BACKGROUND:The online food environment has transformed how individuals access and engage with food by amplifying food availability and promotions. Artificial intelligence (AI) offers scalable solutions to investigate and monitor this space, yet its application across food environment dimensions has not been systematically reviewed. OBJECTIVE:This systematic review examined how AI methods have been applied to assess key online food environment dimensions, including food availability, marketing, price, and nutritional properties. METHODS:We conducted a systematic review, searching Scopus, Web of Science, PubMed, and ABI/INFORM in 2024. Eligible studies were English-language and applied AI to assess online food environment platforms. Titles and abstracts were screened using an AI-based tool (ASReview) and manual review, followed by full-text screening. We assessed food environment setting and dimension, AI method, subfield, algorithm and study quality. RESULTS:In total, 19 studies applied different AI methods, mainly supervised learning, natural language processing (NLP), and computer vision across multiple platforms. Food availability was most frequently assessed, followed by nutritional properties and marketing. Fewer studies addressed pricing or individual exposure. Applications included food and outlet classification, caloric estimation, and consumer engagement analysis. Rule-based NLP enabled large-scale data collection. CONCLUSIONS:AI is increasingly applied to analyze the online food environment, particularly food availability, marketing, and nutrition. Research remains fragmented across platforms and regions, with limited comparative evaluations and standardization. To improve scalability and impact, future efforts should prioritize FAIR-compliant (Findable, Accessible, Interoperable, and Reusable) infrastructure, semantic data integration, multimodal transfer learning, and better alignment with public policy goals.

GDPR defines six legal bases, at least one of which needs to be followed in order to process (or share) personally identifiable data in a lawful manner. Most of the research today is centered around the legal bases of consent and contracts. This limits the options for legal bases that one can select (or use) for data sharing, especially in circumstances where there is a need to use mul-tiple legal bases. For example, one can consent to share data but may want to place restrictions on how it can be used, which requires a license (an extension/add-on to data sharing contracts) in scenarios, where digital assets licensing is involved. Overcoming these limitations and en-abling data sharing via multiple legal bases require combining multiple legal bases. However, incorporating additional (or multiple) legal bases, such as licenses (as an add-on to contracts), in a GDPR-compliant manner remains a challenging task. This is because combining multiple legal bases requires an understanding of each individual legal basis—a task challenging in it-self—and designing a system in a manner that is both compliant with regulatory requirements and practically pertinent. Therefore, in this paper, we present our semantic-based approach and tool that enables GDPR-compliant data sharing via multiple legal bases, consent, and contracts (using licenses as an add-on). This work extends our previous work, GDPR Contract Com-pliance Verification (CCV) tool, which enables GDPR-compliant data sharing via consent and contracts only. We add licenses as a further add-on to contracts, make our previous work more semantically compliant by utilizing SHACL validation for compliance checking, secure the con-tract signing process with digital signatures, introduce SHACL repairs to automatically fix data inconsistencies, and evaluate the performance of the tool and the SHACL components. We demonstrate the effectiveness of SHACL and the enhancement of the tool with GDPR-complaint data sharing based on multiple legal bases by performance testing.

Food and planetary health are interconnected, as the food system is responsible for 26% of greenhouse gas emissions while simultaneously necessary for optimal human health. Optimization and balancing of this relationship in our current and future foods is an active area of research. Integrating different data sources for this purpose remains difficult due to the heterogeneity of food data. To support food experts in complex interdisciplinary information retrieval we developed NICE-Food KG, a knowledge graph and data pipeline integrating data on Nutrition, Ingredients, Contaminants and Environmental impact for the Netherlands. Through data annotation and mapping employing food related ontologies and Resource Description Framework (RDF), NICE-Food KG enhances the FAIR (Findable, Accessible, Interoperable, Reusable) principles of food data. We used NICE-Food KG 1) to assess the data overlap between the different NICE domains 2) for the identification of communal food groups across the integrated data 3) to provide food recommendation based on specific food preferences 4) to infer knowledge on contemporary branded products such as meat and fish alternatives. Although more data is needed and expansions of ontologies are required to confidently bridge the gap between different disciplines in food sciences, we provide a proof of concept of how a knowledge graph supported approach can be used to integrate interdisciplinary food data, harnessing the value of FAIR data.

