Multi-task Learning Model Research Articles

Gas permeability, diffusivity, and solubility in polymers: Simulation-experiment data fusion and multi-task machine learning

Machine learning (ML) models for predicting gas permeability through polymers have traditionally relied on experimental data. While these models exhibit robustness within familiar chemical domains, reliability wanes when applied to new spaces. To address this challenge, we present a multi-tiered multi-task learning framework empowered with advanced machine-crafted polymer fingerprinting algorithms and data fusion techniques. This framework combines scarce “high-fidelity” experimental data with abundant diverse “low-fidelity” simulation or synthetic data, resulting in predictive models that display a high level of generalizability across novel chemical spaces. Additionally, this multi-task scheme capitalizes on known physics and interrelated properties, such as gas diffusivity and solubility, both of which are closely tied to permeability. By amalgamating high throughput generated simulation data with available experimental data for gas permeability, diffusivity, and solubility for various gases, we construct multi-task deep learning models. These models can simultaneously predict all three properties for all gases under consideration, with markedly enhanced predictive accuracy, particularly compared to traditional models reliant solely on experimental data for a singular property. This strategy underscores the potential of coupling high-throughput classical simulations with data fusion methodologies to yield state-of-the-art property predictors, especially when experimental data for targeted properties is scarce.

A comprehensive multi-task deep learning approach for predicting metabolic syndrome with genetic, nutritional, and clinical data

Metabolic syndrome (MetS) is a complex disorder characterized by a cluster of metabolic abnormalities, including abdominal obesity, hypertension, elevated triglycerides, reduced high-density lipoprotein cholesterol, and impaired glucose tolerance. It poses a significant public health concern, as individuals with MetS are at an increased risk of developing cardiovascular diseases and type 2 diabetes. Early and accurate identification of individuals at risk for MetS is essential. Various machine learning approaches have been employed to predict MetS, such as logistic regression, support vector machines, and several boosting techniques. However, these methods use MetS as a binary status and do not consider that MetS comprises five components. Therefore, a method that focuses on these characteristics of MetS is needed. In this study, we propose a multi-task deep learning model designed to predict MetS and its five components simultaneously. The benefit of multi-task learning is that it can manage multiple tasks with a single model, and learning related tasks may enhance the model's predictive performance. To assess the efficacy of our proposed method, we compared its performance with that of several single-task approaches, including logistic regression, support vector machine, CatBoost, LightGBM, XGBoost and one-dimensional convolutional neural network. For the construction of our multi-task deep learning model, we utilized data from the Korean Association Resource (KARE) project, which includes 352,228 single nucleotide polymorphisms (SNPs) from 7729 individuals. We also considered lifestyle, dietary, and socio-economic factors that affect chronic diseases, in addition to genomic data. By evaluating metrics such as accuracy, precision, F1-score, and the area under the receiver operating characteristic curve, we demonstrate that our multi-task learning model surpasses traditional single-task machine learning models in predicting MetS.

Acoustic emission source localization in complex pipe structure using multi-task deep learning models

Localizing defects in long-range pipelines is essential to reduce the inspection time and develop timely repair strategies. The acoustic emission (AE) method is employed to pinpoint the position of defects in pipelines. The conventional 1-D localization algorithm requires time of arrival differences between two sensors, which may not be accurately captured due to the dispersive nature of the pipe structures. The geometric variations such as elbows and welds can influence the propagating elastic waves and, consequently, arrival time. In this study, an AE source localization approach using a deep learning model is developed to tackle the influences of sensor-source distance and geometric variables. The multi-task learning model first identifies the impact of the elbow and subsequently integrates this information when predicting the source location. The proposed model is evaluated on a complex piping system, which features welded elbows in its connections. Incorporating the elbow effect into the model shows a notable improvement in overall accuracy, rising from 53% (conventional method) to 94% (proposed multi-task learning method).

MPEK: a multitask deep learning framework based on pretrained language models for enzymatic reaction kinetic parameters prediction.

