The Discrete Langevin Machine: Bridging the Gap Between Thermodynamic and Neuromorphic Systems
arXiv:1901.05214 · doi:10.1103/PhysRevE.101.063304
Abstract
A formulation of Langevin dynamics for discrete systems is derived as a class of generic stochastic processes. The dynamics simplify for a two-state system and suggest a network architecture which is implemented by the Langevin machine. The Langevin machine represents a promising approach to compute successfully quantitative exact results of Boltzmann distributed systems by LIF neurons. Besides a detailed introduction of the dynamics, different simplified models of a neuromorphic hardware system are studied with respect to a control of emerging sources of errors.
arXiv comment25 pages, 16 figures
References in corpus (20)
- Solving the Quantum Many-Body Problem with Artificial Neural Networks
- Efficient Representation of Quantum Many-body States with Deep Neural Networks
- Event-Driven Contrastive Divergence for Spiking Neuromorphic Systems
- Stochastic Synapses Enable Efficient Brain-Inspired Learning Machines
- Stochastic quantization at finite chemical potential
- Noise Induced Complexity: From Subthreshold Oscillations to Spiking in Coupled Excitable Systems
- A Brief History of Excitable Map-Based Neurons and Neural Networks
- Quenches near Ising quantum criticality as a challenge for artificial neural networks
- Universality away from critical points in two-dimensional phase transitions
- Nonmonotonic Behavior of Spatiotemporal Pattern Formation in a Noisy Lotka-Volterra System
- Stochastic inference with spiking neurons in the high-conductance state
- Stability of complex Langevin dynamics in effective models
- Complex Langevin dynamics and other approaches at finite chemical potential
- Stochastic FitzHugh-Nagumo neuron model in excitable regime embeds a leaky integrate-and-fire model
- Stochasticity from function -- why the Bayesian brain may need no noise
- Critical phenomena and noise-induced phase transitions in neuronal networks
- Deterministic networks for probabilistic computing
- Information Geometry of Quantum Resources
- Probabilistic inference in discrete spaces can be implemented into networks of LIF neurons
- Spike-based probabilistic inference with correlated noise