What question did this study set out to answer?

The aim is to evaluate the effectiveness of three t-norm operators in classifying AI systems for EU compliance.

March 23, 2026Open Access

T-Norm Operators for EU AI Act Compliance Classification: An Empirical Comparison of Łukasiewicz, Product, and Gödel Semantics in a Neuro-Symbolic Reasoning System

Key Points

The aim is to evaluate the effectiveness of three t-norm operators in classifying AI systems for EU compliance.
Benchmark of 1035 annotated AI system descriptions used.
Comparison of three t-norm operators: Łukasiewicz, Product, and Gödel.
Evaluation of classification accuracy across four risk categories.
Utilization of the LGGT+ engine for classification.
T_G achieved the highest accuracy at 84.5%.
All operators showed significant differences in performance (McNemar p<0.001).
T_G introduced 0.8% false positives, impacting reliability.

Abstract

I present a comparative pilot evaluation of three t-norm operators — Łukasiewicz (TL), Product (TP), and Gödel (TG) — as logical conjunction mechanisms in a neuro-symbolic reasoning system for EU AI Act compliance classification. Using the LGGT+ engine and a benchmark of 1035 annotated AI system descriptions, we evaluate classification accuracy across four risk categories. At n=1035, all operators differ significantly (McNemar p<0. 001). TG achieves highest accuracy (84. 5%) but introduces 0. 8% false positives. Code and benchmark dataset: github. com/TriStiX-LS/LggT-core (Apache 2. 0).

T-Norm Operators for EU AI Act Compliance Classification: An Empirical Comparison of Łukasiewicz, Product, and Gödel Semantics in a Neuro-Symbolic Reasoning System

Key Points

Abstract

Cite This Study