*Result*: PLSTM-MTGF: A deep learning fusion model enabling real-time multi-target monitoring of penicillin fermentation via near-infrared spectroscopy.

Title:
PLSTM-MTGF: A deep learning fusion model enabling real-time multi-target monitoring of penicillin fermentation via near-infrared spectroscopy.
Authors:
Shen N; School of Information Science and Engineering, Hebei University of Science and Technology, Shijiazhuang, China. Electronic address: 3524527113@qq.com., Wang J; School of Information Science and Engineering, Hebei University of Science and Technology, Shijiazhuang, China., Wang X; School of Information Science and Engineering, Hebei University of Science and Technology, Shijiazhuang, China., Ma J; School of Information Science and Engineering, Hebei University of Science and Technology, Shijiazhuang, China., Wu S; School of Information Science and Engineering, Hebei University of Science and Technology, Shijiazhuang, China.
Source:
Spectrochimica acta. Part A, Molecular and biomolecular spectroscopy [Spectrochim Acta A Mol Biomol Spectrosc] 2026 Mar 15; Vol. 349, pp. 127358. Date of Electronic Publication: 2025 Dec 14.
Publication Type:
Journal Article
Language:
English
Journal Info:
Publisher: Elsevier Country of Publication: England NLM ID: 9602533 Publication Model: Print-Electronic Cited Medium: Internet ISSN: 1873-3557 (Electronic) Linking ISSN: 13861425 NLM ISO Abbreviation: Spectrochim Acta A Mol Biomol Spectrosc Subsets: MEDLINE
Imprint Name(s):
Publication: : Amsterdam : Elsevier
Original Publication: [Kidlington, Oxford, U.K. ; Tarrytown, NY] : Pergamon, c1994-
Contributed Indexing:
Keywords: Deep learning; Multi-task learning; Near-infrared spectroscopy; Partial least squares regression; Penicillin fermentation
Substance Nomenclature:
0 (Penicillins)
Entry Date(s):
Date Created: 20251217 Date Completed: 20260113 Latest Revision: 20260113
Update Code:
20260130
DOI:
10.1016/j.saa.2025.127358
PMID:
41406794
Database:
MEDLINE

*Further Information*

*Real-time, multi-parameter monitoring is essential for optimizing industrial penicillin fermentation, yet conventional assays are slow and prone to contamination. This study introduces a lightweight dual-stream fusion model-Partial Least Squares Regression-Long Short-Term Memory with Multi-Task Gated Fusion (PLSTM-MTGF)-for online monitoring of four key indicators via near-infrared spectroscopy: residual sugar, amino nitrogen, cell density, and potency. The architecture integrates partial least squares regression for spectral compression with a long short-term memory-self-attention network to capture process dynamics, and employs a multi-task gated fusion mechanism to balance task-specific representations. Evaluated on 24 industrial batches under a strict batch-wise scheme, the model achieved high accuracy across all targets (R<sup>2</sup> = 0.89-0.98; RPD = 3.0-7.1) and outperformed representative linear and deep-learning benchmarks. Ablation studies confirmed the indispensability of both spectral and temporal streams and the gating strategy. With a compact size of 0.97 M parameters and real-time inference on a standard CPU, the model is readily deployable for routine bioprocess monitoring. These results validate that fusing chemometrics with deep learning provides a robust, interpretable strategy for multi-target, online bioprocess analytics.
(Copyright © 2025 Elsevier B.V. All rights reserved.)*

*Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.*