Enzymatic reaction kinetics are central in analyzing enzymatic reaction mechanisms and target-enzyme optimization, and thus in biomanufacturing and other industries. The enzyme turnover number (kcat) and Michaelis constant (Km), key kinetic parameters for measuring enzyme catalytic efficiency, are crucial for analyzing enzymatic reaction mechanisms and the directed evolution of target enzymes. Experimental determination of kcat and Km is costly in terms of time, labor, and cost. To consider the intrinsic connection between kcat and Km and further improve the prediction performance, we propose a universal pretrained multitask deep learning model, MPEK, to predict these parameters simultaneously while considering pH, temperature, and organismal information. Through testing on the same kcat and Km test datasets, MPEK demonstrated superior prediction performance over the previous models. Specifically, MPEK achieved the Pearson coefficient of 0.808 for predicting kcat, improving ca. 14.6% and 7.6% compared to the DLKcat and UniKP models, and it achieved the Pearson coefficient of 0.777 for predicting Km, improving ca. 34.9% and 53.3% compared to the Kroll_model and UniKP models. More importantly, MPEK was able to reveal enzyme promiscuity and was sensitive to slight changes in the mutant enzyme sequence. In addition, in three case studies, it was shown that MPEK has the potential for assisted enzyme mining and directed evolution. To facilitate in silico evaluation of enzyme catalytic efficiency, we have established a web server implementing this model, which can be accessed at http://mathtc.nscc-tj.cn/mpek.

JointNet: Multitask Learning Framework for Denoising and Detecting Anomalies in Hyperspectral Remote Sensing

One of the significant challenges with traditional single-task learning-based anomaly detection using noisy hyperspectral images (HSIs) is the loss of anomaly targets during denoising, especially when the noise and anomaly targets are similar. This issue significantly affects the detection accuracy. To address this problem, this paper proposes a multitask learning (MTL)-based method for detecting anomalies in noisy HSIs. Firstly, a preliminary detection approach based on the JointNet model, which decomposes the noisy HSI into a pure background and a noise–anomaly target mixing component, is introduced. This approach integrates the minimum noise fraction rotation (MNF) algorithm into an autoencoder (AE), effectively isolating the noise while retaining critical features for anomaly detection. Building upon this, the JointNet model is further optimized to ensure that the noise information is shared between the denoising and anomaly detection subtasks, preserving the integrity of the training data during the anomaly detection process and resolving the issue of losing anomaly targets during denoising. A novel loss function is designed to enable the joint learning of both subtasks under the multitask learning model. In addition, a noise score evaluation metric is introduced to calculate the probability of a pixel being an anomaly target, allowing for a clear distinction between noise and anomaly targets, thus providing the final anomaly detection results. The effectiveness of the proposed model and method is validated via testing on the HYDICE and San Diego datasets. The denoising metric results of the PSNR, SSIM, and SAM are 41.79, 0.91, and 4.350 and 42.83, 0.93, and 3.558 on the HYDICE and San Diego datasets, respectively. The anomaly detection ACU is 0.943 and 0.959, respectively. The proposed method outperforms the other algorithms, demonstrating that the reconstructed images using this method exhibited lower noise levels and more complete image information, and the JointNet model outperforms the mainstream HSI anomaly detection algorithms in both the quantitative evaluation and visual effect, showcasing its improved detection capabilities.

Heterogenous biological network multi-task learning model for ncRNA-disease-drug association prediction

Traditional medical research is characterised by lengthy duration, significant financial investment, and substantial risk of failure. In response to these challenges, network medicine, combined with medicine and computer technology, has become an important development direction, and computational methods have been proposed to predict potential associations. However, most of the current computational methods focus on single-potential association prediction tasks, which face issues of association sparsity and weak generalisation ability. To address these challenges, we developed a heterogeneous biological network multi-task learning model (HBNMM). Unlike previous methods based on bipartite graphs, HBNMM constructs a complex heterogeneous biological network, including ncRNA-disease-drug association networks and diverse similarity networks. HBNMM applies graph attention networks to aggregate node neighbourhood information and acquire node feature embeddings, and is then trained with a multi-task learning strategy to simultaneously predict potential ncRNA-disease, ncRNA-drug, and drug-disease associations. As a result, the HBNMM achieves an excellent performance that is higher than that of the state-of-the-art models. Furthermore, five case studies supported by experiments showed powerful predictive ability for drug discovery and disease treatment.

