Building Information Modeling (BIM) has transformed the Architecture, Engineering, Construction, and Operation (AECO) industry by integrating diverse types of information into a unified digital model. While BIM enhances collaboration and decision, making throughout a project’s lifecycle, querying and extracting meaningful insights from BIM models remains a challenge due to their complexity and the technical expertise required to navigate formats like Industry Foundation Classes (IFC). Existing tools provide limited accessibility, and applying state-of-the-art Large Language Models (LLMs) directly to IFC files has proven ineffective due to data volume, lack of semantic structure, and relational complexity. To address these challenges, we introduce ASK-BIM, an approach that combines LLMs, linked data, and knowledge graph (KG) technologies to enable natural language querying of IFC files. By structuring BIM data into a KG before engaging an LLM for reasoning and query resolution, ASK-BIM enhances data accessibility while preserving semantic relationships crucial for complex queries. We evaluate ASK-BIM on a real-world multi-storey building, categorizing questions along two axes and assessing performance in extracting relevant information. Our findings demonstrate the potential of graph-based representations to facilitate AI-driven BIM analysis while also identifying challenges related to the extraction of information from the graph structures. By bridging the gap between BIM data and AI reasoning, ASK-BIM represents a significant step toward intuitive and efficient BIM querying through natural language.

Abstract The growing volume of academic submissions in recent years highlighted the need for scalable and accurate reviewer assignment systems, able to go beyond techniques based on manual processes and basic keyword matching. We propose a novel pipeline that integrates Knowledge Graphs (KGs) and Large Language Models (LLMs) to automate and enhance the reviewer assignment process. Our method extracts meaningful representations of papers and reviewer expertise using Open Information Extraction, the Computer Science Ontology classifier, and GLiNER to build KGs from research content. LLMs are employed to generate targeted keywords through prompt-based synthesis, refining both paper and reviewer profiles. The assignment relies on a hybrid similarity metric combining Cosine and Jaccard similarities to capture both lexical and semantic alignment. We evaluate the pipeline using standard metrics such as Mean Reciprocal Rank, Mean Average Precision, and Precision at K, on a dataset in the Computer Science domain, demonstrating its effectiveness in aligning submissions with appropriate reviewers. This approach offers a scalable and adaptive solution to the complexities of modern peer review.

Abstract Converting natural language text into structured, logically coherent knowledge graphs (KGs) enhances the ability to retrieve, organize, and analyze vast amounts of information at scale. This paper introduces Text2AMR2FRED, a text-to-KG pipeline that converts multilingual natural language text into logically coherent, interoperable KGs. Designed to support large-scale information retrieval and knowledge extraction, this pipeline addresses key limitations of existing semantic parsers and machine readers, including issues with logical consistency and interoperability. By adhering to Semantic Web standards, Text2AMR2FRED systematically structures text-based information and enhances it through integration with external knowledge sources, delivering enriched, semantically sound KGs ready for diverse applications. We obtain the output KGs by leveraging Abstract Meaning Representation (AMR) as an intermediate semantic parsing formalism, exploiting the progress achieved by text-to-AMR parsers employing pre-trained language models. We produce a manually validated KG s bank created by transforming a dataset of natural language sentences into KGs using Text2AMR2FRED and applying an intrinsic evaluation method that leverages Open Knowledge Extraction motifs.

The use of Knowledge Graphs (KGs) which constitute large networks of real-world entities and their interrelationships, has grown rapidly. A substantial body of research has emerged, exploring the integration of deep learning (DL) and large language models (LLMs) with KGs. This workshop aims to bring together leading researchers in the field to discuss and foster collaborations on the intersection of KG and DL/LLMs.

In scientific papers, it is common practice to cite other articles to substantiate claims, provide evidence for factual assertions, reference limitations, and research gaps, and fulfill various other purposes. When authors include a citation in a given sentence, there are two considerations they need to take into account: (i) where in the sentence to place the citation and (ii) which citation to choose to support the underlying claim. In this paper, we focus on the first task as it allows multiple potential approaches that rely on the researcher’s individual style and the specific norms and conventions of the relevant scientific community. We propose two automatic methodologies that leverage transformers architecture for either solving a Mask-Filling problem or a Named Entity Recognition problem. On top of the results of the proposed methodologies, we apply ad-hoc Natural Language Processing heuristics to further improve their outcome. We also introduce s2orc-9K, an open dataset for fine-tuning models on this task. A formal evaluation demonstrates that the generative approach significantly outperforms five alternative methods when fine-tuned on the novel dataset. Furthermore, this model’s results show no statistically significant deviation from the outputs of three senior researchers.

