Federated Learning (FL) has gained prominence as a method for facilitating collaborative and privacy-preserving model training across multiple heterogeneous devices in recent years. In most approaches, the clients are closely deployed to the data source. However, as FL systems are implemented in the industry, multiple platform options can be considered in the design phase.In this paper, we present a novel approach for deploying FL clients to multiple locations considering a multi-platform strategy with cloud, fog, and edge resources. We provide a FL architecture that integrates mechanisms for building cohorts of similar clients and a client selection algorithm for optimizing the performance of all clients with respect to energy consumption, model performance, and FL completion time.We evaluate seven deployment strategies in three scenarios given a real-world use case from the electronics industry and heterogeneous hardware capabilities. Our results show that our approach can improve model performance by up to 15%, while energy consumption and completion time converge to the relatively best deployment.

In industrial settings, traditional centralized ap-proaches for training AI models can be insufficient due to limited training data. Industrial Federated Learning (IFL) offers a promising solution by enabling collaborative training across multiple industrial devices, while keeping data on-premises. In this paper, we propose a novel approach for supporting the development, deployment, integration and execution of IFL solutions. The proposed method provides a lifecycle management of FL artifacts and supports FL as a Service (FlaaS). This enables the extensibility and customizability of FL-based edge applications in industrial settings. Additionally, we introduce a federated clustering algorithm that we have integrated into a condition monitoring app running on client locations to evaluate the proposed lifecycle management. We run two scenarios with four and 33 clients using real-world time series data from industrial pumps. Our results show the applicability of the implemented lifecycle management and demonstrates that privacy-preserving approaches compete well with privacy-disclosing ones.

Artificial Intelligence (AI) is increasingly explored in various domains and industries. Many companies experiment with AI, but too often those experiments are one-off analyses based on outdated data and the resulting models never make it into production. This paper proposes a framework for building and operating AI models at the industrial edge. The center of this framework is the model artifact, a model-generating entity. We analyze three AI model use-cases and user roles involved in industrial AI applications to illustrate the challenges in deploying and operating AI applications in industrial edge scenarios. We propose to structure the AI models into predefined artifacts that enable deployments with only a few clicks. The edge device links sensor data with the model input and returns the model output as feedback back into the industrial process. Model training, deployment, and management can be carried out in a scalable manner even by a non-expert. Several models can be managed in parallel, and data can be linked to the respective sensor or machine.

The digital transformation provides industries with unparalleled opportunities for value creation. AI and Machine learning (AI/ML)-driven approaches for data analysis applied to the massive amounts of data steaming from industrial processes can lead to enhanced operation, costs reduction, and powerful decision-making strategies. In this paper we address the problem of Quality Assurance (QA) in industrial manufacturing. We propose Smart Active Sampling (SAS), a new QA sampling strategy for quality inspection outside the production line. Based on the principles of active learning, an AI/ML model trained for quality prediction decides which produced pieces or samples are sent to quality inspection, to further improve its own prediction accuracy. SAS reduces the production of scrap parts due to earlier detection of quality violations. By inspecting a much lower number of samples as compared to traditional random sampling approaches, SAS improves QA efficiency and cuts down quality inspection costs, resulting in an overall smoother operation. We elaborate on some of the challenges faced in smart sampling strategies for quality inspection, describe the main concepts behind SAS, and showcase its application in a real-world manufacturing QA use case, training an AI/ML model for product defect prediction. Compared to a standard random sampling strategy, widely applied today in industrial QA applications, SAS improves model prediction accuracy requiring a significantly lower number of inspected samples, up to five time less samples in the analyzed dataset.

Monitoring of batch production processes is complex and existing solutions do not offer good performance in providing real-time feedback about the state of the process. Therefore, we introduce an AI system that monitors a fermentation process and detects deviations from the normal process execution directly on the edge and provides real-time feedback to the operator, allowing intervention before the process gets out of control. We analyze the accuracy of the novel AI-based approach by carrying out several experiments and compare the outcome with statistical methods as a baseline. The experiments show that the AI-based approach performs significantly better at detecting anomalies in a fermentation process than the statistical methods.

An increasing number of industrial assets are equipped with IoT sensor platforms and the industry now expects data-driven maintenance strategies with minimal deployment costs. However, gathering labeled training data for supervised tasks such as anomaly detection is costly and often difficult to implement in operational environments. Therefore, this work aims to design and implement a solution that reduces the required amount of data for training anomaly classification models on time series sensor data and thereby brings down the overall deployment effort of IoT anomaly detection sensors. We set up several in-lab experiments using three peristaltic pumps and investigated approaches for transferring trained anomaly detection models across assets of the same type. Our experiments achieved promising effectiveness and provide initial evidence that transfer learning could be a suitable strategy for using pre-trained anomaly classification models across industrial assets of the same type with minimal prior labeling and training effort. This work could serve as a starting point for more general, pre-trained sensor data embeddings, applicable to a wide range of assets.