Soil health is fundamental to environmental sustainability and food security, yet relevant knowledge remains fragmented across diverse sources, hindering its effective application. Knowledge graphs (KGs) offer a robust solution by integrating disparate information into a structured, semantically rich format. Addressing this need, this paper presents an ontology-compliant soil health KG derived from domain literature, and the semi-automated, human-in-the-loop pipeline developed to construct it. Our pipeline leverages large language models to accelerate knowledge extraction, while incorporating expert oversight to ensure ontological compliance and accuracy. The resulting KG integrates unstructured knowledge into 11,715 RDF triples representing 2,018 entities, including 1,786 soil-related concepts. The KG’s fidelity was confirmed by soil scientists through a validation process involving competency questions. The KG, supporting ontology, and the source code of the pipeline are made publicly available.

In recent years, there have been many developments for GDPR-compliant data access and sharing based on consent. For more complex data sharing scenarios, where consent might not be sufficient, many parties rely on contracts. Before a contract is signed, it must undergo the process of contract negotiation within the contract lifecycle, which consists of negotiating the obligations associated with the contract. Contract compliance verification (CCV) provides a means to verify whether a contract is GDPR-compliant, i.e., adheres to legal obligations and there are no violations. The rise of knowledge graph (KG) adoption, enabling semantic interoperability using well-defined semantics, allows CCV to be applied on KGs. In the scenario of different participants negotiating obligations, there is a need for data consistency to ensure that CCV is done correctly. Recent work introduced the automated contracting tool (ACT), a KG-based and ODRL-employing tool for GDPR CCV, which was developed in the Horizon 2020 project smashHit (https://smashhit.eu). Although the tool reports violations with respect to obligations, it had limitations in verifying and ensuring compliance, as it did not use an interoperable semantic formalism, such as SHACL, and did not support users in resolving data inconsistencies. In this work, we propose a novel approach to overcome these limitations of ACT. We semi-automatically resolve CCV inconsistencies by providing repair strategies, which automatically propose (optimal) solutions to the user to re-establish data consistency and thereby support them in managing GDPR-compliant contract lifecycle data. We have implemented the approach, integrated it into ACT and tested its correctness and performance against basic CCV consistency requirements.

We propose KGTN-ens, a framework extending the recent Knowledge Graph Transfer Network (KGTN) in order to incorporate multiple knowledge graph embeddings at a small cost. There are many real-world scenarios in which the amount of data is severely limited (e.g. health industry, rare anomalies). Prior knowledge can be used to tackle this task. In KGTN, one can use a single knowledge source at once. The purpose of this study is to investigate the possibility of combining multiple knowledge sources. We evaluate it with different embeddings in a few-shot image classification task. Our model is partially trained on k ∈{ 1, 2, 5, 10} samples. We also construct a new knowledge source – Wikidata embeddings – and evaluate it with KGTN and KGTN-ens. With ResNet50, our approach outperforms KGTN in terms of the top-5 accuracy on the ImageNet-FS dataset for the majority of tested settings. For k ∈{ 1, 2, 5, 10} respectively, we obtained +0.63/+0.58/+0.43/+0.26 pp. (novel classes) and +0.26/+0.25/+0.32/–0.04 pp. (all classes).

General Data Protection Regulation (GDPR) is compulsory for processing personally identifiable data (PID) across Europe after 2018. Specifically when sharing research data, we cannot process PID without a legal basis defined by GDPR. To facilitate this, we present a scalable and interoperable automated Contract Compliance Verification (CCV) tool that enables GDPR-compliant contract management and data sharing. With the implementation of two scenarios in smart cities and insurance domains, we show how CCV is helpful to process PID and make data collection and integration (e.g., from crowdsourcing) easier.

