arXiv subjects

Niklas Kühl

Publications and source records attributed to Niklas Kühl.

At least 109 records · Page 6Linked to original sources

arXiv2020

Continuous Adjoint Complement to the Blasius Equation

The manuscript is concerned with a continuous adjoint complement to two-dimensional, incompressible, first-order boundary-layer equations for a flat plate boundary-layer. The text is structured into three parts. The first part demonstrates, that the adjoint complement can be derived in two ways, either following a first simplify then derive or a first derive and then simplify strategy. The simplification step comprises the classical boundary-layer (b.-l.) approximation and the derivation step transfers the primal flow equation into a companion adjoint equation. The second part of the paper comprises the analyses of the coupled primal/adjoint b.-l. framework. This leads to similarity parameters, which turn the Partial-Differential-Equation (PDE) problem into a boundary value problem described by a set of Ordinary-Differential-Equations (ODE) and support the formulation of an adjoint complement to the classical Blasius equation. Opposite to the primal Blasius equation, its adjoint complement consists of two ODEs which can be simplified depending on the treatment of advection. It is shown, that the advective fluxes, which are frequently debated in the literature, vanish for the investigated self-similar b.l. flows. Differences between the primal and the adjoint Blasius framework are discussed against numerical solutions, and analytical expressions are derived for the adjoint b.-l. thickness, wall shear stress and subordinated skin friction and drag coefficients. The analysis also provides an analytical expression for the shape sensitivity to shear driven drag objectives. The third part assesses the predictive agreement between the different Blasius solutions and numerical results for Navier-Stokes simulations of a flat plate b.-l. at Reynolds numbers between 1E+03 <= ReL <= 1E+05 .

physics.flu-dyn↗

arXiv2020

Switching Scheme: A Novel Approach for Handling Incremental Concept Drift in Real-World Data Sets

Machine learning models nowadays play a crucial role for many applications in business and industry. However, models only start adding value as soon as they are deployed into production. One challenge of deployed models is the effect of changing data over time, which is often described with the term concept drift. Due to their nature, concept drifts can severely affect the prediction performance of a machine learning system. In this work, we analyze the effects of concept drift in the context of a real-world data set. For efficient concept drift handling, we introduce the switching scheme which combines the two principles of retraining and updating of a machine learning model. Furthermore, we systematically analyze existing regular adaptation as well as triggered adaptation strategies. The switching scheme is instantiated on New York City taxi data, which is heavily influenced by changing demand patterns over time. We can show that the switching scheme outperforms all other baselines and delivers promising prediction results.

cs.LG↗

arXiv2020

DEAL: Deep Evidential Active Learning for Image Classification

Convolutional Neural Networks (CNNs) have proven to be state-of-the-art models for supervised computer vision tasks, such as image classification. However, large labeled data sets are generally needed for the training and validation of such models. In many domains, unlabeled data is available but labeling is expensive, for instance when specific expert knowledge is required. Active Learning (AL) is one approach to mitigate the problem of limited labeled data. Through selecting the most informative and representative data instances for labeling, AL can contribute to more efficient learning of the model. Recent AL methods for CNNs propose different solutions for the selection of instances to be labeled. However, they do not perform consistently well and are often computationally expensive. In this paper, we propose a novel AL algorithm that efficiently learns from unlabeled data by capturing high prediction uncertainty. By replacing the softmax standard output of a CNN with the parameters of a Dirichlet density, the model learns to identify data instances that contribute efficiently to improving model performance during training. We demonstrate in several experiments with publicly available data that our method consistently outperforms other state-of-the-art AL approaches. It can be easily implemented and does not require extensive computational resources for training. Additionally, we are able to show the benefits of the approach on a real-world medical use case in the field of automated detection of visual signals for pneumonia on chest radiographs.

cs.LG↗

arXiv2020

"Healthy surveillance": Designing a concept for privacy-preserving mask recognition AI in the age of pandemics