An important prerequisite for determining whether a certain product is producible in any given production facility is an accurate assessment of which production lines and/or the machines are able to execute the necessary production steps. Not only the static information about the capabilities of the machines, but also the conditions of machines and tools are significant for this analysis. Because of the deviation of machine capabilities with increasing deterioration and weary of the equipment, it is also necessary to continuously monitor the status of the machine and analyze the machine conditions. In this paper, we present an approach for generating production plans across multiple factories, considering both static information and dynamic data analysis. Edge devices constantly monitor high frequency machine data and report condensed machine states to an Industrial IoT platform (IIoT). A marketplace within the cloud-application MindSphere enables us to integrate the requirements of the products and the capabilities of the production sites. Customers are be able to evaluate these production plans based on duration, energy consumption, CO2 footprint etc.

Edge computing describes a paradigm for combining computational resources at the edge of the network with the cloud. Even though complementing the cloud with these resources provides benefits, e.g., low latency, it also introduces new challenges to the operational staff. Such challenges can be: deciding if the applications should be placed in the cloud or at the edge, and monitoring them at runtime to ensure that all the application requirements are met. This becomes more challenging when using microservices due to the complexity of the resulting placement problem. To mitigate such concerns, we introduce an automatic deployment framework along with a prototype implementation, called D-DAD. This framework provides a transparent (to the operational staff) way to deploy applications with respect to all their requirements-including the non-functional-using mechanisms for monitoring and adapting the deployments to the available resources in a cloud-edge environment. For evaluating our framework, we provide results from a series of experiments which show how the adaptation mechanism meets the application requirements, including a ~90% reduction of CPU utilization violations, compared to using only the local resources.

Operators of pipeline infrastructure buried underground are in many countries required to ensure that depth of cover—a measure of the quantity of soil covering a pipeline—lie within prescribed bounds. Traditionally, monitoring depth of cover at scale has been carried out qualitatively by means of visual inspection. We proceed instead to rely on airborne remote sensing techniques to obtain densely sampled ground surface point measurements from the pipeline's right of way, from which we determine depth of cover using automated algorithms. Proceeding in our manner presents a reproducible, quantitative approach to monitoring depth of cover, yet the demands thus made by the scale of real-world pipeline monitoring scenarios on compute and storage resources can be substantial. We show that the scalability afforded by the cloud can be leveraged to address such scenarios, distributing the algorithms we employ to take advantage of multiple compute nodes and exploiting elastic storage. While the use case underlying this paper is monitoring depth of cover, our proposed architecture can be applied more broadly to a wide variety of geospatial analytics tasks carried out 'in the large', including change detection, semantic classification or segmentation, or computation of vegetation indices.

The timely discovery, sharing and integration of architectural knowledge (AK) have become critical aspects in enabling the software architects to make meaningful conceptual and technical design decisions and trade-offs. In large-scale organizations particular obstacles in making AK available to architects are a heterogeneous pool of internal and external knowledge sources, poor interoperability between AK management tools and limited support of computational AK reasoning. Therefore we introduce the Continuous Architectural Knowledge Integration (CAKI) approach that combines the continuous integration of internal and external AK sources together with enhanced semantic reasoning and personalization capabilities dedicated to large organizations. Preliminary evaluation results show that CAKI potentially reduces AK search effort by concurrently yielding more diverse and relevant results.

In this work we advance the understanding of software eco-systems research by examining the structure and evolution of the R statistical computing open-source ecosystem. Our research attempts to shed light on the following intriguing question: what makes software ecosystems successful? The approach we follow is to perform a quantitative analysis of the R ecosystem. R is a well-established and popular ecosystem, whose community and marketplace are steadily growing. We assess and quantify the ecosystem throughout its history, and derive metrics on its core software components, the marketplace as well as its community. We use our insights to make observations that are applicable to ecosystems in general, validate existing theories from the literature, and propose a predictive model for the evolution of software packages. Our results show that the success of the ecosystem relies on a strong commitment by a small core of users who support a large and growing community.