This paper presents our scalable and interoperable tool for GDPR (General Data Protection Regulation) consent management and automated compliance verification. The tool enables GDPR-compliant data sharing and is beneficial to the industries that process personally identifiable data. The tool has been designed following the GDPR data protection by design principles and has been successfully validated against real-world industrial use case scenarios in smart cities and insurance.

Climate change is a heated discussion topic in public arenas such as social media. Both texts and visuals play key roles in the debate, as they can complement, contradict, or reinforce each other in nuanced ways. It is therefore urgently needed to study the messages as multimodal objects to better understand the polarized debate about climate change impacts and policies. Multimodal representation models such as CLIP are known to be able to transfer knowledge across domains and modalities, enabling the investigation of textual and visual semantics together. Yet they are not directly able to distinguish the nuances between supporting and sceptic climate change stances. This paper explores a simple but effective strategy combining modality fusion and domain-knowledge enhancing to prepare CLIP-based models with knowledge of climate change stances. A multimodal Dutch Twitter dataset is collected and experimented with the proposed strategy, which increased the macro-average F1 score across stances from 51% to 86%. The outcomes can be applied in both data science and public policy studies, to better analyse how the combined use of texts and visuals generates meanings during debates, in the context of climate change and beyond.

Sharing Internet of Things (IoT) data across different sectors, such as in smart cities, becomes complex due to heterogeneity. This poses challenges related to a lack of interoperability, data quality issues and lack of context information, and a lack of data veracity (or accuracy). In addition, there are privacy concerns as IoT data may contain personally identifiable information. To address the above challenges, this paper presents a novel semantic technology-based framework that enables data sharing in a GDPR-compliant manner while ensuring that the data shared is interoperable, contains required context information, is of acceptable quality, and is accurate and trustworthy. The proposed framework also accounts for the edge/fog, an upcoming computing paradigm for the IoT to support real-time decisions. We evaluate the performance of the proposed framework with two different edge and fog-edge scenarios using resource-constrained IoT devices, such as the Raspberry Pi. In addition, we also evaluate shared data quality, interoperability and veracity. Our key finding is that the proposed framework can be employed on IoT devices with limited resources due to its low CPU and memory utilization for analytics operations and data transformation and migration operations. The low overhead of the framework supports real-time decision making. In addition, the 100% accuracy of our evaluation of the data quality and veracity based on 180 different observations demonstrates that the proposed framework can guarantee both data quality and veracity.

The General Data Protection Regulation (GDPR) has imposed strict requirements for data sharing, one of which is informed consent. A common way to request consent online is via cookies. However, commonly, users accept online cookies being unaware of the meaning of the given consent and the following implications. Once consent is given, the cookie “disappears”, and one forgets that consent was given in the first place. Retrieving cookies and consent logs becomes challenging, as most information is stored in the specific Internet browser’s logs. To make users aware of the data sharing implied by cookie consent and to support transparency and traceability within systems, we present a knowledge graph (KG) based tool for personalised cookie consent information visualisation. The KG is based on the OntoCookie ontology, which models cookies in a machine-readable format and supports data interpretability across domains. Evaluation results confirm that the users’ comprehension of the data shared through cookies is vague and insufficient. Furthermore, our work has resulted in an increase of 47.5% in the users’ willingness to be cautious when viewing cookie banners before giving consent. These and other evaluation results confirm that our cookie data visualisation approach and tool help to increase users’ awareness of cookies and data sharing.

The acceptance of the GDPR legislation in 2018 started a new technological shift towards achieving transparency. GDPR put focus on the concept of informed consent applicable for data processing, which led to an increase of the responsibilities regarding data sharing for both end users and companies. This paper presents a literature survey of existing solutions that use semantic technology for implementing consent. The main focus is on ontologies, how they are used for consent representation and for consent management in combination with other technologies such as blockchain. We also focus on visualisation solutions aimed at improving individuals’ consent comprehension. Finally, based on the overviewed state of the art we propose best practices for consent implementation.

