Forecasting the outcome of spintronic experiments with Neural Ordinary Differential Equations

Neural Ordinary Differential Equations for Forecasting and Accelerating Photon Correlation Spectroscopy

Evaluating the quantum optical properties of solid-state single-photon emitters is a time-consuming task that typically requires interferometric photon correlation experiments. Photon correlation Fourier spectroscopy (PCFS) is one such technique that measures time-resolved single-emitter line shapes and offers additional spectral information over Hong-Ou-Mandel two-photon interference but requires long experimental acquisition times. Here, we demonstrate a neural ordinary differential equation model, g2NODE, that can forecast a complete and noise-free interferometry experiment from a small subset of noisy correlation functions. We demonstrate this for simulated and experimental data, where g2NODE utilizes 10-20 noisy measured photon correlation functions to create entire denoised interferograms of up to 200 stage positions, enabling up to a 20-fold speedup in experimental acquisition time from hours to minutes. Our work presents a new deep learning approach to greatly accelerate the use of photon correlation spectroscopy as an experimental characterization tool for novel quantum emitter materials.

Predicting tissue distribution and tumor delivery of nanoparticles in mice using machine learning models

Nanoparticles (NPs) can be designed for targeted delivery in cancer nanomedicine, but the challenge is a low delivery efficiency (DE) to the tumor site. Understanding the impact of NPs' physicochemical properties on target tissue distribution and tumor DE can help improve the design of nanomedicines. Multiple machine learning and artificial intelligence models, including linear regression, support vector machine, random forest, gradient boosting, and deep neural networks (DNN), were trained and validated to predict tissue distribution and tumor delivery based on NPs' physicochemical properties and tumor therapeutic strategies with the dataset from Nano-Tumor Database. Compared to other machine learning models, the DNN model had superior predictions of DE to tumors and major tissues. The determination coefficients (R2) for the test datasets were 0.41, 0.42, 0.45, 0.79, 0.87, and 0.83 for DE in tumor, heart, liver, spleen, lung, and kidney, respectively. All the R2 and root mean squared error (RMSE) results of the test datasets were similar to the 5-fold cross validation results. Feature importance analysis showed that the core material of NPs played an important role in output predictions among all physicochemical properties. Furthermore, multiple NP formulations with greater DE to the tumor were determined by the DNN model. To facilitate model applications, the final model was converted to a web dashboard. This model could serve as a high-throughput pre-screening tool to support the design of new and efficient nanomedicines with greater tumor DE and serve as an alternative tool to reduce, refine, and partially replace animal experimentation in cancer nanomedicine research.

Forecasting with an N-dimensional Langevin equation and a neural-ordinary differential equation

Accurate prediction of electricity day-ahead prices is essential in competitive electricity markets. Although stationary electricity-price forecasting techniques have received considerable attention, research on non-stationary methods is comparatively scarce, despite the common prevalence of non-stationary features in electricity markets. Specifically, existing non-stationary techniques will often aim to address individual non-stationary features in isolation, leaving aside the exploration of concurrent multiple non-stationary effects. Our overarching objective here is the formulation of a framework to systematically model and forecast non-stationary electricity-price time series, encompassing the broader scope of non-stationary behavior. For this purpose, we develop a data-driven model that combines an N-dimensional Langevin equation (LE) with a neural-ordinary differential equation (NODE). The LE captures fine-grained details of the electricity-price behavior in stationary regimes but is inadequate for non-stationary conditions. To overcome this inherent limitation, we adopt a NODE approach to learn, and at the same time predict, the difference between the actual electricity-price time series and the simulated price trajectories generated by the LE. By learning this difference, the NODE reconstructs the non-stationary components of the time series that the LE is not able to capture. We exemplify the effectiveness of our framework using the Spanish electricity day-ahead market as a prototypical case study. Our findings reveal that the NODE nicely complements the LE, providing a comprehensive strategy to tackle both stationary and non-stationary electricity-price behavior. The framework's dependability and robustness is demonstrated through different non-stationary scenarios by comparing it against a range of basic naïve methods.