Crowdsourcing is an increasingly used model to outsource certain tasks to be carried out by external experts on the Web. Especially when lacking experience or expertise with certain task types, crowdsourcing offers a convenient way to receive instant support. In this paper, we introduce an in-house enterprise crowdsourcing model, which leverages the crowdsourcing concept and transfers it to traditional organizations. Here, a company's staff is considered a crowd that — besides its regularly assigned tasks — can also receive tasks from colleagues from other departments and across hierarchical structures. The aim is to offer instant support and utilize free capacities throughout a large organization more efficiently. In our work, we describe this concept and supporting mechanisms in context of an agile software development use case. However, in contrast to usually crowdsourced microtasks, complex software architectures usually consist of tens and hundreds of connected modules that can be potentially crowdsourced. These technical dependencies between modules require active coordination and interactions between crowd members that process the single artifacts. Hence, technical dependencies of artifacts result in social dependencies of collaborating crowd members that create them. In order to efficiently discover member compositions based on artifact dependencies, we introduce an indexing and discovery approach based on subgraph matching. Typically, assigning tasks to well-rehearsed teams results in more reliable task processing, faster results, and higher quality of work. We evaluate our approach in terms of system scalability and overall applicability by mining and analyzing the popular SourceForge community. We show that our approach of member composition discovery is feasibly in terms of scalability and quality of discovery results. Our findings deliver important input for the design and implementation of supporting information systems for future large-scale collaboration platforms.

Web-based collaborations and processes have become essential in today's business environments. Such processes typically span interactions between people and services across globally distributed companies. Web services and SOA are the defacto technology to implement compositions of humans and services. The increasing complexity of compositions and the distribution of people and services require adaptive and context-aware interaction models. To support complex interaction scenarios, we introduce a mixed service-oriented system composed of both human-provided and Software-Based Services (SBSs) interacting to perform joint activities or to solve emerging problems. However, competencies of people evolve over time, thereby requiring approaches for the automated management of actor skills, reputation, and trust. Discovering the right actor in mixed service-oriented systems is challenging due to scale and temporary nature of collaborations. We present a novel approach addressing the need for flexible involvement of experts and knowledge workers in distributed collaborations. We argue that the automated inference of trust between members is a key factor for successful collaborations. Instead of following a security perspective on trust, we focus on dynamic trust in collaborative networks. We discuss Human-Provided Services (HPSs) and an approach for managing user preferences and network structures. HPS allows experts to offer their skills and capabilities as services that can be requested on demand. Our main contributions center around a context-sensitive trust-based algorithm called ExpertHITS inspired by the concept of hubs and authorities in web-based environments. ExpertHITS takes trust-relations and link properties in social networks into account to estimate the reputation of users.

Open Web-based and social platforms dramatically influence models of work. Today, there is an increasing interest in outsourcing tasks to crowd sourcing environments that guarantee professional processing. The challenge is to gain the customer's confidence by organizing the crowd's mixture of capabilities and structure to become reliable. This work outlines the requirements for a reliable management in crowd computing environments. For that purpose, distinguished crowd members act as responsible points of reference. These members mediate the crowd's workforce, settle agreements, organize activities, schedule tasks, and monitor behavior. At the center of this work we provide a hard/soft constraints scheduling algorithm that integrates existing agreement models for service-oriented systems with crowd computing environments. We outline an architecture that monitors the capabilities of crowd members, triggers agreement violations, and deploys counteractions to compensate service quality degradation.

Flexible interactions in complex social and service-oriented collaboration systems increasingly demand for automated adaptation techniques to optimize partner discovery and selection. Today, applications of complex service-oriented systems can be found in crowd sourcing environments. In such environments, collaborations are typically short-lived and strongly influenced by incentives and actor behavior. As actors prove their reliable and dependable behavior in jointly performed activities, they become increasingly considered as invaluable partners. A social network builds a strong basis to enable successful collaborations between crowd members. In order to keep track of the dynamics in such systems, it is inevitable to apply an autonomous approach to manage social network structures automatically using captured interaction data. Thus, we introduce an adaptation concept that accounts for emerging social relations based on varying interaction behavior of collaboration partners. We describe the foundational concepts for dynamic social link management in Web-based collaborations. We highlight major concerns of computational models in highly dynamic networks and deal with temporal aspects such as supporting the emergence of relations, efficient update mechanisms, and aging of relations.

Social networks have emerged from niche existence to a mass phenomenon. Nowadays, their fundamental concepts, such as managing personal contacts and sharing profile information, are increasingly harnessed for businesses in professional environments. Similar to service-oriented networks, they allow flexible discovery on demand and loose coupling of participants. Establishing social links facilitates cooperation and enables selective sharing of information. Intuitively, one shares more information with his connected neighbors and less or even none with unrelated individuals. Today, information is one of the most important and valuable goods in business networks. Being informed about ongoing collaborations and upcoming trends is a key success factor. Thus, in professional networks, participants aim at strategically establishing connections to enable reliable information flows. In this paper, we especially highlight an opportunistic model that let mediators connect actually unrelated actors in order to benefit from information mediation. We further discuss a framework that implements this model for service-oriented professional virtual communities.

