Abstract:With the vast amounts of real-time data today, efficient storage and processing are critical to temporal data processing applications for making informed decisions. Physical memory devices provide effective solutions to manage large data volumes, making them ideal for edge systems. These devices typically necessitate compact models for device-circuit co-design. Alternatively, machine learning (ML) can quickly predict the behaviour of novel materials/ devices without explicitly defining any material properties and device physics. However, previously reported ML device models are limited by their fixed hidden layer depth, which restricts their adaptability to predict varying temporal dynamics of a complex system. Here, we propose a novel approach that utilizes a continuous-time model based on neural ordinary differential equations to predict the temporal dynamic behaviour of a charge-based device, a solid electrolyte FET, whose gate current characteristics show a unique negative differential resistance that leads to steep switching beyond the Boltzmann limit. Our model, trained on a minimal experimental dataset successfully captures device transient and steady state behaviour with a high accuracy of 0.06 (root mean squared error) and for previously unseen examples of excitatory postsynaptic current when subject to an input of variable pulse width lasting 20 – 240 milliseconds. Additionally, our model predicts device dynamics in ~ 5 seconds, with 60% reduced error over a physics-based model, which takes nearly a day on an equivalent computer. Moreover, the model can predict the variability of device characteristics from device to device by a simple change in frequency of applied signal, making it a useful tool in the design of neuromorphic systems such as reservoir computing.

Data-Driven Forecasting of Non-Equilibrium Solid-State Dynamics

Time series forecasting of multiphase microstructure evolution using deep learning

Data-Driven Acceleration of Multi-Physics Simulations

Dynamics of supercooled liquids from static averaged quantities using machine learning

Machine-learning-based data-driven discovery of nonlinear phase-field dynamics

Learning nonequilibrium statistical mechanics and dynamical phase transitions

A machine learning-oriented pseudo-field approach to accelerate runtime of molecular dynamics simulation of liquids

Data-driven probability density forecast for stochastic dynamical systems

Generative learning for forecasting the dynamics of high-dimensional complex systems

Forecasting Long-Time Dynamics in Quantum Many-Body Systems by Dynamic Mode Decomposition

Machine Learning Forecasting Of Active Nematics

Optimal Data-Driven Estimation of Generalized Markov State Models for Non-Equilibrium Dynamics.

Extrapolating tipping points and simulating non-stationary dynamics of complex systems using efficient machine learning

Accelerating phase-field-based microstructure evolution predictions via surrogate models trained by machine learning methods

A data‐free, support vector machine‐based physics‐driven estimator for dynamic response computation

Neural Ordinary Differential Equations for Predicting the Temporal Dynamics of a ZnO Solid Electrolyte FETâ€

Physics-aware, probabilistic model order reduction with guaranteed stability

Data-driven modeling and prediction of non-linearizable dynamics via spectral submanifolds

Generative Learning for Forecasting the Dynamics of Complex Systems

Accelerating Electron Dynamics Simulations through Machine Learned Time Propagators