Abstract A primary concern in the realm of mechanical engineering is to ensure the efficient and effective data entry of hardware devices. Fasteners are mechanical tools that rigidly connect or affix two surfaces or objects together. They are small and often different fasteners might look similar; it is therefore a long and prone-to-risk procedure to manually analyze them to classify and store their related information. With the widespread diffusion of AI frameworks in several domains, equipment manufacturers started to rely on AI technologies for these heavy tasks. Automatically classifying fasteners by type and extracting metadata from natural language questions are important tasks that fastener manufacturers and suppliers encounter. In this paper, we address these challenges. To address the first task, we introduce an augmentation methodology that starts with a small set of 3D models representing each of the 21 types of fasteners we aim to classify. This methodology efficiently generates multiple 2D images from these models. Next, we train a vision transformer using the collected data to address a single-label multi-class classification task. For the second task, we introduce a prompt-engineering technique designed for conversational agents. This technique leverages in-context knowledge to extract (metadata field, value) pairs from natural language questions. Subsequently, we tackle a question-answering task to the description fields of the extracted fasteners. Our evaluation demonstrates the effectiveness of both approaches, surpassing the baselines we tested.

The design and management of modern big data platforms are extremely complex. It requires carefully integrating multiple storage and computational platforms as well as implementing approaches to protect and audit data access. Therefore, onboarding new data and implementing new data transformation processes is typically time-consuming and expensive. In many cases, enterprises construct their data platforms without a clear distinction between logical and technical concerns. Consequently, these platforms lack sufficient abstraction and are closely tied to particular technologies, making the adaptation to technological evolution very costly. This paper illustrates a novel approach to designing data platform models based on a formal ontology that structures various domain components into an accessible knowledge graph. We also describe the preliminary version of AGILE-DM, a novel ontology that we built for this purpose. Our solution is flexible, technologically agnostic, and more adaptable to changes and technical advancements.

Scientific knowledge has been traditionally disseminated and preserved through research articles published in journals, conference proceedings, and online archives. However, this article-centric paradigm has been often criticized for not allowing to automatically process, categorize, and reason on this knowledge. An alternative vision is to generate a semantically rich and interlinked description of the content of research publications. In this paper, we present the Artificial Intelligence Knowledge Graph (AI-KG), a large-scale automatically generated knowledge graph that describes 820K research entities. AI-KG includes about 14M RDF triples and 1.2M reified statements extracted from 333K research publications in the field of AI, and describes 5 types of entities (tasks, methods, metrics, materials, others) linked by 27 relations. AI-KG has been designed to support a variety of intelligent services for analyzing and making sense of research dynamics, supporting researchers in their daily job, and helping to inform decision-making in funding bodies and research policymakers. AI-KG has been generated by applying an automatic pipeline that extracts entities and relationships using three tools: DyGIE++, Stanford CoreNLP, and the CSO Classifier. It then integrates and filters the resulting triples using a combination of deep learning and semantic technologies in order to produce a high-quality knowledge graph. This pipeline was evaluated on a manually crafted gold standard, yielding competitive results. AI-KG is available under CC BY 4.0 and can be downloaded as a dump or queried via a SPARQL endpoint.

This paper describes a novel method for generating and integrating knowledge graphs extracted from multiple natural language sources by FRED, a machine reading tool for generating abstract representations of text documents. This is a key problem in human-robot spoken dialogue interaction, issue which arises from a current research project related to active and healthy ageing using caring service robots where we are involved. The problem is also relevant in many application scenarios requiring the creation and dynamic evolution of a knowledge base, such as automatic news summarisation. Solving this problem requires solving sub-tasks that have only been studied individually, so far. We propose a holistic approach to handle FRED's graphs related to different input texts and output a knowledge graph representing the reconciled knowledge.

We address the problem of selecting a banner advertisement, based on the profile of the online user. The profile consists of the set of webpages opened by the online user, optionally clustered. In order to select the banner, we train a classifier with a dataset containing rules of the form P(u) → B(u), where P(u) is the profile of user u, and B(u) is the set of banners clicked by user u. We present two possible transformations that we use in order to train the classifier. In the first transformation, TMax, we put only one line P(u) →b, where b is the most frequently clicked banner by the user u. In the second transformation, TMultiline, we put one line P(u) → b, for each banner b in B(u). Finally, we perform several experiments, which show that there is a strong correlation between the profiles of the user, and the banners clicked by the user.

Range and k-nearest neighbor searching are core problems in pattern recognition. Given a database S of objects in a metric space M and a query object q in M, in a range searching problem the goal is to find the objects of S within some threshold distance to g, whereas in a k-nearest neighbor searching problem, the k elements of S closest to q must be produced. These problems can obviously be solved with a linear number of distance calculations, by comparing the query object against every object in the database. However, the goal is to solve such problems much faster. We combine and extend ideas from the M-tree, the multivantage point structure, and the FQ-tree to create a new structure in the "bisector tree" class, called the Antipole tree. Bisection is based on the proximity to an "Antipole" pair of elements generated by a suitable linear randomized tournament. The final winners a, b of such a tournament is far enough apart to approximate the diameter of the splitting set. If dist(a, b) is larger than the chosen cluster diameter threshold, then the cluster is split. The proposed data structure is an indexing scheme suitable for (exact and approximate) best match searching on generic metric spaces. The Antipole tree outperforms by a factor of approximately two existing structures such as list of clusters, M-trees, and others and, in many cases, it achieves better clustering properties.