Fraunhofer Institute for Intelligent Analysis and Information Systems

In den letzten Jahren bieten Technologien der künstlichen Intelligenz (KI) neue Möglichkeiten zur Steigerung von Qualität und Effizienz in der Produktion. Damit das hiermit verbundene Potenzial vollständig ausgeschöpft werden kann, ist der Nachweis der Zuverlässigkeit von KI unabdingbar, insbesondere für den Einsatz der KI in automatisierten Produktionssystemen. Für einen belastbaren Zuverlässigkeitsnachweis sind Tests erforderlich, bei denen die KI-Modelle auf verschiedene Fehlerarten geprüft werden. Dieser Beitrag stellt anhand von drei Anwendungsfällen aus der industriellen Produktion vor, wie mit spezialisierten Prüfwerkzeugen Tests von KI-Modellen durchgeführt werden können, die systematische Schwächen in den KI-Modellen aufdecken. Damit die KI-Tests effizient durchgeführt und Testresultate automatisiert dokumentiert werden, sind die Prüfwerkzeuge in eine skalierbare Prüfplattform integriert.

The field of Information Assurance (IA) and Cybersecurity has seen substantial evolution, driven by advancements in technology and the increasing sophistication of threats in the digital age. This study employs Large Language Models (LLMs), as well as other advanced NLP techniques, to conduct a comprehensive analysis of literature from 1967 to 2024. By analyzing a corpus of more than 62,000 documents extracted from Scopus, our approach involves a comprehensive methodology that includes two main phases: topic detection using BERTopic and automatic summarization with LLMs across various periods (annual and decades). By designing targeted queries to extract relevant papers, analyzing textual data, and applying advanced prompting techniques for summarization, we integrate computational models to handle large volumes of data. Our results demonstrate that an ensemble of methods (Ev2) outperforms traditional summarization and density-based approaches, with improvements ranging from 16.7% to 29.6% in keyword definition tasks. It generates summaries that outperform in 5 out of the 7 tested metrics while maintaining the logical integrity of bibliographic references. Our results illuminate the shifts in focus within Information Assurance across decades, revealing key breakthroughs and forecasting emerging areas of significance.

Methods of artificial intelligence (AI) and especially machine learning (ML) have been growing ever more complex, and at the same time have more and more impact on people’s lives. This leads to explainable AI (XAI) manifesting itself as an important research field that helps humans to better comprehend ML systems. In parallel, quantum machine learning (QML) is emerging with the ongoing improvement of quantum computing hardware combined with its increasing availability via cloud services. QML enables quantum-enhanced ML in which quantum mechanics is exploited to facilitate ML tasks, typically in the form of quantum-classical hybrid algorithms that combine quantum and classical resources. Quantum gates constitute the building blocks of gate-based quantum hardware and form circuits that can be used for quantum computations. For QML applications, quantum circuits are typically parameterized and their parameters are optimized classically such that a suitably defined objective function is minimized. Inspired by XAI, we raise the question of the explainability of such circuits by quantifying the importance of (groups of) gates for specific goals. To this end, we apply the well-established concept of Shapley values. The resulting attributions can be interpreted as explanations for why a specific circuit works well for a given task, improving the understanding of how to construct parameterized (or variational) quantum circuits, and fostering their human interpretability in general. An experimental evaluation on simulators and two superconducting quantum hardware devices demonstrates the benefits of the proposed framework for classification, generative modeling, transpilation, and optimization. Furthermore, our results shed some light on the role of specific gates in popular QML approaches.

Remote sensing and artificial intelligence are pivotal technologies of precision agriculture nowadays. The efficient retrieval of large-scale field imagery combined with machine learning techniques shows success in various tasks like phenotyping, weeding, cropping, and disease control. This work will introduce a machine learning framework for automatized large-scale plant-specific trait annotation for the use case of disease severity scoring for CLS in sugar beet. With concepts of DLDL, special loss functions, and a tailored model architecture, we develop an efficient Vision Transformer based model for disease severity scoring called SugarViT. One novelty in this work is the combination of remote sensing data with environmental parameters of the experimental sites for disease severity prediction. Although the model is evaluated on this special use case, it is held as generic as possible to also be applicable to various image-based classification and regression tasks. With our framework, it is even possible to learn models on multi-objective problems, as we show by a pretraining on environmental metadata. Furthermore, we perform several comparison experiments with state-of-the-art methods and models to constitute our modeling and preprocessing choices.

