Nearly half of global anthropogenic N2O emissions are accounted for by nitrogen fertiliser application. Therefore, accurate prediction of fertiliser-induced N2O fluxes is crucial for optimising fertiliser strategies and mitigating climate change. In this work, we introduce DA-Mamba: a data augmentation-enhanced state space model that can capture long-range N2O flux dynamics and their interactions with agri-environmental variables, even when data is limited. Using a publicly available dataset of fertiliser-induced N2O emissions, DA-Mamba achieves state-of-the-art performance, outperforming six baseline models. Additionally, we have integrated DA-Mamba as a containerised microservice within ParallelFarm, our cloud-based precision fertilisation and farm management system. The service uses real-time weather, soil and management data to generate optimised fertiliser plans and field-level N2O–yield predictions, thereby supporting sustainable agricultural decision-making.

To tackle the technical challenges of precisely aligning and inserting test tubes into sampling needles, this paper proposes an autonomous operation method for blood gas analyzers based on reinforcement learning. To simplify the complexity of the alignment and insertion task, it is decomposed into two independent subtasks, which are learned in a distributed manner and executed sequentially to complete the overall needle insertion process. To enable accurate perception of dynamic environmental states, a reinforcement learning model is developed that defines the state space, action space, and multi-task reward functions. Building on this framework, we enhance the Actor-Critic structure of the Proximal Policy Optimization (PPO) algorithm by introducing a dual-network architecture that integrates Long Short-Term Memory (LSTM) and Wavelet Transform Convolution (WTC) networks. This enhancement significantly improves both learning efficiency and policy stability. Extensive comparative experiments demonstrate that the proposed LSTM-WTC-PPO algorithm achieves a high success rate, stable convergence, and efficient policy optimization.

In real-world industrial scenarios, fault detection faces the widely recognized challenge of data imbalance, which not only refers to the scarcity of fault data but also includes the imbalance in healthy data. This article is concerned with imbalanced open-set fault detection (IOSFD), a practical yet challenging scenario in industrial applications where multiple healthy operating conditions and multiple fault types are imbalanced. In this article, we propose a new contrastive conditional adversarial autoencoder for IOSFD. It constructs an end-to-end unified model based on multi-class known healthy and faulty data to address the reliance of traditional methods on fault samples, while optimizing with class-specific weighted forces to ensure equal attention to imbalanced known classes. Input and feature reconstruction conditioned on operating modes are utilized to learn a compact decision plane and achieve both unknown fault detection and known data classification. Significantly, we formulate the optimization objective of conditional reconstruction based on contrastive learning and introduce adversarial training to further enhance the model’s performance. The effectiveness of the proposed method is validated through real-world pipeline leak detection and Tennessee-Eastman multi-fault detection.

Lifelong learning (LL) has proven successful in computer vision and natural language processing, but its applications in condition monitoring are still largely understudied. To bridge this gap, this article innovatively proposed a LL-based condition monitoring framework for rotating machinery over the whole life cycle. First, the development of LL is reviewed and summarized. Inspired by the core idea of knowledge maintenance and transfer in LL, the innovation of this article is to integrate data-driven modeling with LL to extract process knowledge-driven features for autonomous condition monitoring. This opens up an emerging lifelong monitoring paradigm for mechanical systems. Based on the extracted features learned from the data-driven model, a threshold construction method and an online monitoring strategy are integrated into the monitoring framework. Finally, the utility of the proposed framework is demonstrated through the rolling bearing performance degradation and shaft fatigue fracture test cases. Compared to the state-of-the-art monitoring methods, the proposed framework shows significantly improved adaptability and reliability.

