Abstract:Kinetic model identification relies on accurate concentration measurements and physical constraints to limit solution multiplicity. Not having these measurements and prior knowledge of species and reactions creates considerable challenges that are currently unresolved. We address these by developing a data-driven framework using realtime spectroscopic data, comprising: (i) multivariate curve resolution to deconvolve the spectra of the reacting mixture into those of its pseudocomponents and their corresponding concentrations, which enables species identification without prior information, (ii) Bayesian structure learning among the pseudocomponent spectra to enable hypothesizing reaction pathways, and (iii) neural ordinary differential equations (ODE) that are physically constrained by the hypothesized reaction network and the laws of mass action and temperature dependence to learn kinetic models from the temporal concentration projections of the realtime spectra. The predictive performance of the constrained neural ODEs is limited by the accuracy of spectral deconvolution in the presence of noise, and has been benchmarked against a constrained regression approach by varying signal:noise ratios in the synthetic spectroscopic data of reacting mixtures. Although the hypothesized reaction network differs from the actual reaction template, owing to noise, the network-constrained neural ODEs are seen to result in a 75.2% and 68.15% decrease in the root mean squared error (RMSE) of the concentration profile predictions as compared to the constrained regression method , when trained on time projected concentration data of the synthetic spectra generated at a signal to noise ratio of 35 and 100, respectively.

A note on rank deficiency and numerical modeling

On Hidden Rank Deficiency in MCR Problems

Local factor analysis of rank-deficient reaction systems

Inference for Stochastic Chemical Kinetics Using Moment Equations and System Size Expansion

Rank Suggestion in Non-negative Matrix Factorization: Residual Sensitivity to Initial Conditions (RSIC)

Chemical Rank Estimation By Noise Perturbation In Functional Principal Component Analysis

Behavior of Spiral Wave Spectra with a Rank-Deficient Diffusion Matrix

Benchmarking chemical neural ordinary differential equations to obtain reaction network-constrained kinetic models from spectroscopic data

A hierarchical dynamical low-rank algorithm for the stochastic description of large reaction networks

Rank-one matrix estimation: analysis of algorithmic and information theoretic limits by the spatial coupling method

A low-rank complexity reduction algorithm for the high-dimensional kinetic chemical master equation

Two new methods for the estimation and interpretation of the range of feasible profiles in multivariate curve resolution and their implications to analytical chemistry

Low rank approximation for the numerical simulation of high dimensional Lindblad and Riccati equations

Rank-deficiencies in a reduced information latent variable model

An Augmented Classical Least Squares Method for Quantitative Raman Spectral Analysis Against Component Information Loss

Modeling of chemical reaction systems with detailed balance using gradient structures

Essential Number of Principal Components and Nearly Training-Free Model for Spectral Analysis

A Stein's Method Approach to the Linear Noise Approximation for Stationary Distributions of Chemical Reaction Networks

Estimation of high-dimensional low-rank matrices

The Reaction–Diffusion Models in Biomedicine: Highly Accurate Calculations Via a Hybrid Matrix Collocation Algorithm