This study introduces an explainable framework for Automated Fact Verification (AFV) systems, integrating a novel Context-Aware Retrieval and Explanation Generation (CARAG) methodology. CARAG enhances evidence retrieval by leveraging thematic embeddings derived from a Subset of Interest (SOI, a focused subset of the fact-verification dataset) to integrate local and global perspectives. The retrieval process combines these thematic embeddings with claim-specific vectors to refine evidence selection. Retrieved evidence is integrated into an explanation-generation pipeline employing a Large Language Model (LLM) in a zero-shot paradigm, ensuring alignment with topic-based thematic contexts. The SOI and its derived thematic embeddings, supported by a visualized SOI graph, provide transparency into the retrieval process and promote explainability in AI by outlining evidence-selection rationale. CARAG is evaluated using FactVer, a novel explanation-focused dataset curated to enhance AFV transparency. Comparative analysis with standard Retrieval-Augmented Generation (RAG) demonstrates CARAG’s effectiveness in generating contextually aligned explanations, underscoring its potential to advance explainable AFV frameworks.

Renewable energy forecasting is crucial for pollution prevention, management, and long-term sustainability. In response to the challenges associated with energy forecasting, the simultaneous deployment of several data-processing approaches has been used in a variety of studies in order to improve the energy–time-series analysis, finding that, when combined with the wavelet analysis, deep learning techniques can achieve high accuracy in energy forecasting applications. Consequently, we investigate the implementation of various wavelets within the structure of a long short-term memory neural network (LSTM), resulting in the new LSTM wavelet (LSTMW) neural network. In addition, and as an improvement phase, we modeled the uncertainty and incorporated it into the forecast so that systemic biases and deviations could be accounted for (LSTMW with luster: LSTMWL). The models were evaluated using data from six renewable power generation plants in Chile. When compared to other approaches, experimental results show that our method provides a prediction error within an acceptable range, achieving a coefficient of determination (R2) between 0.73 and 0.98 across different test scenarios, and a consistent alignment between forecasted and observed values, particularly during the first 3 prediction steps.

The extent and timescale of climate change impacts remain uncertain, including global temperature increase, sea level rise, and more frequent and intense extreme events. Uncertainties are compounded by cascading effects. Nevertheless, decision-makers must take action. Adaptation pathways, an approach for developing dynamic adaptive policymaking, are widely considered suitable for planning urban or regional climate change adaptation, but often lack integration of measures for disaster risk management. This article emphasizes the need to strengthen Adaptation Pathways by bringing together explicitly slow-onset impacts and sudden climate disasters within the framework of Resilience Pathways. It explores key features of Adaptation Pathways—such as thresholds, performance assessments, and visual tools—to enhance their capacity to address extreme events and foster the integration of Climate Change Adaptation and Disaster Risk Management.

Quadratic unconstrained binary optimization (QUBO) problems are well-studied, not least because they can be approached using contemporary quantum annealing or classical hardware acceleration. However, due to limited precision and hardware noise, the effective set of feasible parameter values is severely restricted. As a result, otherwise solvable problems become harder or even intractable. In this work, we study the implications of solving QUBO problems under limited precision. Specifically, it is shown that the problem’s dynamic range has a crucial impact on the problem’s robustness against distortions. We show this by formalizing the notion of preserving optima between QUBO instances and explore to which extend parameters can be modified without changing the set of minimizing solutions. Based on these insights, we introduce techniques to reduce the dynamic range of a given QUBO instance based on the theoretical bounds of the minimal energy value. An experimental evaluation on random QUBO instances as well as QUBO-encoded BinClustering and SubsetSum problems show that our theoretical findings manifest in practice. Results on quantum annealing hardware show that the performance can be improved drastically when following our methodology.