The obligation to wear masks in times of pandemics reduces the risk of spreading viruses. In case of the COVID-19 pandemic in 2020, many governments recommended or even obligated their citizens to wear masks as an effective countermeasure. In order to continuously monitor the compliance of this policy measure in public spaces like restaurants or tram stations by public authorities, one scalable and automatable option depicts the application of surveillance systems, i.e., CCTV. However, large-scale monitoring of mask recognition does not only require a well-performing Artificial Intelligence, but also ensure that no privacy issues are introduced, as surveillance is a deterrent for citizens and regulations like General Data Protection Regulation (GDPR) demand strict regulations of such personal data. In this work, we show how a privacy-preserving mask recognition artifact could look like, demonstrate different options for implementation and evaluate performances. Our conceptual deep-learning based Artificial Intelligence is able to achieve detection performances between 95% and 99% in a privacy-friendly setting. On that basis, we elaborate on the trade-off between the level of privacy preservation and Artificial Intelligence performance, i.e. the "price of privacy".

cs.CY↗

arXiv2020

Adjoint Complement to the Volume-of-Fluid Method for Immiscible Flows

The paper is concerned with an adjoint complement to the Volume-of-Fluid (VoF) method for immiscible two-phase flows, e.g. air and water, which is widely used in marine engineering due to its computational efficiency. The particular challenge of the primal and the corresponding adjoint VoF-approach refers to the sharp interface treatment featuring discontinuous physical properties. Both the continuous adjoint two-phase system (integration-by-parts) and the corresponding dual compressive convection schemes (summation-by-parts) are derived for two prominent compressive convection schemes, namely the High Resolution Interface Capturing Scheme (HRIC) and Compressive Interface Capturing Scheme for Arbitrary Meshes (CICSAM). The dual scheme rigorously mirrors the primal Normalized-Variable-Diagram (NVD) stencils. Attention is restricted to steady state applications. Thus both the primal and the dual procedures are performed in pseudo time and the backward integration of the dual approach is performed around the (pseudo-temporal) converged primal field. The paper analyses the primal and adjoint equations for an engineering model problem. An analytical solution to the model problem is initially presented, which displays that the adjoint part does not offer a unique, non-trivial solution. As a remedy, an additional diffusive concentration term is introduced to the adjoint concentration equation. Numerical results obtained from the modified approach are benchmarked against the analytical solution for the model problem. Supplementary, the influence of the modification on the sensitivities obtained from simulations for the two-dimensional flow around a submerged hydrofoil are discussed. The final application refers to a shape-optimization of a generic 3D underwater vehicle and underlines a negligible influence of the free mobility parameter.

physics.flu-dyn↗

arXiv2020

Towards Leveraging End-of-Life Tools as an Asset: Value Co-Creation based on Deep Learning in the Machining Industry

Sustainability is the key concept in the management of products that reached their end-of-life. We propose that end-of-life products have -- besides their value as recyclable assets -- additional value for producer and consumer. We argue this is especially true for the machining industry, where we illustrate an automatic characterization of worn cutting tools to foster value co-creation between tool manufacturer and tool user (customer) in the future. In the work at hand, we present a deep-learning-based computer vision system for the automatic classification of worn tools regarding flank wear and chipping. The resulting Matthews Correlation Coefficient of 0.878 and 0.644 confirms the feasibility of our system based on the VGG-16 network and Gradient Boosting. Based on these first results we derive a research agenda which addresses the need for a more holistic tool characterization by semantic segmentation and assesses the perceived business impact and usability by different user groups.