Web-based collaboration and virtual environments supported by various Web 2.0 concepts enable the application of numerous monitoring, mining and analysis tools to study human interactions and team formation processes. The composition of an effective team requires a balance between adequate skill fulfillment and sufficient team connectivity. The underlying interaction structure reflects social behavior and relations of individuals and determines to a large degree how well people can be expected to collaborate. In this paper we address an extended team formation problem that does not only require direct interactions to determine team connectivity but additionally uses implicit recommendations of collaboration partners to support even sparsely connected networks. We provide two heuristics based on Genetic Algorithms and Simulated Annealing for discovering efficient team configurations that yield the best trade-off between skill coverage and team connectivity. Our self-adjusting mechanism aims to discover the best combination of direct interactions and recommendations when deriving connectivity. We evaluate our approach based on multiple configurations of a simulated collaboration network that features close resemblance to real world expert networks. We demonstrate that our algorithm successfully identifies efficient team configurations even when removing up to 40% of experts from various social network configurations.

Web services and service-oriented architecture (SOA) have become the de facto standard for designing distributed and loosely coupled applications. Many service-based applications demand for a mix of interactions between humans and Software-Based Services (SBS). An example is a process model comprising SBS and services provided by human actors. Such applications are difficult to manage due to changing interaction patterns, behavior, and faults resulting from varying conditions in the environment. To address these complexities, we introduce a self-healing approach enabling recovery mechanisms to avoid degraded or stalled systems. The presented work extends the notion of self-healing by considering a mixture of human and service interactions observing their behavior patterns. We present the design and architecture of the VieCure framework supporting fundamental principles for autonomic self-healing strategies. We validate our self-healing approach through simulations.

Complex networks consisting of humans and software services, such as Web-based social and collaborative environments, typically require flexible and context-based interaction models. Due to the dynamics in such systems, networks are in a state of constant flux and change. Several fundamental concepts, including discovery, interactions, task delegations and executions are no longer based on static policies, but need periodic readjustments. Sophisticated adaptation techniques for improving collaborations are within the key research areas in service-oriented systems. In this paper, we introduce an adaptation approach that accounts for emerging trust relations based on varying interaction behavior of network members. We describe a science collaboration scenario that applies adaptive information sharing techniques. In our model, trust evolves from cooperative behavior of collaboration partners. This behavioral trust provides an intuitive grounding for adaptations and optimizations of member compositions and sharing policies. As people prove their reliable and dependable behavior in jointly performed activities, they become increasingly considered as invaluable partners. We describe the foundational concepts, including support for ad-hoc and self-managed collaboration scenarios, and dynamic trust determination supported by SOA concepts. Furthermore, we present a reference architecture, and evaluate its applicability for large-scale collaboration networks.

The evolution towards cross-organizational collaboration and interaction patterns has led to the emergence of scalable, Web services-based composition infrastructures. The success of service-oriented architecture (SOA) was mainly influenced by the standardization of composition languages such as BPEL. However, compositions require humans to be in the loop and ways to interface with people in a service-oriented manner. In this paper, we discuss Human-Provided Services (HPS) enabling the seamless integration of human capabilities in SOA. In complex and large-scale environments, processes might span interactions among partially unknown participants residing in different organizational units. To address the problem of trusted selection of participants, we introduce a mining approach for the automatic inference of trust relations. Unlike a security-based view on trust, our approach relates to the emergence of trust across humans and services from a social perspective.

Mashup tools are becoming increasingly important enabling users to compose services and processes on the Web. Most existing tools focus on Web-based interfaces, usability, and visual languages for creating mashups. A major challenge that has received limited attention is context-awareness and adaptivity of service mashups. In this paper we focus on two main aspects: First, a service capability model describing service characteristics that can be tracked and matched against the requirements associated with service mashups and second an algorithm to recommend refinements such as replacing services within mashups. We implemented a set of adaptation algorithms to validate our approach in real service-oriented systems.

In this paper we present the concept of activity-centric collaboration using service-oriented architectures (ACCUSO), which addresses the requirements arising from ad-hoc collaboration in mobile teams. In ACCUSO, activities are used to map human actions to Web services exploiting the potential benefits of SOA, such as service discovery and binding at run time. The possibility to compose activities hierarchically from sub-activities and to redesign running activities provides the process-flexibility required in ad-hoc collaboration. We expand the notion of service orientation by introducing human-provided services (HpS) which provide functionality not realizable through software services. HpS are "implemented" by human actors (possibly being mobile), which remains transparent to the system, thereby allowing for the provisioning of HpS based on conventional WS-infrastructure. The feasibility and applicability of ACCUSO is demonstrated through a proof-of-concept implementation.

In collaborative working environments (CWE), human interaction patterns represent reoccurring situations describing the sequence and type of interactions between individuals. We believe that such patterns provide information that may be used to improve human collaboration. In this paper we introduce interaction patterns to an existing context sharing platform used by distributed teams. We use these patterns to formulate rules that help determining the relevance of context information between users and that raise team awareness between interacting entities. These rules are integrated in an existing platform for context sharing between mobile users which allows us to demonstrate the practical applicability of our approach