It is already known that the diet of the world’s population has a massive impact on climate change. However, how climate change affects the growing conditions of ingredients for different foods and beverages, and emission rates due to, for example, production and logistics are still not known. In this work, different datasets have been explored to study the feasibility of interlinking datasets to automatically generate alternatives for climate change-sensitive food items selection and substitution. A core question to be answered is what the alternatives of the mostly consumed crops in current diets in the Netherlands in case of a climate change can be. The main crop attributes taken into account are nutritional composition and the growing conditions. The growing conditions of three most-consumed crops in the Netherlands have been linked manually to their nutritional composition data and a corresponding knowledge graph is created. This study shows that linking various data semantically promises to generate alternatives automatically.

Smell and taste disorders have become a more prominent issue due to their association with Covid-19, and their impact on quality of life and health outcomes. However, pertinent information regarding these disorders is often inaccessible and poorly organized, with the majority of data stored solely in clinical data repositories. To rectify this, a technological solution capable of digitizing, semantically modeling, and integrating health data is necessary. The knowledge graph, an emerging technology capable of organizing inconsistent and heterogeneous health data and inferring implicit knowledge, presents a viable solution to this problem. In pursuit of the aforementioned goal, an existing ontology pertaining to smell and taste disorders was enriched by introducing additional relevant concepts and relationships. Subsequently, a knowledge graph was constructed based on the defined ontology and patients' data. The resultant knowledge graph was subjected to a rigorous evaluation, encompassing dimensions such as completeness, coherency, coverage, and succinctness. The evaluation established the effectiveness and usability of the knowledge graph, with only minor issues detected through the OOPS! pitfall scanner. Furthermore, as a proof-of-concept for clinical application, a user interface was created, enabling users to access pertinent information concerning smell and taste disorders, including causative factors, medications, and etiology, among others. The interface generates a graph-based structure based on the selected question from a drop-down menu. The end-user can modify the query by merely clicking on the generated graph to ask related questions. This study showcases the potential of knowledge graphs centered on smell and taste disorders to organize and provide accessible health data to end-users.

Food security is currently a major concern due to the growing global population, the exponential increase in food demand, the deterioration of soil quality, the occurrence of numerous diseases, and the effects of climate change on crop yield. Sustainable agriculture is necessary to solve this food security challenge. Disruptive tech-nologies, such as of artificial intelligence, especially, deep learning techniques can contribute to agricultural sustainability. For example, applying deep learning techniques for early disease classification allows us to take timely action, thereby helping to increase the yield without inflicting unnecessary environmental damage, such as excessive use of fertilisers or pesticides. Several studies have been conducted on agricultural sustainability using deep learning techniques and also semantic web technologies such as ontologies and knowledge graphs. However, the three major challenges remain: (i) the lack of explainability of deep learning-based systems (e.g. disease information), especially to non-experts like farmers; (ii) a lack of contextual information (e.g. soil or plant information) and domain-expert knowledge in deep learning-based systems; and (iii) the lack of pattern learning ability of systems based on the semantic web, despite their ability to incorporate domain knowledge. Therefore, this paper presents the work on disease classification, addressing the challenges as mentioned earlier by combining deep learning and semantic web technologies, namely ontologies and knowledge graphs. The findings are: (i) 0.905 (90.5%) prediction accuracy on large noisy dataset; (ii) ability to generate user-level explanations about disease and incorporate contextual and domain knowledge; (iii) the average prediction latency of 3.8514 s on 5268 samples; (iv) 95% of users finding the explanation of the proposed method useful; and (v) 85% of users being able to understand generated explanations easily-show that the proposed method is superior to the state-of-the-art in terms of performance and explainability and is also suitable for real-world scenarios.

