Generative AI has shown remarkable advancements across various tasks, including recommender systems, where recent research leverages generative approaches to provide personalised recommendations based on user-item historical interaction data. However, the inherent sparsity of the interaction data poses a significant challenge to the advancement of generative recommender models. While some discriminative models have explored incorporating knowledge graphs (KGs) to address this issue, they often struggle with noise sensitivity, lack of explainability, and difficulties in handling cold-start scenarios, where new items with little or no historical user interaction data are involved. In this paper, we propose a novel dual-architecture generative model that intuitively integrates a diffusion model with KG-based reasoning, which reflects the propagation of user preference in a KG towards items. Our approach not only improves recommendation accuracy significantly, but also introduces explainability by leveraging the structured insights from KGs. Furthermore, the KG-based reasoning enables our model to effectively address cold-start scenarios. By utilising the semantic connections in the KG, our model can recommend these new items with confidence, overcoming a common limitation of traditional methods. We evaluate our model on three benchmark datasets, demonstrating superior performance (beat SOTA by over 10% in recall@20 in average).

The rapid growth of open and structured RDF data on the Web has promoted the development of dataset search as an important research topic. The core function of existing systems is ad hoc dataset retrieval (AHDR) based on the metadata of datasets, which contains limited information and often suffers from quality issues. To overcome the limitations, in this article, we systematically investigate content-based AHDR to exploit the actual RDF data in datasets. We address three main tasks of content-based AHDR with novel methods for handling the large size and complex structure of RDF data to facilitate dataset retrieval, deduplication, and snippet extraction. These methods are integrated into an online and open-source prototype called Caddie . The effectiveness and practicability of its components are evaluated on a public test collection and by a user study.

Dataset search, or more specifically, ad hoc dataset retrieval which is a trending specialized IR task, has received increasing attention in both academia and industry. While methods and systems continue evolving, existing test collections for this task exhibit shortcomings, particularly suffering from lexical bias in pooling and limited to keyword-style queries for evaluation. To address these limitations, in this paper, we construct ACORDAR 2.0, a new test collection for this task which is also the largest to date. To reduce lexical bias in pooling, we adapt dense retrieval models to large structured data, using them to find an extended set of semantically relevant datasets to be annotated. To diversify query forms, we employ a large language model to rewrite keyword queries into high-quality question-style queries. We use the test collection to evaluate popular sparse and dense retrieval models to establish a baseline for future studies. The test collection and source code are publicly available.

Abstract Knowledge Graph Embedding (KGE) has attracted increasing attention. Relation patterns, such as symmetry and inversion, have received considerable focus. Among them, composition patterns are particularly important, as they involve nearly all relations in KGs. However, prior KGE approaches often consider relations to be compositional only if they are well-represented in the training data. Consequently, it can lead to performance degradation, especially for under-represented composition patterns. To this end, we propose HolmE, a general form of KGE with its relation embedding space closed under composition, namely that the composition of any two given relation embeddings remains within the embedding space. This property ensures that every relation embedding can compose, or be composed by other relation embeddings. It enhances HolmE’s capability to model under-represented (also called long-tail) composition patterns with limited learning instances. To our best knowledge, our work is pioneering in discussing KGE with this property of being closed under composition. We provide detailed theoretical proof and extensive experiments to demonstrate the notable advantages of HolmE in modelling composition patterns, particularly for long-tail patterns. Our results also highlight HolmE’s effectiveness in extrapolating to unseen relations through composition and its state-of-the-art performance on benchmark datasets.

Data analytics including machine learning (ML) is essential to extract insights from production data in modern industries. However, industrial ML is affected by: the low transparency of ML towards non-ML experts; poor and non-unified descriptions of ML practices for reviewing or comprehension; ad-hoc fashion of ML solutions tailored to specific applications, which affects their re-usability. To address these challenges, we propose the concept and a system of executable knowledge graph (KG), which represent KGs that rely on semantic technologies to formally encode ML knowledge and solutions. These KGs can be translated to executable scripts in a reusable and modularised fashion. The demo attendees will use our system to modify, integrate and create executable KGs via a graphic user interface, which offer a user-friendly way to understand, configure, reuse, and create data analytics pipelines.