cs.OH↗

arXiv2020

An Uncertainty-based Human-in-the-loop System for Industrial Tool Wear Analysis

Convolutional neural networks have shown to achieve superior performance on image segmentation tasks. However, convolutional neural networks, operating as black-box systems, generally do not provide a reliable measure about the confidence of their decisions. This leads to various problems in industrial settings, amongst others, inadequate levels of trust from users in the model's outputs as well as a non-compliance with current policy guidelines (e.g., EU AI Strategy). To address these issues, we use uncertainty measures based on Monte-Carlo dropout in the context of a human-in-the-loop system to increase the system's transparency and performance. In particular, we demonstrate the benefits described above on a real-world multi-class image segmentation task of wear analysis in the machining industry. Following previous work, we show that the quality of a prediction correlates with the model's uncertainty. Additionally, we demonstrate that a multiple linear regression using the model's uncertainties as independent variables significantly explains the quality of a prediction (\(R^2=0.718\)). Within the uncertainty-based human-in-the-loop system, the multiple regression aims at identifying failed predictions on an image-level. The system utilizes a human expert to label these failed predictions manually. A simulation study demonstrates that the uncertainty-based human-in-the-loop system increases performance for different levels of human involvement in comparison to a random-based human-in-the-loop system. To ensure generalizability, we show that the presented approach achieves similar results on the publicly available Cityscapes dataset.

cs.CV↗

arXiv2020

AI-based Resource Allocation: Reinforcement Learning for Adaptive Auto-scaling in Serverless Environments

Serverless computing has emerged as a compelling new paradigm of cloud computing models in recent years. It promises the user services at large scale and low cost while eliminating the need for infrastructure management. On cloud provider side, flexible resource management is required to meet fluctuating demand. It can be enabled through automated provisioning and deprovisioning of resources. A common approach among both commercial and open source serverless computing platforms is workload-based auto-scaling, where a designated algorithm scales instances according to the number of incoming requests. In the recently evolving serverless framework Knative a request-based policy is proposed, where the algorithm scales resources by a configured maximum number of requests that can be processed in parallel per instance, the so-called concurrency. As we show in a baseline experiment, this predefined concurrency level can strongly influence the performance of a serverless application. However, identifying the concurrency configuration that yields the highest possible quality of service is a challenging task due to various factors, e.g. varying workload and complex infrastructure characteristics, influencing throughput and latency. While there has been considerable research into intelligent techniques for optimizing auto-scaling for virtual machine provisioning, this topic has not yet been discussed in the area of serverless computing. For this reason, we investigate the applicability of a reinforcement learning approach, which has been proven on dynamic virtual machine provisioning, to request-based auto-scaling in a serverless framework. Our results show that within a limited number of iterations our proposed model learns an effective scaling policy per workload, improving the performance compared to the default auto-scaling configuration.

cs.DC↗

arXiv2020

Handling Concept Drift for Predictions in Business Process Mining

Predictive services nowadays play an important role across all business sectors. However, deployed machine learning models are challenged by changing data streams over time which is described as concept drift. Prediction quality of models can be largely influenced by this phenomenon. Therefore, concept drift is usually handled by retraining of the model. However, current research lacks a recommendation which data should be selected for the retraining of the machine learning model. Therefore, we systematically analyze different data selection strategies in this work. Subsequently, we instantiate our findings on a use case in process mining which is strongly affected by concept drift. We can show that we can improve accuracy from 0.5400 to 0.7010 with concept drift handling. Furthermore, we depict the effects of the different data selection strategies.

cs.LG↗

arXiv2020

How to Learn from Others: Transfer Machine Learning with Additive Regression Models to Improve Sales Forecasting

In a variety of business situations, the introduction or improvement of machine learning approaches is impaired as these cannot draw on existing analytical models. However, in many cases similar problems may have already been solved elsewhere-but the accumulated analytical knowledge cannot be tapped to solve a new problem, e.g., because of privacy barriers. For the particular purpose of sales forecasting for similar entities, we propose a transfer machine learning approach based on additive regression models that lets new entities benefit from models of existing entities. We evaluate the approach on a rich, multi-year dataset of multiple restaurant branches. We differentiate the options to simply transfer models from one branch to another ("zero shot") or to transfer and adapt them. We analyze feasibility and performance against several forecasting benchmarks. The results show the potential of the approach to exploit the collectively available analytical knowledge. Thus, we contribute an approach that is generalizable beyond sales forecasting and the specific use case in particular. In addition, we demonstrate its feasibility for a typical use case as well as the potential for improving forecasting quality. These results should inform academia, as they help to leverage knowledge across various entities, and have immediate practical application in industry.

