Automatic coherence-driven inference on arguments
This work addresses inconsistencies in legal and policy domains, offering a technological remedy for legislative analysis and legal reasoning.
The paper tackles the problem of inconsistencies in law and administration by proposing a neurosymbolic architecture that uses large language models to extract propositions and perform coherence-driven inference via combinatorial optimization, enabling meaningful judgments about argument coherence.
Inconsistencies are ubiquitous in law, administration, and jurisprudence. Though a cure is too much to hope for, we propose a technological remedy. Large language models (LLMs) can accurately extract propositions from arguments and compile them into natural data structures that enable coherence-driven inference (CDI) via combinatorial optimization. This neurosymbolic architecture naturally separates concerns and enables meaningful judgments about the coherence of arguments that can inform legislative and policy analysis and legal reasoning.