Machine-learning assisted analysis on coupled fluid-dynamics and electrochemical processes in interdigitated channel for iron-chromium flow batteries

This study explores the impact of interdigitated flow channel spacing on electrolyte distribution and flow velocity in porous electrodes, thereby affecting pump consumption and system efficiency. Utilizing a 3D electrochemical flow-coupled model, the research investigates the electrochemical performance and energy efficiency across various channel spacings, specific flow rates, and current densities. It elucidates the mechanisms by which interdigitated channel spacing influences battery performance. Through controlled trial-and-error, the optimal spacing for flow channels in Iron-Chromium Redox Flow Batteries (ICRFBs) was determined to be 4 mm. At a current density of 140 mA/cm2, the voltage efficiency reached 86.3 %, and the pump-based voltage efficiency achieved 85.9 %. To quickly and efficiently explore the impact of each individual parameter on battery efficiency and identify the optimal operating conditions, this study further developed a multi-task machine learning (ML) model. Initially trained on simulation data, the model achieved high predictive accuracy (R2 > 0.88) and effectively linked the interdigitated flow field design of ICRFBs with current density, specific flow rate, and channel spacing. The ML model was then validated through further investigation to ensure its accuracy and to achieve optimum performance efficiency as recommended by the model. This integrated approach offers a comprehensive understanding of RFBs performance under varying conditions, driving advancements in RFBs technology.

Radiomic analysis reveals diverse prognostic and molecular insights into the response of breast cancer to neoadjuvant chemotherapy: a multicohort study

BackgroundBreast cancer patients exhibit various response patterns to neoadjuvant chemotherapy (NAC). However, it is uncertain whether diverse tumor response patterns to NAC in breast cancer patients can predict survival outcomes. We aimed to develop and validate radiomic signatures indicative of tumor shrinkage and therapeutic response for improved survival analysis.MethodsThis retrospective, multicohort study included three datasets. The development dataset, consisting of preoperative and early NAC DCE-MRI data from 255 patients, was used to create an imaging signature-based multitask model for predicting tumor shrinkage patterns and pathological complete response (pCR). Patients were categorized as pCR, nonpCR with concentric shrinkage (CS), or nonpCR with non-CS, with prediction performance measured by the area under the curve (AUC). The prognostic validation dataset (n = 174) was used to assess the prognostic value of the imaging signatures for overall survival (OS) and recurrence-free survival (RFS) using a multivariate Cox model. The gene expression data (genomic validation dataset, n = 112) were analyzed to determine the biological basis of the response patterns.ResultsThe multitask learning model, utilizing 17 radiomic signatures, achieved AUCs of 0.886 for predicting tumor shrinkage and 0.760 for predicting pCR. Patients who achieved pCR had the best survival outcomes, while nonpCR patients with a CS pattern had better survival than non-CS patients did, with significant differences in OS and RFS (p = 0.00012 and p = 0.00063, respectively). Gene expression analysis highlighted the involvement of the IL-17 and estrogen signaling pathways in response variability.ConclusionsRadiomic signatures effectively predict NAC response patterns in breast cancer patients and are associated with specific survival outcomes. The CS pattern in nonpCR patients indicates better survival.

Multi-Task ADME/PK prediction at industrial scale: leveraging large and diverse experimental datasets.

ADME (Absorption, Distribution, Metabolism, Excretion) properties are key parameters to judge whether a drug candidate exhibits a desired pharmacokinetic (PK) profile. In this study, we tested multi-task machine learning (ML) models to predict ADME and animal PK endpoints trained on in-house data generated at Boehringer Ingelheim. Models were evaluated both at the design stage of a compound (i. e., no experimental data of test compounds available) and at testing stage when a particular assay would be conducted (i. e., experimental data of earlier conducted assays may be available). Using realistic time-splits, we found a clear benefit in performance of multi-task graph-based neural network models over single-task model, which was even stronger when experimental data of earlier assays is available. In an attempt to explain the success of multi-task models, we found that especially endpoints with the largest numbers of data points (physicochemical endpoints, clearance in microsomes) are responsible for increased predictivity in more complex ADME and PK endpoints. In summary, our study provides insight into how data for multiple ADME/PK endpoints in a pharmaceutical company can be best leveraged to optimize predictivity of ML models.

GAN-Based Multi-Task Learning Approach for Prognostics and Health Management of IIoT

Health assessment (HA) & remaining useful life (RUL) estimation, the two essential pillars of the prognostics and health management (PHM) paradigm, help improve industrial equipment reliability while reducing maintenance costs. However, reported works treat HA and RUL estimation as disjoint problems though there exist unexploited similarities among these related issues. Additionally, in practical industrial working scenarios, equipment(s) stay in a normal condition for most of its lifespan, leading to a disproportionate training dataset, hampering the prediction accuracy. To overcome the above problems, we propose a data-driven multi-task learning framework aided by a novel least squares recurrent auxiliary classifier generative adversarial network (LS-R-ACGAN). LS-R-ACGAN employs recurrent neural networks (RNNs) in its generator & discriminator networks for multi-variate fault data generation while overcoming the vanishing gradient problem of ACGANs. Post-data-augmentation, a balanced training dataset, trains a multi-task learning model based on a deep gated RNN (DGRU) for joint HA and RUL estimation. Our simulations use the C-MAPSS dataset for testing the proposed approach’s accuracy. The final results showcase improvements by at least <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\sim$</tex-math> </inline-formula> 3.54% and <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\sim$</tex-math> </inline-formula> 2.38% on the RMSE and Score metric over existing state-of-the-art works suggesting its competitiveness and competence for real-world implementations. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Note to Practitioners</i> —This study attacks multiple presumptions concerning the maintenance of complex systems including, ample availability of fault samples alongside normal samples and solving only one of the problems, i.e., health assessment (HA) or remaining useful life (RUL) estimations rather than both. Popular artificial intelligence (AI) based solutions rely heavily upon the quantity of training data for desired performance outcomes. However, in real-life industrial scenarios, fault instances are rare, inherently resulting in an imbalanced dataset. Additionally, it is both labor-intensive and costly to repeat multiple experiments for gathering required data as the equipment(s) pass through short-lived fault states. Thus conventional AI-based solutions may end up producing biased predictions. This work proposes to solve the imbalanced training dataset problem along with joint HA and RUL estimations, in one solution pipeline, currently unexplored in majority of the existing works in the literature. LS-R-ACGAN generates ample diverse fault samples for reducing the degree of imbalance between normal and fault classes. Subsequently, after data augmentation, the balanced training dataset helps train a multi-task learning model for joint HA and RUL estimations. The proposed framework has been implemented and tested on a benchmark dataset proving its superiority over multiple existing methods in the literature.

End-to-End Multitask Learning With Vision Transformer.

Multitask learning (MTL) is a challenging puzzle, particularly in the realm of computer vision (CV). Setting up vanilla deep MTL requires either hard or soft parameter sharing schemes that employ greedy search to find the optimal network designs. Despite its widespread application, the performance of MTL models is vulnerable to under-constrained parameters. In this article, we draw on the recent success of vision transformer (ViT) to propose a multitask representation learning method called multitask ViT (MTViT), which proposes a multiple branch transformer to sequentially process the image patches (i.e., tokens in transformer) that are associated with various tasks. Through the proposed cross-task attention (CA) module, a task token from each task branch is regarded as a query for exchanging information with other task branches. In contrast to prior models, our proposed method extracts intrinsic features using the built-in self-attention mechanism of the ViT and requires just linear time on memory and computation complexity, rather than quadratic time. Comprehensive experiments are carried out on two benchmark datasets, including NYU-Depth V2 (NYUDv2) and CityScapes, after which it is found that our proposed MTViT outperforms or is on par with existing convolutional neural network (CNN)-based MTL methods. In addition, we apply our method to a synthetic dataset in which task relatedness is controlled. Surprisingly, experimental results reveal that the MTViT exhibits excellent performance when tasks are less related.

Multi-task deep learning model for quantitative volatile organic compounds analysis by feature fusion of electronic nose sensing

In exploring pattern recognition for electronic noses via deep neural networks, traditional networks encounter key challenges, such as low training efficiency, and neglect of spatial-temporal attributes of gas sensor response sequences. In this study, an unmanned gas-sensing test system is used to generate a large dataset to ensure robust model training. The Gramian angular field-Markov transition field is utilized to convert time sequences into images. Using advanced image processing tools, the images are then subsequently compressed with data augmentation. This method fully preserves temporal and spatial features within the sequences, thus enhancing model performance. Furthermore, the proposed multi-task learning (MTL) framework competently performs simultaneous classification and regression tasks. The primary component of the MTL network, majorly constituted of convolutional neural networks, emphasizes the spatial features of the sequences. The integration of a long short-term memory layer ensures the preservation of temporal feature analysis of the input data, thereby enhancing predictive performance. When images are compressed to only 3.9 % of the original data, substantial information can still be preserved. Subsequently, the model trained by such compressed images attains an accuracy of 95.31 % and an R2 score of 0.9510 for classification and regression tasks, respectively. Our work reveals the remarkable potential of integrating temporal and spatial features in pattern recognition, promoting the potency of multi-tasking deep learning networks in electronic nose technology.