cs.LG↗

arXiv2020

A network-based transfer learning approach to improve sales forecasting of new products

Data-driven methods -- such as machine learning and time series forecasting -- are widely used for sales forecasting in the food retail domain. However, for newly introduced products insufficient training data is available to train accurate models. In this case, human expert systems are implemented to improve prediction performance. Human experts rely on their implicit and explicit domain knowledge and transfer knowledge about historical sales of similar products to forecast new product sales. By applying the concept of Transfer Learning, we propose an analytical approach to transfer knowledge between listed stock products and new products. A network-based Transfer Learning approach for deep neural networks is designed to investigate the efficiency of Transfer Learning in the domain of food sales forecasting. Furthermore, we examine how knowledge can be shared across different products and how to identify the products most suitable for transfer. To test the proposed approach, we conduct a comprehensive case study for a newly introduced product, based on data of an Austrian food retailing company. The experimental results show, that the prediction accuracy of deep neural networks for food sales forecasting can be effectively increased using the proposed approach.

cs.LG↗

arXiv2020

A New Metric for Lumpy and Intermittent Demand Forecasts: Stock-keeping-oriented Prediction Error Costs

Forecasts of product demand are essential for short- and long-term optimization of logistics and production. Thus, the most accurate prediction possible is desirable. In order to optimally train predictive models, the deviation of the forecast compared to the actual demand needs to be assessed by a proper metric. However, if a metric does not represent the actual prediction error, predictive models are insufficiently optimized and, consequently, will yield inaccurate predictions. The most common metrics such as MAPE or RMSE, however, are not suitable for the evaluation of forecasting errors, especially for lumpy and intermittent demand patterns, as they do not sufficiently account for, e.g., temporal shifts (prediction before or after actual demand) or cost-related aspects. Therefore, we propose a novel metric that, in addition to statistical considerations, also addresses business aspects. Additionally, we evaluate the metric based on simulated and real demand time series from the automotive aftermarket.

cs.LG↗

arXiv2020

Handling Concept Drifts in Regression Problems -- the Error Intersection Approach

Machine learning models are omnipresent for predictions on big data. One challenge of deployed models is the change of the data over time, a phenomenon called concept drift. If not handled correctly, a concept drift can lead to significant mispredictions. We explore a novel approach for concept drift handling, which depicts a strategy to switch between the application of simple and complex machine learning models for regression tasks. We assume that the approach plays out the individual strengths of each model, switching to the simpler model if a drift occurs and switching back to the complex model for typical situations. We instantiate the approach on a real-world data set of taxi demand in New York City, which is prone to multiple drifts, e.g. the weather phenomena of blizzards, resulting in a sudden decrease of taxi demand. We are able to show that our suggested approach outperforms all regarded baselines significantly.

cs.LG↗

arXiv2020

Half-empty or half-full? A Hybrid Approach to Predict Recycling Behavior of Consumers to Increase Reverse Vending Machine Uptime

Reverse Vending Machines (RVMs) are a proven instrument for facilitating closed-loop plastic packaging recycling. A good customer experience at the RVM is crucial for a further proliferation of this technology. Bin full events are the major reason for Reverse Vending Machine (RVM) downtime at the world leader in the RVM market. The paper at hand develops and evaluates an approach based on machine learning and statistical approximation to foresee bin full events and, thus increase uptime of RVMs. Our approach relies on forecasting the hourly time series of returned beverage containers at a given RVM. We contribute by developing and evaluating an approach for hourly forecasts in a retail setting - this combination of application domain and forecast granularity is novel. A trace-driven simulation confirms that the forecasting-based approach leads to less downtime and costs than naive emptying strategies.

