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.

With the recent surge in popularity of Large Language Models (LLMs), there is the rising risk of users blindly trusting the information in the response. Nevertheless, there are cases where the LLM recommends actions that have potential legal implications and this may put the user in danger. We provide an empirical analysis on multiple existing LLMs showing the urgency of the problem. Hence, we propose a first short-term solution, consisting in an approach for isolating these legal issues through prompt engineering. We prove that this solution is able to stem some risks related to legal implications, nonetheless we also highlight some limitations. Hence, we argue on the need for additional knowledge-intensive resources and specifically Knowledge Graphs for fully solving these limitations. For the purpose, we draw our proposal aiming at designing and developing a solution powered by a legal Knowledge Graph (KG) that, besides capturing and alerting the user on possible legal implications coming from the LLM answers, is also able to provide actual evidence for them by supplying citations of the interested laws. We conclude with a brief discussion on the issues that may be needed to solve for building a comprehensive legal Knowledge Graph