Reducing Overhead for Low-Power Optical Wireless Communications
We demonstrate on-off-keying optical wireless transmissions according to the IEEE P802.15.13 PM-PHY, in which we replace 8b10b line-coding by guided and non-guided data scramblers and compensate the remaining high-pass distortions through a...
Explaining the Decisions of Convolutional and Recurrent Neural Networks
In this chapter we discuss the algorithmic and theoretical underpinnings of layer-wise relevance propagation (LRP), apply the method to a complex model trained for the task of visual question answering (VQA), and demonstrate that it produces...
Beyond Explaining: Opportunities and Challenges of XAI-Based Model Improvement
Explainable Artificial Intelligence (XAI) is an emerging research field bringing transparency to highly complex and opaque machine learning (ML) models. This paper offers a comprehensive overview over techniques that apply XAI practically to...
150 GBd PAM-4 Electrical Signal Generation using SiGe-Based Analog Multiplexer Ic
Analog Multiplexers in Silicon-Germanium technology allow increasing the analog bandwidth of Digital-to-Analog-Converters and enable faster data transfers. We demonstrated 4-level pulse-amplitude modulation (PAM-4) signal transmission at 150 GBd...
Towards the Interpretability of Deep Learning Models for Multi-Modal Neuroimaging: Finding Structural Changes of the Ageing Brain
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...
New Definitions of Human Lymphoid and Follicular Cell Entities in Lymphatic Tissue by Machine Learning
Histological sections of the lymphatic system are usually the basis of static (2D) morphological investigations. Here, we performed a dynamic (4D) analysis of human reactive lymphoid tissue using confocal fluorescent laser microscopy in...
Selection of XAI Methods Matters: Evaluation of Feature Attribution Methods for Oculomotoric Biometric Identification
Substantial advances in oculomotoric biometric identification have been made due to deep neural networks processing non-aggregated time series data that replace methods processing theoretically motivated engineered features. In this work, we...
Time Adaptive Probabilistic Shaping for Combined Optical/THz Links
We investigate the applicability of PAS for outdoor THz wireless links in simulations with weather-dependent loss models. Link performances are evaluated and optimal shaping entropies are determined to adjust error rates to a given FEC threshold....
Edge Cloud based Visual Inspection for Automatic Quality Assurance in Production
We present a remote quality assurance use-case in distributed production sites that can be realized with the powerful capabilities of Artificial Intelligence (AI) combined with real-time video streaming systems and high-speed, low-latency...
Ultra-Broadband Optical Wavelength-Conversion using Nonlinear Multi-Modal Optical Waveguides
Ultra-Broadband Wavelength Conversion is one of the key issues of future optical networks. The physical background of ultra-broadband optical wavelength conversion in a multi-modal Silicon waveguide and methods to optimize its functionality are...