Springer
Refine
Year of publication
- 2023 (15) (remove)
Institute
- Fachbereich Elektrotechnik und Informationstechnik (4)
- Fachbereich Medizintechnik und Technomathematik (4)
- Fachbereich Luft- und Raumfahrttechnik (3)
- Fachbereich Wirtschaftswissenschaften (3)
- Fachbereich Chemie und Biotechnologie (2)
- MASKOR Institut für Mobile Autonome Systeme und Kognitive Robotik (2)
- ECSM European Center for Sustainable Mobility (1)
- Fachbereich Energietechnik (1)
- IfB - Institut für Bioengineering (1)
- Institut fuer Angewandte Polymerchemie (1)
Language
- English (15)
Document Type
- Conference Proceeding (7)
- Article (4)
- Part of a Book (3)
- Book (1)
Keywords
- Natural language processing (3)
- Information extraction (2)
- Active learning (1)
- Agent-based simulation (1)
- Autonomy (1)
- Best practice sharing (1)
- CFD (1)
- Charging station (1)
- Chondroitin sulfate (1)
- Clustering (1)
Modern implementations of driver assistance systems are evolving from a pure driver assistance to a independently acting automation system. Still these systems are not covering the full vehicle usage range, also called operational design domain, which require the human driver as fall-back mechanism. Transition of control and potential minimum risk manoeuvres are currently research topics and will bridge the gap until full autonomous vehicles are available. The authors showed in a demonstration that the transition of control mechanisms can be further improved by usage of communication technology. Receiving the incident type and position information by usage of standardised vehicle to everything (V2X) messages can improve the driver safety and comfort level. The connected and automated vehicle’s software framework can take this information to plan areas where the driver should take back control by initiating a transition of control which can be followed by a minimum risk manoeuvre in case of an unresponsive driver. This transition of control has been implemented in a test vehicle and was presented to the public during the IEEE IV2022 (IEEE Intelligent Vehicle Symposium) in Aachen, Germany.
In recent years, the development of large pretrained language models, such as BERT and GPT, significantly improved information extraction systems on various tasks, including relation classification. State-of-the-art systems are highly accurate on scientific benchmarks. A lack of explainability is currently a complicating factor in many real-world applications. Comprehensible systems are necessary to prevent biased, counterintuitive, or harmful decisions.
We introduce semantic extents, a concept to analyze decision patterns for the relation classification task. Semantic extents are the most influential parts of texts concerning classification decisions. Our definition allows similar procedures to determine semantic extents for humans and models. We provide an annotation tool and a software framework to determine semantic extents for humans and models conveniently and reproducibly. Comparing both reveals that models tend to learn shortcut patterns from data. These patterns are hard to detect with current interpretability methods, such as input reductions. Our approach can help detect and eliminate spurious decision patterns during model development. Semantic extents can increase the reliability and security of natural language processing systems. Semantic extents are an essential step in enabling applications in critical areas like healthcare or finance. Moreover, our work opens new research directions for developing methods to explain deep learning models.
Supervised machine learning and deep learning require a large amount of labeled data, which data scientists obtain in a manual, and time-consuming annotation process. To mitigate this challenge, Active Learning (AL) proposes promising data points to annotators they annotate next instead of a subsequent or random sample. This method is supposed to save annotation effort while maintaining model performance.
However, practitioners face many AL strategies for different tasks and need an empirical basis to choose between them. Surveys categorize AL strategies into taxonomies without performance indications. Presentations of novel AL strategies compare the performance to a small subset of strategies. Our contribution addresses the empirical basis by introducing a reproducible active learning evaluation (ALE) framework for the comparative evaluation of AL strategies in NLP.
The framework allows the implementation of AL strategies with low effort and a fair data-driven comparison through defining and tracking experiment parameters (e.g., initial dataset size, number of data points per query step, and the budget). ALE helps practitioners to make more informed decisions, and researchers can focus on developing new, effective AL strategies and deriving best practices for specific use cases. With best practices, practitioners can lower their annotation costs. We present a case study to illustrate how to use the framework.
Messenger apps like WhatsApp and Telegram are frequently used for everyday communication, but they can also be utilized as a platform for illegal activity. Telegram allows public groups with up to 200.000 participants. Criminals use these public groups for trading illegal commodities and services, which becomes a concern for law enforcement agencies, who manually monitor suspicious activity in these chat rooms. This research demonstrates how natural language processing (NLP) can assist in analyzing these chat rooms, providing an explorative overview of the domain and facilitating purposeful analyses of user behavior. We provide a publicly available corpus of annotated text messages with entities and relations from four self-proclaimed black market chat rooms. Our pipeline approach aggregates the extracted product attributes from user messages to profiles and uses these with their sold products as features for clustering. The extracted structured information is the foundation for further data exploration, such as identifying the top vendors or fine-granular price analyses. Our evaluation shows that pretrained word vectors perform better for unsupervised clustering than state-of-the-art transformer models, while the latter is still superior for sequence labeling.
Like all preceding transformations of the manufacturing industry, the large-scale usage of production data will reshape the role of humans within the sociotechnical production ecosystem. To ensure that this transformation creates work systems in which employees are empowered, productive, healthy, and motivated, the transformation must be guided by principles of and research on human-centered work design. Specifically, measures must be taken at all levels of work design, ranging from (1) the work tasks to (2) the working conditions to (3) the organizational level and (4) the supra-organizational level. We present selected research across all four levels that showcase the opportunities and requirements that surface when striving for human-centered work design for the Internet of Production (IoP). (1) On the work task level, we illustrate the user-centered design of human-robot collaboration (HRC) and process planning in the composite industry as well as user-centered design factors for cognitive assistance systems. (2) On the working conditions level, we present a newly developed framework for the classification of HRC workplaces. (3) Moving to the organizational level, we show how corporate data can be used to facilitate best practice sharing in production networks, and we discuss the implications of the IoP for new leadership models. Finally, (4) on the supra-organizational level, we examine overarching ethical dimensions, investigating, e.g., how the new work contexts affect our understanding of responsibility and normative values such as autonomy and privacy. Overall, these interdisciplinary research perspectives highlight the importance and necessary scope of considering the human factor in the IoP.