Probabilistic latent variable models (PLVMs), such as probabilistic principal component analysis (PPCA), are widely employed in process monitoring and fault detection of industrial processes. This article proposes a novel deep PPCA (DePPCA) model, which has the advantages of both probabilistic modeling and deep learning. The construction of DePPCA includes a greedy layer-wise pretraining phase and a unified end-to-end fine-tuning phase. The former establishes a hierarchical deep structure based on cascading multiple layers of the PPCA module to extract high-level features. The latter builds an end-to-end connection between the raw inputs and the final outputs to further improve the representation of the model to high-level features. After constructing the model structure of DePPCA, we first present the detailed training processes of the pretraining and fine-tuning stages, then clarify the theoretical merits of the proposed model from the perspective of variational inference. For process monitoring purposes, we develop two statistics based on the established DePPCA. The monitoring performance of these two statistics can remain superior even if the features extracted by DePPCA are significantly compressed to univariate. This makes the feature extraction process and online monitoring procedure of DePPCA quite fast. In other words, the proposed DePPCA can achieve accurate and efficient process monitoring by only extracting one feature for each sample. Finally, the effectiveness of DePPCA is evaluated on the Tennessee Eastman (TE) process and the multiphase flow (MPF) facility.

Promoting sustainable agriculture plays a crucial role in reducing greenhouse gas (GHG) emissions, lowering the carbon footprint, and improving farm resilience. Three challenges must be overcome to achieve sustainable agriculture management. Firstly, there is a lack of reliable and sustainable fertiliser solutions to improve fertiliser use efficiency, reduce GHG emissions while maintaining crop production. In addition, how to cost-effectively quantify the response of soil carbon and GHG fluxes to different fertilisation practices. Thirdly, there is a requirement to integrate multi-source farming data and AI models into a farm management information system (FMIS) to support intelligent decisions for farmers. To address these challenges, we developed the ParellelFarm, an AI-enabled sustainable farming management system that integrates multi-source farming data and AI-driven fertiliser and soil carbon models into a multi-tenant cloud platform, to support sustainable farming. It also provides remote field visualisation and management as well as instant messaging via web and mobile clients, supporting fast and accurate labour allocation with fewer resources. It is a potential solution for a cost-effective, highly productive and sustainable modern net-zero farm.

The uneven distribution of process industrial data poses a significant challenge for soft sensor modeling. Hence, it is necessary to employ generative models to generate some new data used for augmenting the distribution fitting ability of the model by the full utilization of sparse regions. Existing generative models are mostly applied in the image and text generation fields, which are more suitable for discrete data where each variable only consists of integers. To enhance the applicability of generative models in industrial data modeling, a novel vector-quantized weighted-Wasserstein variational autoencoder based soft sensor is developed in this article to solve uneven distribution data. The proposed model further enhances the advantages of present generative models in generating qualified data by redesigning the model structure for clustering and training the data, thereby greatly alleviating poor generative performance caused by sparse regions. Specifically, first, the vector quantization strategy is designed to address the problem of sparse regions. After that, the sparse regions are transferred into the boundary of each group to generate more well-proportioned points. Then, the weighted and continuous variation distance mixture strategy is introduced to better evaluate the divergence between the real and learned distribution without mutation in the training part. Finally, a soft sensor model with the above data augmentation strategies is developed. By comparing it with three state-of-the-art augmentation strategies on a real polyester polymerization dataset, the superiority of the proposed soft sensor model with the novel data augmentation strategies is unequivocally verified.

As an emerging learning paradigm, lifelong learning solves multiple consecutive tasks based upon previously accumulated knowledge. When facing with a new task, existing lifelong learning approaches need both input and desired output data to construct task models before knowledge transfer can succeed. However, labeling each task requires extensive labors and time, which can be prohibitive for real-world lifelong regression problems. To reduce this burden, we propose to incorporate unsupervised feature into lifelong regression via coupled dictionary learning, enabling to learn new tasks without target output data. Specifically, the input data for each task is encoded as unsupervised feature while both input and output data are used to construct task predictor. The unsupervised feature is linked with task predictor through two dictionaries that are coupled by a joint sparse representation. Because of the learned coupling between the two spaces, the task predictor for the new coming task can be recovered given only the input data. We further incorporate active task selection into this framework, enabling actively choosing tasks to learn in a task-efficient manner. Three case studies are used to evaluate the effectiveness of our method, in comparison with existing lifelong learning approaches. Results show that our method is able to accurately predict new tasks through unsupervised transfer, eliminating the need to label tasks before constructing the predictor.