With the development of Industry 4.0 technology, modern industries such as Bosch's welding monitoring witnessed the rapid widespread of machine learning (ML) based data analytical applications, which in the case of welding monitoring has led to more efficient and accurate welding monitoring quality. However, industrial ML is affected by the low transparency of ML towards non-ML experts needs. The lack of understanding by domain experts of ML methods hampers the application of ML methods in industry and the reuse of developed ML pipelines, as ML methods are often developed in an ad hoc manner for specific problems. To address these challenges, we propose the concept and a system of executable Knowledge Graph (KG), which formally encode ML knowledge and solutions in KGs, which serve as common language between ML experts and non-ML experts, thus facilitate their communication and increase the transparency of ML methods. We evaluated our system extensively with an industrial use case at Bosch, showing promising results.

Automatic knowledge graph (KG) construction is widely used for e.g. data integration, question answering and semantic search. There are many approaches of automatic KG construction. Among which, an important approach is to map the raw data to a given domain KG schema, e.g., domain ontology or conceptual graph, and construct the entities and properties according to the domain KG schema. However, the existing approaches to construct KGs are not always efficient enough and the resulting KGs are not sufficiently application and user-friendly. The main challenge arises from the trade-off: the domain KG schema should be domain-generic and knowledge-oriented, to reflect the general domain knowledge rather than data particularities; while a KG schema should be data-oriented, to cover all data features. If the former is directly used for KG construction, this can cause issues like a high load of blank nodes, which are technical nodes in the KGs that represent unknown entities. To this end, we propose our ScheRe system in the demo, which relies on a schema reshaping algorithm and other two semantic modules for enhancing KG construction. The demo attendees will use ScheRe to reshape a domain KG schema to data specific KG schema, build KGs with industrial data, and experience more user-friendly querying.

Monitoring of the state, performance, quality of operations and other parameters of equipment and production processes, which is typically referred to as condition monitoring, is an important common practice in many industries including manufacturing, oil and gas, chemical and process industry. In the age of Industry 4.0, where the aim is a deep degree of production automation, unprecedented amounts of data are generated by equipment and processes, and this enables adoption of Machine Learning (ML) approaches for condition monitoring. Development of such ML models is challenging. On the one hand, it requires collaborative work of experts from different areas, including data scientists, engineers, process experts, and managers with asymmetric backgrounds. On the other hand, there is high variety and diversity of data relevant for condition monitoring. Both factors hampers ML modelling for condition monitoring. In this work, we address these challenges by empowering ML-based condition monitoring with semantic technologies. To this end we propose a software system SemML that allows to reuse and generalise ML pipelines for conditions monitoring by relying on semantics. In particular, SemML has several novel components and relies on ontologies and ontology templates for ML task negotiation and for data and ML feature annotation. SemML also allows to instantiate parametrised ML pipelines by semantic annotation of industrial data. With SemML, users do not need to dive into data and ML scripts when new datasets of a studied application scenario arrive. They only need to annotate data and then ML models will be constructed through the combination of semantic reasoning and ML modules. We demonstrate the benefits of SemML on a Bosch use-case of electric resistance welding with very promising results.

Manufacturing of car bodies heavily relies on demanding welding processes of joining body parts together that introduce thousands of joining welding spots in each car. Quality monitoring for these spots impacts production efficiency and cost. In this paper we develop an ML pipeline to predict the spot quality before the actual welding happens. This pipeline is based on a Feature Engineering~(FE) approach to manually design features using domain knowledge. We evaluated the pipeline with two datasets from industrial plants, achieving very promising results with prediction errors around 2%. Then, we develop an approach to semantically enhance FE pipelines in order to automate the ML process without compromising the prediction accuracy and to facilitate generalisation and transfer of FE-based models to other datasets and processes. Our ML pipeline has been deployed offline on various Bosch manufacturing datasets in a controlled environment since early 2019 and evaluated.

Machine learning (ML) based data analysis has attracted an increasing attention in the manufacturing industry, however, many challenges hamper their wide spread adoption. The main challenges are the high costs of labour-intensive data preparation from diverse sources and processes, the asymmetrical backgrounds of the experts involved in manufacturing analyses that impede efficient communication between them, and the lack of generalisability of ML models tailored to specific applications. Our semantically enhanced ML pipeline, SemFE, with feature engineering addresses these challenges, serving as a bridge to bring the endeavours of experts together, and making data science accessible to non-ML-experts. SemFE relies on ontologies for discrete manufacturing monitoring that encapsulate domain and ML knowledge; it has five novel semantic modules for automation of ML-pipeline development and user-friendly GUIs. The demo attendees will be able to use our system to build manufacturing monitoring ML pipelines, and to design their own pipelines with minimal prior knowledge of machine learning.