Open-domain visual entity recognition aims to identify and link entities depicted in images to a vast and evolving set of real-world concepts, such as those found in Wikidata. Unlike conventional classification tasks with fixed label sets, it operates under open-set conditions, where most target entities are unseen during training and exhibit long-tail distributions. This makes the task inherently challenging due to limited supervision, high visual ambiguity, and the need for semantic disambiguation. We propose a Knowledge-guided Contrastive Learning (KnowCoL) framework that combines both images and text descriptions into a shared semantic space grounded by structured information from Wikidata. By abstracting visual and textual inputs to a conceptual level, the model leverages entity descriptions, type hierarchies, and relational context to support zero-shot entity recognition. We evaluate our approach on the OVEN benchmark, a large-scale open-domain visual recognition dataset with Wikidata IDs as the label space. Our experiments show that using visual, textual, and structured knowledge greatly improves accuracy, especially for rare and unseen entities. Our smallest model improves the accuracy on unseen entities by 10.5% compared to the state-of-the-art, despite being 35 times smaller.

Precise optical inspection in industrial applications is crucial for minimizing scrap rates and reducing the associated costs. Besides merely detecting if a product is anomalous or not, it is crucial to know the distinct type of defect, such as a bent, cut, or scratch. The ability to recognize the "exact" defect type enables automated treatments of the anomalies in modern production lines. Current methods are limited to solely detecting whether a product is defective or not without providing any insights on the defect type, nevertheless detecting and identifying multiple defects. We propose MultiADS, a zero-shot learning approach, able to perform Multi-type Anomaly Detection and Segmentation. The architecture of MultiADS comprises CLIP and extra linear layers to align the visual- and textual representation in a joint feature space. To the best of our knowledge, our proposal, is the first approach to perform a multi-type anomaly segmentation task in zero-shot learning. Contrary to the other baselines, our approach i) generates specific anomaly masks for each distinct defect type, ii) learns to distinguish defect types, and iii) simultaneously identifies multiple defect types present in an anomalous product. Additionally, our approach outperforms zero/few-shot learning SoTA methods on image-level and pixel-level anomaly detection and segmentation tasks on five commonly used datasets: MVTec-AD, Visa, MPDD, MAD and Real-IAD.

Knowledge graphs (KGs) have transformed data management within the manufacturing industry, offering effective means for integrating disparate data sources through shared and structured conceptual schemas. However, harnessing the power of KGs can be daunting for non-experts, as it often requires formulating complex SPARQL queries to retrieve specific information. With the advent of Large Language Models (LLMs), there is a growing potential to automatically translate natural language queries into the SPARQL format, thus bridging the gap between user-friendly interfaces and the sophisticated architecture of KGs. The challenge remains in adequately informing LLMs about the relevant context and structure of domain-specific KGs, e.g., in manufacturing, to improve the accuracy of generated queries. In this paper, we evaluate multiple strategies that use LLMs as mediators to facilitate information retrieval from KGs. We focus on the manufacturing domain, particularly on the Bosch Line Information System KG and the I40 Core Information Model. In our evaluation, we compare various approaches for feeding relevant context from the KG to the LLM and analyze their proficiency in transforming real-world questions into SPARQL queries. Our findings show that LLMs can significantly improve their performance on generating correct and complete queries when provided only the adequate context of the KG schema. Such context-aware prompting techniques help LLMs to focus on the relevant parts of the ontology and reduce the risk of hallucination. We anticipate that the proposed techniques help LLMs to democratize access to complex data repositories and empower informed decision-making in manufacturing settings.

Over the years, the demand for interoperability support between diverse applications has significantly in- creased. The Resource Definition Framework (RDF), among other solutions, is utilized as a data modeling language which allows for encoding the knowledge from various domains in a unified representation. More- over, a vast amount of data from heterogeneous data sources are continuously published in documents using the RDF format. Therefore, these RDF documents should be syntactically correct in order to enable software agents performing further processing. Albeit, a number of approaches have been proposed for ensuring error-free RDF documents, commonly they are not able to identify all syntax errors at once by failing on the first encountered error. In this paper, we tackle the problem of simultaneous error identification, and propose RDF-Doctor, a holistic approach for detecting and resolving syntactic errors in a semi-automatic fashion. First, we define a comprehensive list of errors that can be detected along with customized error messages to allow users for a better understanding of the actual errors. Next, a subset of syntactic errors is corrected automatically based on matching them with predefined error messages. Finally, for a particular number of errors, customized and meaningful messages are delivered to users to facilitate the manual corrections process. The results from empirical evaluations provide evidence that the presented approach is able to effectively detect a wide range of syntax errors and automatically correct a large subset of them.

Interoperability among actors, sensors, and heterogeneous systems is a crucial factor for realizing the Industry 4.0 vision, i.e., the creation of Smart Factories by enabling intelligent human-to-machine and machine-to-machine cooperation. In order to empower interoperability in Smart Factories, standards and reference architectures have been proposed. Standards allow for the description of components, systems, and processes, as well as interactions among them. Reference architectures classify, align, and integrate industrial standards according to their purposes and features. Industrial communities in Europe, the United States, and Asia have proposed various reference architectures. However, interoperability among analogous standards in these reference architectures is hampered due to different granularity representation of similar processes or production parts. In this paper, we survey the landscape of Industry 4.0 standards from a semantic perspective. To tackle the problem of interoperability between standards, we developed STO, an ontology for describing standards and their relations. Characteristics of I4.0 standards are described using STO, and these descriptions are exploited for classifying standards from different perspectives according to the reference architectures. Moreover, the semantics encoded in STO allows for the discovery of relations between I4.0 standards, and for mappings across reference architectures proposed by different industrial communities.

Industry 4.0 is a global endeavor of automation and data exchange to create smart factories maximizing production capabilities and allowing for new business models. The Reference Architecture Model for Industry 4.0 (RAMI 4.0) describes the core aspects of Industry 4.0 and defines Administration Shells as digital representations of Industry 4.0 components. In this paper, we present an approach to model and implement Industry 4.0 components with the Resource Description Framework (RDF). The approach addresses the challenges of interoperable communication and machine comprehension in Industry 4.0 settings using semantic technologies. We show how related standards and vocabularies, such as IEC 62264, eCl@ss, and the Ontology of Units of Measure (OM), can be utilized along with the RDF-based representation of the RAMI 4.0 concepts. Finally, we demonstrate the applicability and benefits of the approach using an example from a real-world use case.

S.43-51

Unification and automation of RESTful web services' documentation and descriptions is currently receiving increasing attention. The open-source OpenAPI Specification (formerly known as Swagger) has become core of this effort and has been adopted by a number of major companies. It allows the description of RESTful web services using objects represented in JSON or YAML file formats. As a result, the created descriptions are human and machine-readable, but not machine-understandable. In this paper, we propose a nonintrusive approach for the addition of semantic annotations (similar to RDFa and JSON-LD for HTML) to specific fields of the OpenAPI Specification. We created a lightweight vocabulary for describing RESTful web services using this specification. Furthermore, we practically demonstrate how OpenAPI objects can be enriched with semantic descriptions in a minimally invasive way by adding URIs in the values of chosen OpenAPI properties.

S.334-343