Valentina Presutti

Background: Neuro-symbolic methods enhance the reliability of neural network classifiers through logical constraints, but they lack native support for ontologies. Objectives: We aim to develop a neuro-symbolic method that reliably outputs predictions consistent with a Description Logic ontology that formalizes domain-specific knowledge. Methods: We encode a Description Logic ontology as a circuit, a feed-forward differentiable computational graph that supports tractable execution of queries and transformations. We show that the circuit can be used to (i) generate synthetic datasets that capture the semantics of the ontology; (ii) efficiently perform deductive reasoning on a GPU; (iii) implement neuro-symbolic models whose predictions are approximately or provably consistent with the knowledge defined in the ontology. Results We show that the synthetic dataset generated using the circuit qualitatively captures the semantics of the ontology while being challenging for Machine Learning classifiers, including neural networks. Moreover, we show that compiling the ontology into a circuit is a promising approach for scalable deductive reasoning, with runtimes up to three orders of magnitude faster than available reasoners. Finally, we show that our neuro-symbolic classifiers reliably produce consistent predictions when compared to neural network baselines, maintaining competitive performances or even outperforming them. Conclusions By compiling Description Logic ontologies into circuits, we obtain a tighter integration between the Deep Learning and Knowledge Representation fields. We show that a single circuit representation can be used to tackle different challenging tasks closely related to real-world applications.

Extracting event knowledge from unstructured text is a well-known challenge in Natural Language Processing (NLP) and is particularly difficult when dealing with fiction. Subtext, rather than explicit information, and figurative style in fictional narratives, complicate event extraction. Recent advances in Large Language Models (LLMs) have improved performance across various NLP tasks. However, their effectiveness in extracting events from fiction remains underexplored. In this article, we evaluate the performance of open-weights LLMs to extract character death events from fictional narratives in English. These events are defined as triples consisting of Victim, Perpetrator, and Mode of Demise. We cast Knowledge Extraction (KE) as a zero-shot task and evaluate our approach on a manually annotated benchmark of fanfiction stories. Our results show that LLMs struggle with KE from fiction, with a maximum F1-score of 0.45 across the elements constituting the triples and, at most, 25% of death events correctly extracted. A detailed error analysis reveals that most errors stem from missed death events and from direct presentation modes, such as direct speech, which significantly impair extraction performance. Moreover, KE accuracy declines as the story length increases, while LLMs’ background knowledge leakage contributes to false positives. These findings provide domain-specific insights into the challenges of KE in fiction.

Linking named entities occurring in text to their corresponding entity in a Knowledge Base (KB) is challenging, especially when dealing with historical texts. In this work, we introduce Musical Heritage named Entities Recognition, Classification and Linking (mhercl), a novel benchmark consisting of manually annotated sentences extrapolated from historical periodicals of the music domain. mhercl contains named entities under-represented or absent in the most famous KBs. We experiment with several State-of-the-Art models on the Entity Linking (EL) task and show that mhercl is a challenging dataset for all of them. We propose a novel unsupervised EL model and a method to extend supervised entity linkers by using Knowledge Graphs (KGs) to tackle the main difficulties posed by historical documents. Our experiments reveal that relying on unsupervised techniques and improving models with logical constraints based on KGs and heuristics to predict entities (entities not represented in the KB of reference) results in better EL performance on historical documents.

Linking named entities occurring in text to their corresponding entity in a Knowledge Base (KB) is challenging, especially when dealing with historical texts. In this work, we introduce Musical Heritage named Entities Recognition, Classification and Linking (MHERCL), a novel benchmark consisting of manually annotated sentences extrapolated from historical periodicals of the music domain. MHERCL contains named entities under-represented or absent in the most famous KBs. We experiment with several State-of-the-Art models on the Entity Linking (EL) task and show that MHERCL is a challenging dataset for all of them. We propose a novel unsupervised EL model and a method to extend supervised entity linkers by using Knowledge Graphs (KGs) to tackle the main difficulties posed by historical documents. Our experiments reveal that relying on unsupervised techniques and improving models with logical constraints based on KGs and heuristics to predict NIL entities (entities not represented in the KB of reference) results in better EL performance on historical documents.

This work explores the integration of ontology-based reasoning and Machine Learning techniques for explainable value classification. By relying on an ontological formalization of moral values as in the Moral Foundations Theory, relying on the DnS Ontology Design Pattern, the \textit{sandra} neuro-symbolic reasoner is used to infer values (fomalized as descriptions) that are \emph{satisfied by} a certain sentence. Sentences, alongside their structured representation, are automatically generated using an open-source Large Language Model. The inferred descriptions are used to automatically detect the value associated with a sentence. We show that only relying on the reasoner's inference results in explainable classification comparable to other more complex approaches. We show that combining the reasoner's inferences with distributional semantics methods largely outperforms all the baselines, including complex models based on neural network architectures. Finally, we build a visualization tool to explore the potential of theory-based values classification, which is publicly available at http://xmv.geomeaning.com/.