Topological magnetic structure generation using VAE-GAN hybrid model and discriminator-driven latent sampling

Recently, deep generative models using machine intelligence are widely utilized to investigate scientific systems by generating scientific data. In this study, we experiment with a hybrid model of a variational autoencoder (VAE) and a generative adversarial network (GAN) to generate a variety of plausible two-dimensional magnetic topological structure data. Due to the topological properties in the system, numerous and diverse metastable magnetic structures exist, and energy and topological barriers separate them. Thus, generating a variety of plausible spin structures avoiding those barrier states is a challenging problem. The VAE-GAN hybrid model can present an effective approach to this problem because it brings the advantages of both VAE's diversity and GAN's fidelity. It allows one to perform various applications including searching a desired sample from a variety of valid samples. Additionally, we perform a discriminator-driven latent sampling (DDLS) using our hybrid model to improve the quality of generated samples. We confirm that DDLS generates various plausible data with large coverage, following the topological rules of the target system.

Experimental demonstration of a skyrmion-enhanced strain-mediated physical reservoir computing system

Physical reservoirs holding intrinsic nonlinearity, high dimensionality, and memory effects have attracted considerable interest regarding solving complex tasks efficiently. Particularly, spintronic and strain-mediated electronic physical reservoirs are appealing due to their high speed, multi-parameter fusion and low power consumption. Here, we experimentally realize a skyrmion-enhanced strain-mediated physical reservoir in a multiferroic heterostructure of Pt/Co/Gd multilayers on (001)-oriented 0.7PbMg1/3Nb2/3O3-0.3PbTiO3 (PMN-PT). The enhancement is coming from the fusion of magnetic skyrmions and electro resistivity tuned by strain simultaneously. The functionality of the strain-mediated RC system is successfully achieved via a sequential waveform classification task with the recognition rate of 99.3% for the last waveform, and a Mackey-Glass time series prediction task with normalized root mean square error (NRMSE) of 0.2 for a 20-step prediction. Our work lays the foundations for low-power neuromorphic computing systems with magneto-electro-ferroelastic tunability, representing a further step towards developing future strain-mediated spintronic applications.

Machine learning and artificial intelligence in physiologically based pharmacokinetic modeling

Physiologically based pharmacokinetic (PBPK) models are useful tools in drug development and risk assessment of environmental chemicals. PBPK model development requires the collection of species-specific physiological, and chemical-specific absorption, distribution, metabolism, and excretion (ADME) parameters, which can be a time-consuming and expensive process. This raises a need to create computational models capable of predicting input parameter values for PBPK models, especially for new compounds. In this review, we summarize an emerging paradigm for integrating PBPK modeling with machine learning (ML) or artificial intelligence (AI)-based computational methods. This paradigm includes 3 steps (1) obtain time-concentration PK data and/or ADME parameters from publicly available databases, (2) develop ML/AI-based approaches to predict ADME parameters, and (3) incorporate the ML/AI models into PBPK models to predict PK summary statistics (eg, area under the curve and maximum plasma concentration). We also discuss a neural network architecture "neural ordinary differential equation (Neural-ODE)" that is capable of providing better predictive capabilities than other ML methods when used to directly predict time-series PK profiles. In order to support applications of ML/AI methods for PBPK model development, several challenges should be addressed (1) as more data become available, it is important to expand the training set by including the structural diversity of compounds to improve the prediction accuracy of ML/AI models; (2) due to the black box nature of many ML models, lack of sufficient interpretability is a limitation; (3) Neural-ODE has great potential to be used to generate time-series PK profiles for new compounds with limited ADME information, but its application remains to be explored. Despite existing challenges, ML/AI approaches will continue to facilitate the efficient development of robust PBPK models for a large number of chemicals.

Pattern recognition with neuromorphic computing using magnetic field-induced dynamics of skyrmions

Nonlinear phenomena in physical systems can be used for brain-inspired computing with low energy consumption. Response from the dynamics of a topological spin structure called skyrmion is one of the candidates for such a neuromorphic computing. However, its ability has not been well explored experimentally. Here, we experimentally demonstrate neuromorphic computing using nonlinear response originating from magnetic field-induced dynamics of skyrmions. We designed a simple-structured skyrmion-based neuromorphic device and succeeded in handwritten digit recognition with the accuracy as large as 94.7% and waveform recognition. Notably, there exists a positive correlation between the recognition accuracy and the number of skyrmions in the devices. The large degrees of freedom of skyrmion systems, such as the position and the size, originate from the more complex nonlinear mapping, the larger output dimension, and, thus, high accuracy. Our results provide a guideline for developing energy-saving and high-performance skyrmion neuromorphic computing devices.