Corvinus and the Italian Society of Neural Networks Launch Collaboration on Neural Mesh paradigm

The collaboration builds upon research conducted by the Corvinus Institute for Advanced Studies (CIAS) and the Institute of Data Analytics and Information Systems (DAIS) at Corvinus, with the support of the Hungarian NKFIH HU-RIZONT program. Péter Baranyi and Ádám Csapó, professors of Corvinus CIAS and DAIS, invented the Neural Mesh architecture as an alternative to conventional neural network models that currently dominate many AI applications. 

The mathematical foundations of the Neural Mesh were inspired by two complementary sources: the organizational principles of the cerebellum and the system and control modeling methodologies that have become central components of tensor product model transformation theories. 

The Neural Mesh introduces the concept of a Logical Synthesis Model (LSM), designed as a complement to Large Language Models (LLMs). While LLMs primarily rely on statistical representations and language-based reasoning to generate and process knowledge, LSMs focus on structured representations and engineering reasoning for the representation, analysis, synthesis, optimization, and control of complex systems. 

Rather than competing with LLMs, the Neural Mesh seeks to complement them. LLMs are highly effective in learning and reasoning over knowledge expressed through language, whereas LSMs aim to learn structured representations of physical and engineered systems, enabling their interpretation, verification, optimization, and control. In this sense, the relationship between LLMs and LSMs may be viewed as analogous to the complementary roles of the cerebrum and cerebellum in biological intelligence. 

Through the newly established partnership, Corvinus and SIREN will collaborate on the scientific evaluation, practical testing, and further advancement of the Neural Mesh framework, contributing to the development of next-generation AI technologies that combine learning capabilities with explainability, verifiability, and engineering design principles.