1: \begin{abstract}

2: 	Neural networks (NN) are extensively studied in cutting-edge soft sensor models due to their  feature extraction and function approximation capabilities. 

3: 	Current research into network-based methods primarily focuses on models' offline accuracy. 

4: 	Notably, in industrial soft sensor context, online optimizing stability and interpretability are prioritized, followed by accuracy. 

5: 	This requires a clearer understanding of network's training process. 

6: 	To bridge this gap, we propose a novel NN named the Approximated Orthogonal Projection Unit (AOPU) which has solid mathematical basis and presents superior training stability. 

7: 	AOPU truncates the gradient backpropagation at dual parameters, optimizes the trackable parameters updates, and enhances the robustness of training. 

8: 	We further prove that AOPU attains minimum variance estimation (MVE) in NN, wherein the truncated gradient approximates the natural gradient (NG). 

9: 	Empirical results on two chemical process datasets clearly show that AOPU outperforms other models in achieving stable convergence, marking a significant advancement in soft sensor field.

10: \end{abstract}

11: