled by a Photonic Transmitter

Simon Nellen, Sebastian Lauck, Emilien Peytavit, Pascal Szriftgiser, Martin Schell, Guillaume Ducourn, Björn Globisch

We demonstrate a wireless link at 300 GHz using a fiber-coupled PIN photodiode as the transmitter. We achieved a maximum line rate of 160 Gbit/s with 32QAM modulation. The highest spectral efficiency was achieved with 64QAM at 8 GBaud, i.e. 48...


May 2022

Overview of the Neural Network Compression and Representation (NNR) Standard

Heiner Kirchhoffer, Karsten Müller, Werner Bailer, Fabien Racape, Wojciech Samek, Shan Liu, Miska M. Hannuksela, Paul Haase, Hamed Rezazadegan-Tavakoli, Francesco Cricri, Emre Aksu, Wei Jiang, Wei Wang, Swayambhoo Jain, Shahab Hamidi-Rad

Neural Network Coding and Representation (NNR) is the first international standard for efficient compression of neural networks. The NNR standard contains quantization and an arithmetic coding scheme as core encoding and decoding technologies, as...


May 2022

Towards Auditable AI Systems: From Principles to Practice

Christian Berghoff, Thomas Wiegand, Wojciech Samek, Markus Wenzel, Jona Böddinghaus, Vasilios Danos, Gabrielle Davelaar, Thomas Doms, Heiko Ehrich, Alexandru Forrai, Radu Grosu, Ronan Hamon, Henrik Junklewitz, Matthias Neu, Simon Romanski, Dirk Schlesinger, Jan-Eve Stavesand, Sebastian Steinbach, Arndt von Twickel, Robert Walter, Johannes Weissenböck

Auditing AI systems is a complex endeavour since multiple aspects have to be considered along the AI lifecycle that require multi-disciplinary approaches. AI audit methods and tools are in many cases subject of research and not practically...


May 2022

Low-Loss Bragg-ReflectionWaveguides for On-Chip Time-Bin Entanglement

Hannah Thiel, Moritz Kleinert, Hauke Conradi, Lennart Jehle, Robert Chapman, Stefan Frick, Gregor Weihs, Marita Wagner, Bianca Nardi, Alexander Schlager, Holger Suchomel, Martin Kamp, Sven Hofling, Christian Schneider

We fabricate low-loss AlGaAs Bragg-reflection waveguides for the creation of C-band photon pairs via parametric down-conversion. These photon pairs are used in a hybrid on-chip time-bin entanglement scheme.


April 2022

xxAI - Beyond Explainable AI

Andreas Holzinger, Klaus-Robert Müller, Wojciech Samek, Randy Goebel, Ruth Fong, Taesup Moon

This book takes next steps towards a broader vision for explainable AI in moving beyond explaining classifiers, to include explaining other kinds of models (e.g., unsupervised and reinforcement learning models) via a diverse array of XAI...


April 2022

Explaining the Predictions of Unsupervised Learning Models

Grégoire Montavon, Klaus-Robert Müller, Wojciech Samek, Jacob R. Kauffmann

In this chapter, we review our recently proposed "neuralization-propagation" (NEON) approach for bringing XAI to workhorses of unsupervised learning. NEON first converts the unsupervised model into a functionally equivalent neural network so...


April 2022

ECQx: Explainability-Driven Quantization for Low-Bit and Sparse DNNs

Daniel Becking, Karsten Müller, Wojciech Samek, Sebastian Lapuschkin, Maximilian Dreyer

In this chapter, we develop and describe a novel quantization paradigm for DNNs: Our method leverages concepts of explainable AI (XAI) and concepts of information theory: Instead of assigning weight values based on their distances to the...



Items per page10ǀ20ǀ30
Results 121-150 of 263
<< < 2 3 4 5 6 7 8 9 > >>