Integer Factorization with Stochastic Spiking in Percolating Networks of Nanoparticles.

Area of Science:

• Nanotechnology
• Computer Science
• Physics

Background:

• Global demand for computing power is increasing, driving the need for novel hardware solutions beyond traditional transistors.
• Stochastic nanoscale devices are being explored to accelerate probabilistic computing algorithms.
• Percolating Networks of Nanoparticles (PNNs) exhibit stochastic spiking and criticality, offering potential computational advantages.

Purpose of the Study:

• To investigate the use of spiking PNNs as core stochastic components in coupled networks.
• To demonstrate the capability of PNNs for performing integer factorization.
• To explore the operational characteristics and efficiency of PNNs for computational tasks.

Main Methods:

• Developing and configuring coupled networks of Percolating Networks of Nanoparticles (PNNs).
• Utilizing the stochastic spiking behavior of PNNs to perform probabilistic computing.
• Testing the networks' ability to factor integers up to 945.
• Demonstrating asynchronous operation and the capacity for simultaneous solution generation.

Main Results:

• Spiking PNNs were successfully employed as the core stochastic components in coupled networks.
• These networks achieved the factorization of integers up to 945.
• Asynchronous operation was demonstrated.
• A single PNN device proved sufficient for solving all tested factorization tasks and generating multiple solutions concurrently.

Conclusions:

• Spiking Percolating Networks of Nanoparticles (PNNs) are a viable approach for accelerating probabilistic computing.
• PNNs demonstrate significant potential for integer factorization and other computational tasks requiring stochasticity.
• The criticality and asynchronous operation of PNNs offer unique advantages for future computing paradigms.