Collective dynamics and long-range order in thermal neuristor networks

Yuan-Hang Zhang,Chesson Sipling,Erbin Qiu,Ivan K. Schuller,Massimiliano Di Ventra

2024-07-11

Abstract:In the pursuit of scalable and energy-efficient neuromorphic devices, recent research has unveiled a novel category of spiking oscillators, termed "thermal neuristors." These devices function via thermal interactions among neighboring vanadium dioxide resistive memories, emulating biological neuronal behavior. Here, we show that the collective dynamical behavior of networks of these neurons showcases a rich phase structure, tunable by adjusting the thermal coupling and input voltage. Notably, we identify phases exhibiting long-range order that, however, does not arise from criticality, but rather from the time non-local response of the system. In addition, we show that these thermal neuristor arrays achieve high accuracy in image recognition and time series prediction through reservoir computing, without leveraging long-range order. Our findings highlight a crucial aspect of neuromorphic computing with possible implications on the functioning of the brain: criticality may not be necessary for the efficient performance of neuromorphic systems in certain computational tasks.

Disordered Systems and Neural Networks,Materials Science,Adaptation and Self-Organizing Systems

What problem does this paper attempt to address?

This paper explores the collective dynamics and long-range order of a thermal neuron network. The research team discovered a new type of pulse oscillator called "thermally assisted resistive switch" when searching for scalable and energy-efficient neural morphic devices. These devices operate through thermal interaction between adjacent vanadium oxide resistive memory cells, mimicking the behavior of biological neurons.

In the thermal neuron network, its rich phase structure can be modulated by adjusting thermal coupling and input voltage. The researchers identified phases that exhibit long-range order, but this order is not due to criticality, but rather due to the non-local time response of the system. Furthermore, they demonstrated that these arrays of thermal resistive switches achieve high accuracy in tasks such as image recognition and time series prediction through reservoir computing, without relying on long-range order.

The paper indicates that criticality may not be a necessary condition for efficient operation of neural morphic systems in certain computational tasks, providing a new perspective on understanding brain functionality, namely that efficient brain operation does not necessarily require being in a critical state. The researchers verified the thermal resistive switches based on VO2 through experiments, and analyzed their collective dynamic behavior through numerical simulations. They found that different phases can be tuned by adjusting thermal coupling and input voltage, and these arrays can also perform image recognition tasks without the need for long-range order.

In conclusion, this paper challenges the importance hypothesis of criticality in neural morphic computing and brain functionality, revealing the potential of non-critical systems in certain computational tasks.