ST-HCSS: Deep Spatio-Temporal Hypergraph Convolutional Neural Network for Soft Sensing

Hwa Hui Tew 1 , Fan Ding 1 , Gaoxuan Li 1 , Junn Yong Loo 1,* , Chee-Ming Ting 1 , Ze Yang Ding 2 , Chee Pin Tan 2 · 2025

Read ST-HCSS: Deep Spatio-Temporal Hypergraph Convolutional Neural Network for Soft Sensing by Hwa Hui Tew 1 , Fan Ding 1 , Gaoxuan Li 1 , Junn Yong Loo 1,* , Chee-Ming Ting 1 , Ze Yang Ding 2 , Chee Pin Tan 2 (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-02016.txt.

Description

Hwa Hui Tew and six co-authors build a neural network that estimates hard-to-measure industrial quantities from the sensors a plant already has. Such soft sensing matters because key process variables often cannot be probed directly. Ordinary graph networks link sensors only in pairs, which misses the group interactions that drive nonlinear plant behavior.

Their model, ST-HCSS, learns a hypergraph over the sensors, letting one edge join many nodes at once, without needing the plant's structure in advance. Stacked gated temporal layers and hypergraph convolution layers then pass information across time and nodes. In tests the network beat existing state-of-the-art soft sensors, and its learned groupings lined up well with the true correlations in the sensor data. The authors have posted the code publicly.

Catalog record

CREATORHwa Hui Tew 1 , Fan Ding 1 , Gaoxuan Li 1 , Junn Yong Loo 1,* , Chee-Ming Ting 1 , Ze Yang Ding 2 , Chee Pin Tan 2
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE20 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-02016

Browse the full library · About the Tower of Babel