Standards and specifications (henceforth standards) greatly simplify our everyday lives and the economy in many areas. By establishing quality benchmarks and promoting interoperability, they form a key pillar for international trade and competition. As artificial intelligence (AI) emerges as one of the key technologies of the future, two fundamental questions arise regarding standards: (i) What new standards must be developed specifically for AI? (ii) What modifications to existing standards are necessary to fully leverage the potential of AI? The first question is being addressed actively through a variety of standardisation activities and analyses of the status quo. However, much of the second question remains unanswered. Although the problem is discussed in the literature, there is a lack of both a definition of AI readiness and a systematic analysis of the existing body of standards. This chapter, which originates in the project ‘AI Readiness of Standards’, aims to close this gap. Due to the complexity of the topic, we shall start by focusing on the German body of standards. Nonetheless, we believe our findings and proposals are extrapolatable to other, related contexts. This chapter represents the findings of the abovementioned project: presenting a conceptual definition of AI readiness and, a validated systematic approach for classifying whether an existing standard is AI ready. The procedure outlined here also serves as the foundation for an automated Natural Language Processing (NLP) analysis of the entire German body of standards regarding AI readiness.

Resilience, initially a concept rooted in psychology, has traversed disciplinary boundaries, finding application in fields such as urban planning and development since the 2010s. Despite its broad application, most definitions remain too abstract to allow their practical integration into urban planning and development contexts. Addressing this challenge, the European research projects SHELTER and ARCH offer a practicable integration of resilience with planning and development practices surrounding urban heritage. Following a systemic approach to resilience, both projects integrate perspectives from urban development, climate change adaptation, disaster risk management, and heritage management, supported with tools and guidance to anchor resilience in existing practices. This paper presents the results from both projects, including similarities and differences.

Background: The widespread prevalence of obstructive sleep apnea (OSA) underscores the necessity for effective therapies. Mandibular advancement devices (MADs) have emerged as valid treatment for mild to moderate cases, despite the associated dental side effects. Methods: This study evaluates the nature, onset, and long-term manifestation of these side effects. In the prospective group (n = 12), dental impressions were taken pre-MAD-insertion and at intervals of three, six, nine, and twelve months post-insertion to monitor occlusal alterations. In the retrospective group, participants (n = 8) wearing MADs for 7 years at average underwent lateral cephalogram assessments to compare with pre-treatment X-rays. All participants completed a specific questionnaire. Statistical analysis was performed via t-test and with p < 0.05 as the significance level. Results: The majority of participants in both groups consistently used MADs and reported significant sleep quality improvements, rating common symptoms like jaw tension as negligible. In both the prospective group and the retrospective group, significant reductions in overjet were observed at multiple time points, with the prospective group showing reductions at six months (p = 0.001), nine months (p > 0.001), and twelve months (p = 0.019), while the retrospective group indicated a notable decrease between baseline and follow-up assessments after a mean of seven years of device wear (p = 0.004). A slight overbite increase of 0.2 mm was prospectively observed after one year, whereas a trend towards a minimal decrease over the long term was observed in the retrospective sample (p = 0.003). Noteworthy changes in angle class or lower incisor inclination were absent. Cephalograms revealed significant IOK-NL angle alterations with a mean of 98.2° before and 95.2° upon long-term treatment (p = 0.020). Conclusions: These findings suggest that MADs are effective in treating OSA with minor adverse effects. This study advocates for moderate mandibular protrusion to balance therapeutic efficacy with dental health considerations, crucial for optimizing treatment outcomes. Nonetheless, the limited sample size warrants caution when generalizing these results to the broader population.