Alzheimer's disease (AD), the most prevalent dementia, gradually reduces the cognitive abilities of patients while also posing a significant financial burden on the healthcare system. A variety of multi-task learning methods have recently been proposed in order to identify potential MRI-related biomarkers and accurately predict the progression of AD. These methods, however, all use a predefined task relation structure that is rigid and insufficient to adequately capture the intricate temporal relations among tasks. Instead, we propose a novel mechanism for directly and automatically learning the temporal relation and constructing it as an Automatic Temporal relation Graph (AutoTG). We use the sparse group Lasso to select a universal MRI feature set for all tasks and particular sets for various tasks in order to find biomarkers that are useful for predicting the progression of AD. To solve the biconvex and non-smooth objective function, we adopt the alternating optimization and show that the two related sub-optimization problems are amenable to closed-form solution of the proximal operator. To solve the two problems efficiently, the accelerated proximal gradient method is used, which has the fastest convergence rate of any first-order method. We have preprocessed three latest AD datasets, and the experimental results verify our proposed novel multi-task approach outperforms several baseline methods. To demonstrate the high interpretability of our approach, we visualise the automatically learned temporal relation graph and investigate the temporal patterns of the important MRI features. The implementation source can be found athttps://github.com/menghui-zhou/MAGPP.

Multi-task learning models based on temporal smoothness assumption, in which each time point of a sequence of time points concerns a task of prediction, assume the adjacent tasks are similar to each other. However, the effect of outliers is not taken into account. In this paper, we show that even only one outlier task will destroy the performance of the entire model. To solve this problem, we propose two Robust Temporal Smoothness (RoTS) frameworks. Compared with the existing models based on temporal relation, our methods not only chase the temporal smoothness information but identify outlier tasks, however, without increasing the computational complexity. Detailed theoretical analyses are presented to evaluate the performance of our methods. Experimental results on synthetic and real-life datasets demonstrate the effectiveness of our frameworks. We also discuss several potential specific applications and extensions of our RoTS frameworks.

Identifying and utilising various biomarkers for tracking Alzheimer’s disease (AD) progression have received many recent attentions and enable helping clinicians make the prompt decisions. Traditional progression models focus on extracting morphological biomarkers in regions of interest (ROIs) from MRI/PET images, such as regional average cortical thickness and regional volume. They are effective but ignore the relationships between brain ROIs over time, which would lead to synergistic deterioration. For exploring the synergistic deteriorating relationship between these biomarkers, in this paper, we propose a novel spatio-temporal similarity measure based multi-task learning approach for effectively predicting AD progression and sensitively capturing the critical relationships between biomarkers. Specifically, we firstly define a temporal measure for estimating the magnitude and velocity of biomarker change over time, which indicate a changing trend(temporal). Converting this trend into the vector, we then compare this variability between biomarkers in a unified vector space(spatial). The experimental results show that compared with directly ROI based learning, our proposed method is more effective in predicting disease progression. Our method also enables performing longitudinal stability selection to identify the changing relationships between biomarkers, which play a key role in disease progression. We prove that the synergistic deteriorating biomarkers between cortical volumes or surface areas have a significant effect on the cognitive prediction.

Alzheimer’s disease (AD), the most prevalent dementia, gradually reduces the cognitive abilities of patients while also posing a significant financial burden on the healthcare system. A variety of multi-task learning methods have recently been proposed to identify potential MRI-related biomarkers and accurately predict the progression of AD. These methods, however, all use a predefined task relation structure that is rigid and insufficient to adequately capture the intricate temporal relations among tasks. Instead, we propose a novel mechanism for directly and automatically learning the temporal relation and constructing it as an Automatic Temporal relation Graph (AutoTG). We use the sparse group Lasso to select a universal MRI feature set for all tasks and particular sets for various tasks in order to find biomarkers that are useful for predicting the progression of AD. To solve the biconvex and nonsmooth objective function, we adopt the alternating optimization and show that the two related suboptimization problems are amenable to closed-form solution of the proximal operator. To solve the two problems efficiently, the accelerated proximal gradient method is used, which has the fastest convergence rate of first-order method. We have preprocessed two latest AD datasets, and the experimental results verify our proposed novel multi-task approach outperforms several baseline methods. To demonstrate the high interpretability of our approach, we visualize the automatically learned temporal relation graph and investigate the temporal patterns of the important MRI features. The implementation source is at https://github.com/menghui-zhou/MAGPP.

