Prompt2DeModel: Declarative Neuro-Symbolic Modeling with Natural Language

TL;DR

Prompt2DeModel introduces a natural language-based pipeline for creating declarative neuro-symbolic models, enabling domain experts to formalize knowledge without deep ML expertise, using large language models and interactive refinement.

Contribution

It presents a novel conversational pipeline that leverages large language models to generate declarative programs for neuro-symbolic modeling from natural language prompts.

Findings

Effective generation of domain knowledge representations

Enhanced accessibility for non-ML experts

Integration of symbolic constraints with neural models

Abstract

This paper presents a conversational pipeline for crafting domain knowledge for complex neuro-symbolic models through natural language prompts. It leverages large language models to generate declarative programs in the DomiKnowS framework. The programs in this framework express concepts and their relationships as a graph in addition to logical constraints between them. The graph, later, can be connected to trainable neural models according to those specifications. Our proposed pipeline utilizes techniques like dynamic in-context demonstration retrieval, model refinement based on feedback from a symbolic parser, visualization, and user interaction to generate the tasks' structure and formal knowledge representation. This approach empowers domain experts, even those not well-versed in ML/AI, to formally declare their knowledge to be incorporated in customized neural models in the DomiKnowS framework.