Search

​© 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., Service customization is a central issue in socio- technical service ecosystems, enabled and fueled by new data- driven approaches, and with the goal of increasing value creation for the customer, and value capture for the provider. In this paper, we address the question of how to design service customization within the provider-customer interaction. We propose a novel quantitative approach for modeling the relation between cus- tomization level at the various steps of the customer journey on the one hand, and its effect on the value created for customer and provider on the other hand. Combining this model with approaches from multi-objective optimization, optimum levels of customization from both the customer and the provider perspective can be determined. Thus, the proposed model allows the identification of service designs which are optimized in terms of their value creation and value capture.

While the field of algorithmic fairness has brought forth many ways to measure and improve the fairness of machine learning models, these findings are still not widely used in practice. We suspect that one reason for this is that the field of algorithmic fairness came up with a lot of definitions of fairness, which are difficult to navigate. The goal of this paper is to provide data scientists with an accessible introduction to group fairness metrics and to give some insight into the philosophical reasoning for caring about these metrics. We will do this by considering in which sense socio-demographic groups are compared for making a statement on fairness.

While the field of algorithmic fairness has brought forth many ways to measure and improve the fairness of machine learning models, these findings are still not widely used in practice. We suspect that one reason for this is that the field of algorithmic fairness came up with a lot of definitions of fairness, which are difficult to navigate. The goal of this paper is to provide data scientists with an accessible introduction to group fairness metrics and to give some insight into the philosophical reasoning for caring about these metrics. We will do this by considering in which sense socio-demographic groups are compared for making a statement on fairness.

​© 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., Service customization is a central issue in socio- technical service ecosystems, enabled and fueled by new data- driven approaches, and with the goal of increasing value creation for the customer, and value capture for the provider. In this paper, we address the question of how to design service customization within the provider-customer interaction. We propose a novel quantitative approach for modeling the relation between cus- tomization level at the various steps of the customer journey on the one hand, and its effect on the value created for customer and provider on the other hand. Combining this model with approaches from multi-objective optimization, optimum levels of customization from both the customer and the provider perspective can be determined. Thus, the proposed model allows the identification of service designs which are optimized in terms of their value creation and value capture.

© 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., The optimization of organic semiconductor devices relies on the determination of material and device parameters. However, these parameters are often not directly measurable or accessible and may change depending on the neighboring materials in the layered stack. Once the parameters are known, devices can be optimized in order to maximize a certain target, e.g. the brightness of a LED. Here, we combine the use of machine learning and a semiconductor device modelling tool to extract the material parameters from measurements. Therefore, we train our machine learning model with synthetic training data originating from a semiconductor simulator. In a second step, the machine learning model is applied to a measured data set and determines the underlying material parameters. This novel and reliable method for the determination of material parameters paves the way to further device performance optimization.

​© 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., When applying deep learning methods in an industrial vision application, they often fall short of the performance shown in a clean and controlled lab environment due to data quality issues. Few would consider the actual labels as a driving factor, yet inaccurate label data can impair model performance significantly. However, being able to mitigate inaccurate or incomplete labels might also be a cost-saver for real-world projects. Here, we survey state-of-the-art deep learning approaches to resolve such missing labels, noisy labels, and partially labeled data in the prospect of an industrial vision application. We systematically present un-, weakly, and semi-supervised approaches from ’A’ like anomaly detection to ’Z’ like zero-shot classification to resolve these challenges by embracing them.

​© 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., When applying deep learning methods in an industrial vision application, they often fall short of the performance shown in a clean and controlled lab environment due to data quality issues. Few would consider the actual labels as a driving factor, yet inaccurate label data can impair model performance significantly. However, being able to mitigate inaccurate or incomplete labels might also be a cost-saver for real-world projects. Here, we survey state-of-the-art deep learning approaches to resolve such missing labels, noisy labels, and partially labeled data in the prospect of an industrial vision application. We systematically present un-, weakly, and semi-supervised approaches from ’A’ like anomaly detection to ’Z’ like zero-shot classification to resolve these challenges by embracing them.

© 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., The optimization of organic semiconductor devices relies on the determination of material and device parameters. However, these parameters are often not directly measurable or accessible and may change depending on the neighboring materials in the layered stack. Once the parameters are known, devices can be optimized in order to maximize a certain target, e.g. the brightness of a LED. Here, we combine the use of machine learning and a semiconductor device modelling tool to extract the material parameters from measurements. Therefore, we train our machine learning model with synthetic training data originating from a semiconductor simulator. In a second step, the machine learning model is applied to a measured data set and determines the underlying material parameters. This novel and reliable method for the determination of material parameters paves the way to further device performance optimization.