Assigning semantic labels to table columns and identifying relations between columns pose significant challenges in data management. Automatic column annotation has been widely treated as classification, with recent works using language models trained on annotated tables with type and property labels. While these language models have effectively modeled individual tables, they often overlook the underlying graph structure of the label space, where constraints can exist between certain types and properties within and across tables. To fill this gap, we propose RODEO, a two-tower architecture that integrates a language model and a graph neural network (GNN) to model the table and semantic labels, respectively. We reformulate column annotation tasks from classification to matching problems, where column and column-pair embeddings are aligned with embeddings that represent their corresponding semantic types (nodes) and properties (edges) within the graph. These embeddings, derived from the language model and GNN, are co-trained end-to-end using triplet loss with an online negative mining strategy. The training process brings semantically related columns and labels closer in the embedding space by minimizing their distances. Our approach, evaluated on publicly available benchmark datasets, outperforms state-of-the-art methods in both column type and column property annotation, highlighting that modeling label constraints through the graph significantly improves overall performance. Ablation studies on the triplet loss and GNN show the robustness of our framework's training procedure.

Ishikawa diagrams, also known as fishbone or cause-and-effect diagrams, are a widely known visual tool for performing root cause analysis (RCA). Although Ishikawa diagrams originated in the manufacturing sector, the tool is also actively used in other areas such as healthcare or business due to its simple structure, which requires little or no training beforehand. Though Ishikawa diagrams are valuable sources of knowledge, they lack rich semantics to effectively process them. As a result, knowledge engineers tend to ignore Ishikawa diagrams and choose other means to collect knowledge, although domain experts are familiar with the RCA tool and it is highly accepted. This paper presents the Ishikawa diagram ontology which enables the explicit modeling of Ishikawa diagrams as visual artifacts, their encoded knowledge and the process of their creation by reusing and extending existing ontologies. The ontology was developed using the LOT methodology. We have created a dataset of Ishikawa diagrams and describe a fictional use case to illustrate the intended use of the presented ontology.