Ontologies are essential for structuring domain knowledge, improving accessibility, sharing, and reuse. However, traditional ontology construction relies on manual annotation and conventional natural language processing (NLP) techniques, making the process labour-intensive and costly, especially in specialised fields like casting manufacturing. The rise of Large Language Models (LLMs) offers new possibilities for automating knowledge extraction. This study investigates three LLM-based approaches, including pre-trained LLM-driven method, in-context learning (ICL) method and fine-tuning method to extract terms and relations from domain-specific texts using limited data. We compare their performances and use the best-performing method to build a casting ontology that validated by domian expert.

Automatic Compliance Checking (ACC) within the Architecture, Engineering, and Construction (AEC) sector necessitates automating the interpretation of building regulations to achieve its full potential. Converting textual rules into machine-readable formats is challenging due to the complexities of natural language and the scarcity of resources for advanced Machine Learning (ML). Addressing these challenges, we introduce CODE-ACCORD, a dataset of 862 sentences from the building regulations of England and Finland. Only the self-contained sentences, which express complete rules without needing additional context, were considered as they are essential for ACC. Each sentence was manually annotated with entities and relations by a team of 12 annotators to facilitate machine-readable rule generation, followed by careful curation to ensure accuracy. The final dataset comprises 4,297 entities and 4,329 relations across various categories, serving as a robust ground truth. CODE-ACCORD supports a range of ML and Natural Language Processing (NLP) tasks, including text classification, entity recognition, and relation extraction. It enables applying recent trends, such as deep neural networks and large language models, to ACC.

Traditional manufacturing sectors, particularly aluminium casting, face growing pressure to improve knowledge transfer to new generations and customers. The complexity and volume of production information often make conventional methods inefficient. While Large Language Models (LLMs) and Knowledge Graphs (KGs) show promise in knowledge management, applying them to domain-specific contexts like aluminium casting processes presents persistent challenges. This paper draws on practical implementation experience to highlight key obstacles encountered in deploying an LLM-KG system for knowledge transfer in this setting.

Understanding vulnerable road user (VRU) behavior is critical for designing safer and more inclusive urban infrastructure. This paper presents a structured and explainable framework using Spatio-Temporal Knowledge Graphs (STKGs) to model and analyze VRU crossing behaviors. By constructing knowledge graphs from real-world urban traffic datasets, we capture dynamic interactions between pedestrians, cyclists, and vehicles in different crossing scenarios. Through query-based analysis, our approach extracts insights that provide decision support to traffic engineers, addressing key safety-related concerns such as unsafe crossing patterns, pedestrian-vehicle interactions, and speed-related risks. This work contributes to responsible AI by enabling transparent, explainable, and data efficient decision-making support for real-world traffic planning and infrastructure design. 

The significant impact of the built environment on resource consumption and waste production has led to calls for a shift towards a circular economy model that maximizes the efficient use of resources. This study explores the use of digital product passports (DPPs) to improve how we manage products throughout their lifecycle. However, dealing with the complexity and large volume of data in DPPs can be challenging in terms of effective information management and utilization. We address this issue by adopting a modular ontological approach to systematically capture product lifecycle information from its origin to its end-of-life phase. To ensure interoperability and reusability of the ontology, we annotate key concepts and relationships using International Organization for Standardization (ISO) standards that promote circular economy. Our research led to the development of several ontology modules derived from literature reviews and interviews conducted with industry and academia experts who specialize in sustainability. These modules were then integrated to create a digital product passport ontology. The study demonstrates the feasibility of using a modular ontology approach to manage the complex information inherent in DPPs paving the way for more sustainable management practices in the built environment sector. 

In this paper, we focus on the crucial task of understanding and modeling pedestrian behavior, which is essential for numerous applications. We introduce a semantic representation of pedestrian crossing behavior. The representation is to capture the sub-events within a behavior and the spatial-temporal evolution of interactions between pedestrians and other objects involved in crossing events. We demonstrate its practical application by utilizing it to analyze pedestrian crossing behavior from road user movement data (i.e. trajectories). By constructing a knowledge graph from detailed road user dynamics data using this representation, we enable queries that address safety concerns related to pedestrian crossing behavior, aiding traffic engineers in their work on urban traffic infrastructure design.

The Scheil–Gulliver equation is essential for assessing solid fractions during alloy solidification in materials science. Despite the prevalent use of the Calculation of Phase Diagrams (CALPHAD) method, its computational intensity and time are limiting the simulation efficiency. Recently, Artificial Intelligence has emerged as a potent tool in materials science, offering robust and reliable predictive modeling capabilities. This study introduces an ensemble-based method that has the potential to enhance the prediction of the partitioning coefficient (k) in the Scheil equation by inputting various alloy compositions. The findings demonstrate that this approach can predict the temperature and solid fraction at the eutectic temperature with an accuracy exceeding 90%, while the accuracy for k prediction surpasses 70%. Additionally, a case study on a commercial alloy revealed that the model's predictions are within a 5°C deviation from experimental results, and the predicted solid fraction at the eutectic temperature is within a 15% difference of the values obtained from the CALPHAD model.

Both standards and ontologies are among the important components to realize the vision of BIM (Building Information Modeling). They provide a community consensus for interpretation, communication and interoperability of building data. This consensus is pivotal in enabling diverse stakeholders and systems to seamlessly collaborate and share data across the entire building life cycle. In this paper we describe the development of ontologies for building product properties, aligning with standards, and demonstrate their usage in achieving semantic interoperability. First, a top-domain ontology, BPPO (Buiding Product Property Ontology), is developed for building product properties. This top-domain ontology is used to guide the development of domain ontologies for properties in different categories of products or groups of product categories. Subsequently, a domain ontology, LPPO (Lighting Product Property Ontology), is built for lighting product properties, with guidance from BPPO, in this work. The ontological terminologies of both BPPO and LPPO are aligned with the standards set forth by the BIM community. Furthermore, the ontologies have been used in an application to support and enhance the interoperability between the manufacturer’s product database and the BIM platform.

To satisfy the rising demand for higher product quality and giga-casting requirements, the casting process is undergoing significant changes. However, current control methods rely significantly on human expertise and experience, making process availability and stability difficult to ensure. The semisolid casting process is more complicated than conventional liquid casting due to the additional casting parameters incorporated during the slurry preparation, which can have an effect on the quality of the final product. Therefore, an efficient tool is required to simplify the complete process of semisolid casting. The introduction of an AI system to aid in the supervision of the casting manufacturing procedure is one potential solution. This paper introduces a new casting system named”Smart-Cast” developed for this specific purpose. The paper describes the functions of the system and its current development process. Using an AI system as an assistant can help to achieve the goal of enhancing the efficacy of casting process control, and it can also help foundries step into the Industry 4.0 era.

The built environment is heavily dependent on wasteful linear economic models and needs to transition to the circular economy (CE). One of the key enablers of CE is Digital Product Passports (DPPs). However, determining the necessary information and selecting suitable technologies remains to be challenging in practical implementations. This research aims to present a framework for implementing DPPs using Knowledge Graphs (KGs). A literature review was conducted to identify the key components of the framework. The result shows that the key elements encompass use cases identification, data collection, modelling, integration, governance, access and querying, and maintenance and updating.