Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces a neurosymbolic approach for validating offer documents in regulated public institutions by combining language model-based information extraction with Logic Tensor Networks (LTNs) for decision-making. The system leverages LLMs to extract predicates from text, which are then fed into an LTN to enforce logical rules and make auditable decisions. Experiments on a real-world corpus demonstrate comparable performance to existing models, with the added benefit of interpretability and explainability through predicate values, rule truth values, and supporting text passages.
Neurosymbolic AI offers a transparent and auditable alternative to black-box models for regulated decision-making, enabling verifiable rule checking directly on real-world documents.
We present a neurosymbolic approach, i.e., combining symbolic and subsymbolic artificial intelligence, to validating offer documents in regulated public institutions. We employ a language model to extract information and then aggregate with an LTN (Logic Tensor Network) to make an auditable decision. In regulated public institutions, decisions must be made in a manner that is both factually correct and legally verifiable. Our neurosymbolic approach allows existing domain-specific knowledge to be linked to the semantic text understanding of language models. The decisions resulting from our pipeline can be justified by predicate values, rule truth values, and corresponding text passages, which enables rule checking based on a real corpus of offer documents. Our experiments on a real corpus show that the proposed pipeline achieves performance comparable to existing models, while its key advantage lies in its interpretability, modular predicate extraction, and explicit support for XAI (Explainable AI).