Large language models (LLMs) have demonstrated impressive capabilities in natural language understanding and generation, but they exhibit problems with logical consistency in the output they generate. How can we harness LLMs’ broad-coverage parametric knowledge in formal reasoning despite their inconsistency? We present a method for directly integrating an LLM into the interpretation function of the formal semantics for a paraconsistent logic. We provide experimental evidence for the feasibility of the method by evaluating the function using datasets created from several short-form factuality benchmarks. Unlike prior work, our method offers a theoretical framework for neurosymbolic reasoning that leverages an LLM’s knowledge while preserving the underlying logic’s soundness and completeness properties.

A number of systems have been developed that demonstrate the utility of knowledge-based techniques in the construction of software reuse systems. The amount of manual effort expended in the development of such systems can be reduced through the use of a knowledge acquisition environment tailored for a reuse-oriented model of the software development process. We present a model of the software development process and then describe the Bauhaus, a knowledge-based user interface to Ada software libraries.