In diesem Kapitel wird eines der Basiskonzepte von Industrie 4.0 vorgestellt, die Verwaltungsschale. Die Verwaltungsschale schafft die für Industrie 4.0 notwendige herstellerübergreifende Interoperabilität und stellt die digitalen Informationen für intelligente und nicht-intelligente Assets bereit. Sie bildet insbesondere den kompletten Lebenszyklus der repräsentierten Produkte, Geräte, Maschinen und Anlagen ab und schafft dadurch durchgängige Wertschöpfungsketten. Als digitale Repräsentation bildet die Verwaltungsschale die Basis für autonome Systeme und Anwendungen der Künstlichen Intelligenz im Rahmen der Industrie 4.0. Die Verwaltungsschale kann auch als Umsetzung des Digitalen Zwillings für Industrie 4.0 betrachtet werden.

This study delves into the challenge of efficiently digitalising wind turbine maintenance data, traditionally hindered by non‐standardised formats necessitating manual, expert intervention. Highlighting the discrepancies in past reliability studies based on different key performance indicators (KPIs), the paper underscores the importance of consistent standards, like RDS‐PP, for maintenance data categorisation. Leveraging on established digitalisation workflows, we investigate the efficacy of text classifiers in automating the categorisation process against conventional manual labelling. Results indicate that while classifiers exhibit high performance for specific datasets, their general applicability across diverse wind farms is limited at the present stage. Furthermore, differences in failure rate KPIs derived from manual versus classifier‐processed data reveal uncertainties in both methods. The study suggests that enhanced clarity in maintenance reporting and refined designation systems can lead to more accurate KPIs.

Background: The healthcare sector is currently undergoing a significant transformation, driven by an increased utilization of data. In this evolving landscape, surveys are of pivotal importance to the comprehension of patient needs and preferences. Moreover, the digital affinity of patients and physicians within the healthcare system is reforming the manner in which healthcare services are accessed and delivered. The utilization and donation of data are influencing the future of medical research and treatment, while artificial intelligence (AI) is empowering patients and physicians with knowledge and improving healthcare delivery. Methods: In order to evaluate the opinions of patients and physicians regarding the management of personal health data and the functionality of upcoming data management devices in the context of healthcare digitization, we conducted an exploratory study and designed a survey. The survey focused on a number of key areas, including demographics, experience with digitization, data handling, the identification of needs for upcoming digitization, and AI in healthcare. Results: A total of 40 patients and 15 physicians participated in the survey. The results indicate that data security, timesaving/administrative support, and digital communication are aspects that patients associate with patient-friendly digitization. Based on the responses provided by physicians, it might be concluded that future digital platforms should prioritize usability, time efficacy, data security, and interoperability. Conclusions: In terms of expectations for future digital platforms, there is a notable overlap between the needs expressed by patients and those identified by physicians, particularly in relation to usability, time management, data security, and digital communication. This suggests that the requirements of different stakeholders can be combined in a future system, although individual issues may still require attention.

Recent advancements in speech-based topic segmentation have highlighted the potential of pretrained speech encoders to capture semantic representations directly from speech. Traditionally, topic segmentation has relied on a pipeline approach in which transcripts of the automatic speech recognition systems are generated, followed by text-based segmentation algorithms. In this paper, we introduce an end-to-end scheme that bypasses this conventional two-step process by directly employing semantic speech encoders for segmentation. Focused on the broadcasted news domain, which poses unique challenges due to the diversity of speakers and topics within single recordings, we address the challenge of accessing topic change points efficiently in an end-to-end manner. Furthermore, we propose a new benchmark for spoken news topic segmentation by utilizing a dataset featuring approximately 1000 hours of publicly available recordings across six European languages and including an evaluation set in Hindi to test the model’s cross-domain performance in a cross-lingual, zero-shot scenario. This setup reflects real-world diversity and the need for models adapting to various linguistic settings. Our results demonstrate that while the traditional pipeline approach achieves a state-of-the-art Pk score of 0.2431 for English, our end-to-end model delivers a competitive Pk score of 0.2564. When trained multilingually, these scores further improve to 0.1988 and 0.2370, respectively. To support further research, we release our model along with data preparation scripts, facilitating open research on multilingual spoken news topic segmentation.

