The advent of advanced hardware and software solutions, which can partially emulate the behavior of the brain in simple data processing and in more complex classification and pattern recognition tasks, is a turning point in the development of neuromorphic artificial intelligent systems. Such a breakthrough provided the possibility to further exploit and engineer also the biological-artificial interface and the mutual interactions between the two “intelligent” systems. In particular, in-materia computing devices, based either on memristors or physical reservoir systems, can satisfy the necessity to reduce the energy consumption by performing computation efficiently, by exploiting intrinsic characteristics of the physical substrates.
The fascinating structural and functional interplay between the extraordinary complexity and computational capabilities of a biological neural network and in-materia artificial data processing systems at different length scale draws the attention of both fundamental studies and technological applications, in a multidisciplinary context ranging from physics, biology, material and computer science. Memristors and more in general materials based on non-linear dynamics (physical reservoirs) can be exploited for the development of intelligent devices which perform data processing beyond Von-Neumann architecture, by exploiting spontaneous activity of the materials and employing low power consumption. A case of interest is the biological-artificial system interface: in particular the possibility to engineer the development of in real-time processing systems which allow the transfer of the information encoded in neuronal network signaling into a response of the neuromorphic artificial material.
The advent of advanced hardware and software solutions, which can partially emulate the behavior of the brain in simple data processing and in more complex classification and pattern recognition tasks, is a turning point in the development of neuromorphic artificial intelligent systems. Such a breakthrough provided the possibility to further exploit and engineer also the biological-artificial interface and the mutual interactions between the two “intelligent” systems. In particular, in-materia computing devices, based either on memristors or physical reservoir systems, can satisfy the necessity to reduce the energy consumption by performing computation efficiently, by exploiting intrinsic characteristics of the physical substrates.
The fascinating structural and functional interplay between the extraordinary complexity and computational capabilities of a biological neural network and in-materia artificial data processing systems at different length scale draws the attention of both fundamental studies and technological applications, in a multidisciplinary context ranging from physics, biology, material and computer science. Memristors and more in general materials based on non-linear dynamics (physical reservoirs) can be exploited for the development of intelligent devices which perform data processing beyond Von-Neumann architecture, by exploiting spontaneous activity of the materials and employing low power consumption. A case of interest is the biological-artificial system interface: in particular the possibility to engineer the development of in real-time processing systems which allow the transfer of the information encoded in neuronal network signaling into a response of the neuromorphic artificial material.