Real-time optimization plays a key role in improving energy efficiency and the operational effectiveness of industrial systems. To deal with unknown process characteristics and safety constraints, a novel safe adaptive real-time exploration and optimization (ARTEO) algorithm is proposed recently for safety-critical industrial systems. ARTEO utilizes the Gaussian process (GP) regression to model unknown plant characteristics and enforces safety constraints using confidence intervals provided by the GP models. Due to changing process characteristics, the GP models need to be updated online by incorporating new observations and the computational complexity of model adaptation increases with a growing dataset. This work proposes an alternative ARTEO implementation by using ensemble learning, namely Ensemble-ARTEO. The Ensemble-ARTEO learns unknown plant characteristics through an ensemble of parametric regression models and calculates uncertainty by the variance of ensemble predictions. The predictive uncertainty is integrated into the optimization objective to further drive exploration. The ensemble members are updated efficiently online to capture the changing process characteristics. We demonstrate the effectiveness of our proposed Ensemble-ARTEO approach in an industrial refrigeration process. Experimental results show that our method enables tracking the desired cooling demand while satisfying the safety constraints.

In this paper, we propose a novel efficient multi-task learning formulation for the class of progression problems in which its state will continuously change over time. To use the shared knowledge information between multiple tasks to improve performance, existing multi-task learning methods mainly focus on feature selection or optimizing the task relation structure. The feature selection methods usually fail to explore the complex relationship between tasks and thus have limited performance. The methods centring on optimizing the relation structure of tasks are not capable of selecting meaningful features and have a bi-convex objective function which results in high computation complexity of the associated optimization algorithm. Unlike these multi-task learning methods, motivated by a simple and direct idea that the state of a system at the current time point should be related to all previous time points, we first propose a novel relation structure, termed adaptive global temporal relation structure (AGTS). Then we integrate the widely used sparse group Lasso, fused Lasso with AGTS to propose a novel convex multi-task learning formulation that not only performs feature selection but also adaptively captures the global temporal task relatedness. Since the existence of three non-smooth penalties, the objective function is challenging to solve. We first design an optimization algorithm based on the alternating direction method of multipliers (ADMM). Considering that the worst-case convergence rate of ADMM is only sub-linear, we then devise an efficient algorithm based on the accelerated gradient method which has the optimal convergence rate among first-order methods. We show the proximal operator of several non-smooth penalties can be solved efficiently due to the special structure of our formulation. Experimental results on four real-world datasets demonstrate that our approach not only outperforms multiple baseline MTL methods in terms of effectiveness but also has high efficiency.

The main challenge for industrial predictive models is how to effectively deal with big data from high-dimensional processes with nonstationary characteristics. Although deep networks, such as the stacked autoencoder (SAE), can learn useful features from massive data with multilevel architecture, it is difficult to adapt them online to track fast time-varying process dynamics. To integrate feature learning and online adaptation, this article proposes a deep cascade gradient radial basis function (GRBF) network for online modeling and prediction of nonlinear and nonstationary processes. The proposed deep learning method consists of three modules. First, a preliminary prediction result is generated by a GRBF weak predictor, which is further combined with raw input data for feature extraction. By incorporating the prior weak prediction information, deep output-relevant features are extracted using a SAE. Online prediction is finally produced upon the extracted features with a GRBF predictor, whose weights and structure are updated online to capture fast time-varying process characteristics. Three real-world industrial case studies demonstrate that the proposed deep cascade GRBF network outperforms existing state-of-the-art online modeling approaches as well as deep networks, in terms of both online prediction accuracy and computational complexity.

Multioutput regression of nonlinear and nonstationary data is largely understudied in both machine learning and control communities. This article develops an adaptive multioutput gradient radial basis function (MGRBF) tracker for online modeling of multioutput nonlinear and nonstationary processes. Specifically, a compact MGRBF network is first constructed with a new two-step training procedure to produce excellent predictive capacity. To improve its tracking ability in fast time-varying scenarios, an adaptive MGRBF (AMGRBF) tracker is proposed, which updates the MGRBF network structure online by replacing the worst performing node with a new node that automatically encodes the newly emerging system state and acts as a perfect local multioutput predictor for the current system state. Extensive experimental results confirm that the proposed AMGRBF tracker significantly outperforms existing state-of-the-art online multioutput regression methods as well as deep-learning-based models, in terms of adaptive modeling accuracy and online computational complexity.