Artificial intelligence may greatly enhance the capabilities of law enforcement agencies. However, different (emerging) AI applications create different technical, ethical, legal, and organisational risks. To identify the most promising AI technologies for law enforcement use, the H2020 project ALIGNER has developed a sound methodology consisting of a collaborative technology watch process, followed by three cumulative assessments. This book chapter presents ALIGNER’s AI Technology Impact Assessment, Risk Assessment, and Fundamental Rights Impact Assessment.

Use of techniques derived from generative artificial intelligence (AI), specifically large language models (LLMs), offer a transformative potential on the management of multiple sclerosis (MS). Recent LLMs have exhibited remarkable skills in producing and understanding human-like texts. The integration of AI in imaging applications and the deployment of foundation models for the classification and prognosis of disease course, including disability progression and even therapy response, have received considerable attention. However, the use of LLMs within the context of MS remains relatively underexplored. LLMs have the potential to support several activities related to MS management. Clinical decision support systems could help selecting proper disease-modifying therapies; AI-based tools could leverage unstructured real-world data for research or virtual tutors may provide adaptive education materials for neurologists and people with MS in the foreseeable future. In this focused review, we explore practical applications of LLMs across the continuum of MS management as an initial scope for future analyses, reflecting on regulatory hurdles and the indispensable role of human supervision.

This paper presents the topic review (TR), a novel semi-automatic framework designed to enhance the efficiency and accuracy of literature reviews. By leveraging the capabilities of large language models (LLMs), TR addresses the inefficiencies and error-proneness of traditional review methods, especially in rapidly evolving fields. The framework significantly improves literature review processes by integrating advanced text mining and machine learning techniques. Through a case study approach, TR offers a step-by-step methodology that begins with query generation and refinement, followed by semi-automated text mining to identify relevant articles. LLMs are then employed to extract and categorize key themes and concepts, facilitating an in-depth literature analysis. This approach demonstrates the transformative potential of natural language processing in literature reviews. With an average similarity of 69.56% between generated and indexed keywords, TR effectively manages the growing volume of scientific publications, providing researchers with robust strategies for complex text synthesis and advancing knowledge in various domains. An expert analysis highlights a positive Fleiss’ Kappa score, underscoring the significance and interpretability of the results.

Background Medical text, as part of an electronic health record, is an essential information source in healthcare. Although natural language processing (NLP) techniques for medical text are developing fast, successful transfer into clinical practice has been rare. Especially the hospital domain offers great potential while facing several challenges including many documents per patient, multiple departments and complex interrelated processes. Methods In this work, we survey relevant literature to identify and classify approaches which exploit NLP in the clinical context. Our contribution involves a systematic mapping of related research onto a prototypical patient journey in the hospital, along which medical documents are created, processed and consumed by hospital staff and patients themselves. Specifically, we reviewed which dataset types, dataset languages, model architectures and tasks are researched in current clinical NLP research. Additionally, we extract and analyze major obstacles during development and implementation. We discuss options to address them and argue for a focus on bias mitigation and model explainability. Results While a patient’s hospital journey produces a significant amount of structured and unstructured documents, certain steps and documents receive more research attention than others. Diagnosis, Admission and Discharge are clinical patient steps that are researched often across the surveyed paper. In contrast, our findings reveal significant under-researched areas such as Treatment, Billing, After Care, and Smart Home. Leveraging NLP in these stages can greatly enhance clinical decision-making and patient outcomes. Additionally, clinical NLP models are mostly based on radiology reports, discharge letters and admission notes, even though we have shown that many other documents are produced throughout the patient journey. There is a significant opportunity in analyzing a wider range of medical documents produced throughout the patient journey to improve the applicability and impact of NLP in healthcare. Conclusions Our findings suggest that there is a significant opportunity to leverage NLP approaches to advance clinical decision-making systems, as there remains a considerable understudied potential for the analysis of patient journey data.