Logic Augmented Generation
Aldo Gangemi, Andrea Giovanni Nuzzolese
TL;DR
Logic Augmented Generation (LAG) addresses the gap between symbolic Semantic Knowledge Graphs (SKGs) and opaque Large Language Models (LLMs) by combining a discrete, verifiable knowledge layer with reactive, continuous knowledge synthesis. It treats LLMs as Reactive Continuous Knowledge Graphs (RCKGs) that can surface tacit knowledge on demand; outputs extend from input knowledge $K_{inp}$ to $K_{ext}$ via in-context learning, while SKGs constrain results with logical boundaries. The approach is demonstrated in medical diagnostics and climate projections through a multimodal-to-SKG pipeline and prompt-engineered interactions that harmonize expert viewpoints. This neuro-symbolic framework supports interpretable, context-aware reasoning over evolving knowledge spaces, enabling collective intelligence and decision-support in complex domains.
Abstract
Semantic Knowledge Graphs (SKG) face challenges with scalability, flexibility, contextual understanding, and handling unstructured or ambiguous information. However, they offer formal and structured knowledge enabling highly interpretable and reliable results by means of reasoning and querying. Large Language Models (LLMs) overcome those limitations making them suitable in open-ended tasks and unstructured environments. Nevertheless, LLMs are neither interpretable nor reliable. To solve the dichotomy between LLMs and SKGs we envision Logic Augmented Generation (LAG) that combines the benefits of the two worlds. LAG uses LLMs as Reactive Continuous Knowledge Graphs that can generate potentially infinite relations and tacit knowledge on-demand. SKGs are key for injecting a discrete heuristic dimension with clear logical and factual boundaries. We exemplify LAG in two tasks of collective intelligence, i.e., medical diagnostics and climate projections. Understanding the properties and limitations of LAG, which are still mostly unknown, is of utmost importance for enabling a variety of tasks involving tacit knowledge in order to provide interpretable and effective results.