In this paper, we propose a multi-task learning approach for predicting the progression of Alzheimer's disease (AD), known as the most common form of dementia. The vital challenge is to identify how the tasks are related and build learning models to capture such task relatedness. Unlike previous methods that assume low-rank structure, chase the predefined local temporal relatedness or utilize local approximation, we propose a novel penalty termed L ongitudinal S tability A djustment (LSA) to adaptively capture the intrinsic global temporal correlation among multiple time points and thus utilize the accumulated disease progression information. We combine LSA with sparse group Lasso to present a novel multi-task learning formulation to identify biomarkers closely related to cognitive measurement and predict AD progression. Two efficient algorithms are designed for large-scale dataset. Experimental results conducted on two AD data sets demonstrate our framework outperforms competing methods in terms of overall and each task performances. We also perform stability selection to identify stable biomarkers from the MRI feature set and analyze their temporal patterns in disease progression.

Precision fertilisation is the application of target variable fertilisation techniques based on soil fertility variations in specific regions. Precise fertilisation can help to balance soil nutrients, conserve fertiliser, prevent pollution, and boost crop yields. The lack of agricultural data is a key reason limiting the application of machine learning methods in agriculture. Due to the low-level network technology in farms, it is difficult to obtain diverse and complete agricultural data. The existing agricultural data is typically unstructured and difficult to mine. In this article, we extracted real-world agricultural dataset from four real farms with winter wheat and it includes different types of factors describing agriculture, such as climate, soil nutrients, crop yield information. Moreover, we present a novel multi-task learning (MTL) approach based on a tensor built of farm data to efficiently prediction both the amount and time of base fertiliser and topdressing. Specifically, real-world agricultural measurements (such as climate data, soil nutrients, etc.) are encoded into a three-dimensional tensor, and a set of interpretable temporal and spatial latent factors is extracted from the raw data through tensor decomposition. The latent factors are then utilised to train the spatio-temporal tensor prediction model. We have conducted extensive experiments utilising the real-world agricultural dataset. The experimental results show that our proposed methods have superior accuracy and stability in fertilisation prediction compared to state-of-the-art regression methods.

A key characteristic of biological systems is the ability to update the memory by learning new knowledge and removing out-of-date knowledge so that intelligent decision can be made based on the relevant knowledge acquired in the memory. Inspired by this fundamental biological principle, this article proposes a multi-output selective ensemble regression (SER) for online identification of multi-output nonlinear time-varying industrial processes. Specifically, an adaptive local learning approach is developed to automatically identify and encode a newly emerging process state by fitting a local multi-output linear model based on the multi-output hypothesis testing. This growth strategy ensures a highly diverse and independent local model set. The online modeling is constructed as a multi-output SER predictor by optimizing the combining weights of the selected local multi-output models based on a probability metric. An effective pruning strategy is also developed to remove the unwanted out-of-date local multi-output linear models in order to achieve low online computational complexity without scarifying the prediction accuracy. A simulated two-output process and two real-world identification problems are used to demonstrate the effectiveness of the proposed multi-output SER over a range of benchmark schemes for real-time identification of multi-output nonlinear and nonstationary processes, in terms of both online identification accuracy and computational complexity.

This paper proposes a multiple local model learning approach for nonlinear and nonstationary microwave heating process (MHP). The proposed local learning framework performs model adaption at two levels: (1) adaptation of the local linear model set, which adaptively partitions the process's data into multiple process states, each fitted with a local linear model; (2) online adaptation of model prediction, which selects a subset of candidate local linear models and linearly combines them to produce the model prediction. Adaptive process state partition and fitting a new local linear model to the newly emerging process state is based on statistical hypothesis testing, and the optimal combining coefficients of the selected subset linear models are obtained by minimizing the mean square error with the constraint that the sum of these coefficients is unity. A case study involving a real-world industrial MHP is used to demonstrate the superior performance of the proposed multiple local model learning approach, in terms of online modeling accuracy and computational efficiency.