Neuro-Symbolic Hybrid Systems

Neuro-Symbolic Hybrid Systems

Description of the activity

The research group has been dealing for years with the integration of symbolic methodologies (classical artificial intelligence) with sub-symbolic methodologies (artificial neural networks). In particular, the neural network models adopted, studied and modified over time (new models introduced) belong to the family of weightless neural networks. In addition to the theoretical contributions related to this particular family of neural networks, the research group has deposited various open source software and designed and developed several applications. They range from active video surveillance to real-time processing of images and videos; from cognitive and behavioural robotics to spatial reasoning in two and three dimensions; from classification to clustering; from abductive inferential systems to natural language processing.

Involved personnel

M. de Gregorio | A. Calabrese | O. Talamo | A. Sorgente | G. Coda | P. Vanacore | F. Mele

National and International Collaborations

National Research Council – CNR (Institute for high performance computing and networking – ICAR)
University of Napoli “Federico II” (Dept of Physics, DIETI)
UFRJ – Universidade Federal do Rio de Janeiro (COPPE – Instituto Alberto Luiz Coimbra de PĂłs-Graduação e Pesquisa de Engenharia, NCE – Instituto TĂ©rcio Pacitti de Aplicações e Pesquisas Computacionais), UFMG-Universidade Federal de Minas Gerais (Computational Intelligence Laboratory), UFRPE- Universidade Federal Rural de Pernambuco (DEINFO – Departamento de EstatĂ­stica e informática), Imperial College London (DEEE – Department of Electrical and Electronic Engineering and Dept. Computing)
Neatec Srl