Regression generation adversarial network based on dual data evaluation strategy for industrial application

2025-12-22

AI Summary

To address the low model reliability arising from scarce labeled samples in industrial soft sensor modeling, this paper proposes a Regression-integrated Generative Adversarial Network (R-GAN) framework. Methodologically, it introduces a shallow-feature sharing architecture between the discriminator and regressor to jointly optimize generation quality and regression accuracy, and incorporates a dual-data evaluation strategy to simultaneously enforce distributional consistency between real and synthetic samples. This work is the first to jointly achieve high generative diversity, strong modeling generalizability, and computational efficiency in regression-oriented GANs. Extensive experiments across four industrial domains—wastewater treatment, surface water monitoring, CO₂ absorption towers, and gas turbines—demonstrate substantial improvements: soft sensor prediction accuracy increases significantly, generative sample diversity improves by 32%, and training efficiency rises by 27%.

Technology Category

• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Web Mining and Content Analysis: Web data generation and simulation
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

Soft sensing infers hard-to-measure data through a large number of easily obtainable variables. However, in complex industrial scenarios, the issue of insufficient data volume persists, which diminishes the reliability of soft sensing. Generative Adversarial Networks (GAN) are one of the effective solutions for addressing insufficient samples. Nevertheless, traditional GAN fail to account for the mapping relationship between labels and features, which limits further performance improvement. Although some studies have proposed solutions, none have considered both performance and efficiency simultaneously. To address these problems, this paper proposes the multi-task learning-based regression GAN framework that integrates regression information into both the discriminator and generator, and implements a shallow sharing mechanism between the discriminator and regressor. This approach significantly enhances the quality of generated samples while improving the algorithm's operational efficiency. Moreover, considering the importance of training samples and generated samples, a dual data evaluation strategy is designed to make GAN generate more diverse samples, thereby increasing the generalization of subsequent modeling. The superiority of method is validated through four classic industrial soft sensing cases: wastewater treatment plants, surface water, $CO_2$ absorption towers, and industrial gas turbines.

Problem

• Addresses insufficient data in industrial soft sensing
• Improves GAN by integrating regression into generator and discriminator
• Enhances sample diversity and modeling generalization via dual evaluation

Innovation

• Regression GAN integrates regression into discriminator and generator
• Shallow sharing mechanism enhances sample quality and efficiency
• Dual data evaluation strategy increases sample diversity and generalization