Abstract:Accurate quality variable inference by process variables is the core of industrial inferential sensor modeling, where recent advancements have seen deep learning (DL) models achieving remarkable success. However, integrating knowledge of unit operations is critical for improving inferential sensor performance, yet it has received little attention. The main challenge lies in the incompleteness and correctness of industrial knowledge due to its semi-empirical nature and inevitable engineering errors. Addressing this, this article introduces the gradient knowledge network based on the graph neural network’s message-passing mechanism within the variational Bayesian inference framework, which naturally copes with the abovementioned issues by fusing observational data. Initially, the prior knowledge about the process variables, which mirrors the graph in graph neural network, is parameterized as Dirichlet distribution based on the analysis of message-passing mechanism. However, the divergence computation and normalization constraints are challenging for model implementation. To navigate these challenges, the Bayesian inference problem is transformed into an optimization problem, subsequently recast as a simulation problem induced by the gradient field, ensuring compatibility with DL backends. Furthermore, a theoretical iteration equation is derived to maintain the normalization constraint. The architecture of the proposed model and its learning algorithm are then detailed. Finally, various experiments are conducted on two real industrial processes to demonstrate the model’s efficacy from the perspective of prediction accuracy, sensitivity analysis, and ablation study.

A comparison of Single- and Double-generator formalisms for Thermodynamics-Informed Neural Networks

A comparison of single and double generator formalisms for thermodynamics-informed neural networks

Graph neural networks informed locally by thermodynamics

Advancements in Thermochemical Predictions: A Multi-Output Thermodynamics-Informed Neural Network Approach

Improving Data-Driven Inferential Sensor Modeling by Industrial Knowledge: A Bayesian Perspective

A statistical mechanics framework for Bayesian deep neural networks beyond the infinite-width limit

Gibbs-Duhem-Informed Neural Networks for Binary Activity Coefficient Prediction

Thermodynamic Overfitting and Generalization: Energetic Limits on Predictive Complexity

Physically optimizing inference

A theory of data variability in Neural Network Bayesian inference

The principles of adaptation in organisms and machines II: Thermodynamics of the Bayesian brain

Thermodynamic Fidelity of Generative Models for Ising System

Thermodynamics-informed super-resolution of scarce temporal dynamics data

Partitioned integrators for thermodynamic parameterization of neural networks

On Inductive Biases That Enable Generalization of Diffusion Transformers

InVAErt networks: A data-driven framework for model synthesis and identifiability analysis

Thermodynamic Bayesian Inference

Learning Feynman integrals from differential equations with neural networks

The twin peaks of learning neural networks

From statistical inference to a differential learning rule for stochastic neural networks