Artificial intelligence (AI) is a crucial technology of industrial digitalization. Especially in the production industry, a great potential is present in optimizing existing processes, e.g., concerning resource consumption, emission reduction, process and product quality improvements, predictive maintenance, and so on. Some of this potential is addressed by methods of industrial analytics beyond specific production technology. Furthermore, particular technological aspects in production systems address another part of this potential, e.g., mechatronics, robotics and motion control, automation systems, and so on. The problem is that the field of AI includes many research areas and methods, and many companies are losing the overview of the necessary and appropriate methods for solving the company problems. The reasons for this are, on the one hand, a lack of expertise in AI and, on the other hand, high complexity and risks of use for the companies (especially for SMEs). As a result, many potentials cannot yet be exploited. The KI-NET project aims to fill this gap, whereby a project overview is presented in this contribution.

In manufacturing industry, product failure is costly, as it results in financial and time losses. Understanding the causes of product failure is critical for reducing the occurrence of failure and optimising the manufacturing process. As a result, a number of studies utilising data-driven approaches such as machine learning have been conducted to reduce the occurrence of this failure and to improve the manufacturing process. While these data-driven approaches enable pattern recognition, they lack the advantages associated with knowledge-driven approaches, such as knowledge representation and deductive reasoning. Similarly, knowledge-driven approaches lack the pattern-learning capabilities inherent in data-driven approaches such as machine learning. Therefore, in this paper, leveraging the advantages of both data-driven and knowledge-driven approaches, we present a strategy with a prototype implementation to reduce manufacturing product failure. The proposed strategy combines a data-driven technique, Bayesian structural learning, with a knowledge-based technique, knowledge graphs.

The adoption of the General Data Protection Regulation (GDPR) has resulted in a significant shift in how the data of European Union citizens is handled. A variety of data sharing challenges in scenarios such as smart cities have arisen, especially when attempting to semantically represent GDPR legal bases, such as consent, contracts and the data types and specific sources related to them. Most of the existing ontologies that model GDPR focus mainly on consent. In order to represent other GDPR bases, such as contracts, multiple ontologies need to be simultaneously reused and combined, which can result in inconsistent and conflicting knowledge representation. To address this challenge, we present the smashHitCore ontology. smashHitCore provides a unified and coherent model for both consent and contracts, as well as the sensor data and data processing associated with them. The ontology was developed in response to real-world sensor data sharing use cases in the insurance and smart city domains. The ontology has been successfully utilised to enable GDPR-complaint data sharing in a connected car for insurance use cases and in a city feedback system as part of a smart city use case.

We present Polite Teacher, a simple yet effective method for the task of semi-supervised instance segmentation. The proposed architecture relies on the Teacher-Student mutual learning framework. To filter out noisy pseudo-labels, we use confidence thresholding for bounding boxes and mask scoring for masks. The approach has been tested with CenterMask, a single-stage anchor-free detector. Tested on the COCO 2017 val dataset, our architecture significantly (approx. +8 pp. in mask AP) outperforms the baseline at different supervision regimes. To the best of our knowledge, this is one of the first works tackling the problem of semi-supervised instance segmentation and the first one devoted to an anchor-free detector.

Identifying and anticipating potential failures in the cloud is an effective method for increasing cloud reliability and proactive failure management. Many studies have been conducted to predict potential failure, but none have combined SMART (self-monitoring, analysis, and reporting technology) hard drive metrics with other system metrics, such as central processing unit (CPU) utilisation. Therefore, we propose a combined system metrics approach for failure prediction based on artificial intelligence to improve reliability. We tested over 100 cloud servers’ data and four artificial intelligence algorithms: random forest, gradient boosting, long short-term memory, and gated recurrent unit, and also performed correlation analysis. Our correlation analysis sheds light on the relationships that exist between system metrics and failure, and the experimental results demonstrate the advantages of combining system metrics, outperforming the state-of-the-art.