The rise of AI has profoundly pervaded our society; however, its development poses significant challenges in terms of energy sustainability.
Current digital computing technologies require large amounts of energy and have a considerable environmental impact. In the race toward future computing technologies, bio-inspired systems based on self-organizing memristive networks represents promising unconventional computing hardware platforms for emulating information processing capabilities of our brain. However, the relationship between dynamics of these physical complex systems and their information processing capabilities remained a challenge.
A new study, published in the prestigious journal Nature Communications marks a decisive step in this direction.
Researchers from the INRiM, Politecnico di Torino and Universitat Autònoma de Barcelona have shown that self-assembled memristive networks can be described as stochastic dynamical systems, presenting for the first time a framework capable of linking the dynamics of these complex systems with their computational properties.
“These results, obtained in the framework of my ERC starting grant, shed new light on how the physics of nanoscale systems can be leveraged for the realization of computing architectures with high energy efficiency”, says Gianluca Milano, researcher of the group Advanced Materials & Devices at INRiM and responsible of the ERC MEMBRAIN project and the MEMQUD project that financed the research.
These results pave the way for the development of new hardware architectures exploiting deterministic and stochastic dynamics in the same physical substrate in a similar way to what our brain does.