cs.LG↗

arXiv2020

Sequential Transfer Machine Learning in Networks: Measuring the Impact of Data and Neural Net Similarity on Transferability

In networks of independent entities that face similar predictive tasks, transfer machine learning enables to re-use and improve neural nets using distributed data sets without the exposure of raw data. As the number of data sets in business networks grows and not every neural net transfer is successful, indicators are needed for its impact on the target performance-its transferability. We perform an empirical study on a unique real-world use case comprised of sales data from six different restaurants. We train and transfer neural nets across these restaurant sales data and measure their transferability. Moreover, we calculate potential indicators for transferability based on divergences of data, data projections and a novel metric for neural net similarity. We obtain significant negative correlations between the transferability and the tested indicators. Our findings allow to choose the transfer path based on these indicators, which improves model performance whilst simultaneously requiring fewer model transfers.

cs.LG↗

arXiv2020

Comfort-as-a-Service: Designing a User-Oriented Thermal Comfort Artifact for Office Buildings

Most people spend up to 90 % of their time indoors. However, literature in the field of facility management and related disciplines mostly focus on energy and cost saving aspects of buildings. Especially in the area of commercial buildings, only few articles take a user-centric perspective and none of them considers the subjectivity of thermal comfort. This work addresses this research gap and aims to optimize individual environmental comfort in open office environments, taking advantage of changes in modern office infrastructure and considering actual user feedback without interfering with existing systems. Based on a Design Science Research approach, we first perform a user experience testing in an exemplary corporate office building. Furthermore, we build a mechanism to gather user feedback on environmental comfort. Based on this, we build a machine learning model including different IoT data sources (e.g. building data and weather data) with an average coefficient of determination of 41.5%. Using these insights, we are able to suggest current individual comfort zones within the building and help employees to make better informed decisions on where to sit or what to wear, to feel comfortable and work productively. Therefore, we contribute to the body of knowledge by proposing a user-centric design within a cross-disciplinary context on the basis of analytical processes.

cs.HC↗

arXiv2020

Machine Learning in Artificial Intelligence: Towards a Common Understanding

The application of "machine learning" and "artificial intelligence" has become popular within the last decade. Both terms are frequently used in science and media, sometimes interchangeably, sometimes with different meanings. In this work, we aim to clarify the relationship between these terms and, in particular, to specify the contribution of machine learning to artificial intelligence. We review relevant literature and present a conceptual framework which clarifies the role of machine learning to build (artificial) intelligent agents. Hence, we seek to provide more terminological clarity and a starting point for (interdisciplinary) discussions and future research.

cs.LG↗

arXiv2020

Cahn-Hilliard Navier-Stokes Simulations for Marine Free-Surface Flows

The paper is devoted to the simulation of maritime two-phase flows of air and water. Emphasis is put on an extension of the classical Volume-of-Fluid (VoF) method by a diffusive contribution derived from a Cahn-Hilliard (CH) model and its benefits for simulating immiscible, incompressible two-phase flows. Such flows are predominantly simulated with implicit VoF schemes, which mostly employ heuristic downwind-biased approximations for the concentration transport to mimic a sharp interface. This strategy introduces a severe time step restriction and requires pseudo time-stepping of steady flows. Our overall goal is a sound description of the free-surface region that alleviates artificial time-step restrictions, facilitates an efficient and robust numerical framework and inherently includes surface tension effects when needed. The approach is verified for an analytical Couette-flow example and the bubble formation under the influence of surface tension forces. 2D Validation examples are concerned with laminar standing waves reaching from gravity to capillary scale as well as a submerged hydrofoil flow. The final application refers to the 3D flow around an experimentally investigated container vessel at fixed floatation for Re=1.4E+07 and Fn=0.26. Results are compared with data obtained from VoF approaches, supplemented by analytical solutions and measurements. The study indicates the superior efficiency, resharpening capability and wider predictive realm of the CH-based extension for free surface flows with a confined spatial range of interface Courant numbers.

physics.comp-ph↗