From March 1, the Spanish National Research Council (CSIC) will lead an international project called “Neuropic” to manufacture neural networks that will be used to develop sustainable artificial intelligence. The main objective of the project is to create a network of artificial neurons using the thermal energy of the devices themselves, which is innovative. These neural networks will be used in integrated photonic circuits with lower energy consumption, which is a breakthrough in artificial intelligence. Nowadays, AI has become ubiquitous in our daily lives, but its ability to solve algorithms consumes a large amount of energy, so it is crucial to research and develop projects that optimize its energy use.
The “Neuropic” project aims to address this energy problem and make more efficient use of information. The project coordinator, Pedro David Garca, a CSIC researcher at the Institute of Materials Science of Madrid (ICMM), points out that “we propose to use the thermal energy of the system to create this type of artificial neurons.” The project will be carried out using silicon, a material that is abundant in the environment and is closely linked to technological developments, which will reduce its environmental impact.
The “Neuropic” project has been funded with 3 million euros through Pathfinder grants from the European Union and involves the participation of two research institutes and a small and medium-sized Spanish company that will focus on the design and characterization of the system. The researcher Cefe Lopez, from PC Group, will also participate in the project from ICMM. The duration of the project will be four months.