ARGENT: Multi-task learning model for predicting autism-related genes and drug targets using heterogeneous graph convolutional network

Autism Spectrum Disorder (ASD) is a multifactorial-driven neurodevelopmental disorder, the pathophysiological mechanisms of which remain largely elusive, significantly hampering the development of effective therapeutic strategies. MicroRNAs (miRNAs) are emerging as critical regulators in the molecular etiology of autism, influencing gene expression by interacting with target mRNAs involved in neural development and synaptic function. To introduce miRNA-gene regulatory information to identify potential autism-related genes and predict drug targets, we introduce a multitask learning approach named ARGENT based on the heterogeneous graph convolutional network (HGCN). The heterogeneous graph in ARGENT includes nodes representing genes, miRNAs, drugs and integrates the regulatory relationships between miRNA-gene, gene-gene, and the associations between genes and drugs. Using the HGCN, the proposed model is able to learn embeddings for these nodes by aggregating information from different adjacent node types, enabling the prediction of autism-related genes and their potential drug candidates. The efficacy of ARGENT is substantiated through multiple validation experiments, demonstrating that it not only identifies known autism-related genes but also reveals novel candidate genes and potential drug targets. Additionally, KEGG and GO analysis results of known genes related to ASD and genes predicted by ARGENT as potentially related to ASD reveal that these genes are significantly enriched in pathways associated with neural signal transmission. This suggests that these genes may play crucial roles in the core functions and structure of neural circuits, potentially influencing the neurodevelopmental aspects of ASD, providing new insights for prevention and drug repositioning.

Explainable deep learning for disease activity prediction in chronic inflammatory joint diseases

Analysing complex diseases such as chronic inflammatory joint diseases (CIJDs), where many factors influence the disease evolution over time, is a challenging task. CIJDs are rheumatic diseases that cause the immune system to attack healthy organs, mainly the joints. Different environmental, genetic and demographic factors affect disease development and progression. The Swiss Clinical Quality Management in Rheumatic Diseases (SCQM) Foundation maintains a national database of CIJDs documenting the disease management over time for 19’267 patients. We propose the Disease Activity Score Network (DAS-Net), an explainable multi-task learning model trained on patients’ data with different arthritis subtypes, transforming longitudinal patient journeys into comparable representations and predicting multiple disease activity scores. First, we built a modular model composed of feed-forward neural networks, long short-term memory networks and attention layers to process the heterogeneous patient histories and predict future disease activity. Second, we investigated the utility of the model’s computed patient representations (latent embeddings) to identify patients with similar disease progression. Third, we enhanced the explainability of our model by analysing the impact of different patient characteristics on disease progression and contrasted our model outcomes with medical expert knowledge. To this end, we explored multiple feature attribution methods including SHAP, attention attribution and feature weighting using case-based similarity. Our model outperforms temporal and non-temporal neural network, tree-based, and naive static baselines in predicting future disease activity scores. To identify similar patients, a k-nearest neighbours regression algorithm applied to the model’s computed latent representations outperforms baseline strategies that use raw input features representation.

MTL-TRANSFER: Leveraging Multi-task Learning and Transferred Knowledge for Improving Fault Localization and Program Repair

Fault localization (FL) and automated program repair (APR) are two main tasks of automatic software debugging. Compared with traditional methods, deep learning-based approaches have been demonstrated to achieve better performance in FL and APR tasks. However, the existing deep learning-based FL methods ignore the deep semantic features or only consider simple code representations. And for APR tasks, existing template-based APR methods are weak in selecting the correct fix templates for more effective program repair, which are also not able to synthesize patches via the embedded end-to-end code modification knowledge obtained by training models on large-scale bug-fix code pairs. Moreover, in most of FL and APR methods, the model designs and training phases are performed separately, leading to ineffective sharing of updated parameters and extracted knowledge during the training process. This limitation hinders the further improvement in the performance of FL and APR tasks. To solve the above problems, we propose a novel approach called MTL-TRANSFER, which leverages a multi-task learning strategy to extract deep semantic features and transferred knowledge from different perspectives. First, we construct a large-scale open-source bug datasets and implement 11 multi-task learning models for bug detection and patch generation sub-tasks on 11 commonly used bug types, as well as one multi-classifier to learn the relevant semantics for the subsequent fix template selection task. Second, an MLP-based ranking model is leveraged to fuse spectrum-based, mutation-based and semantic-based features to generate a sorted list of suspicious statements. Third, we combine the patches generated by the neural patch generation sub-task from the multi-task learning strategy with the optimized fix template selecting order gained from the multi-classifier mentioned above. Finally, the more accurate FL results, the optimized fix template selecting order, and the expanded patch candidates are combined together to further enhance the overall performance of APR tasks. Our extensive experiments on widely-used benchmark Defects4J show that MTL-TRANSFER outperforms all baselines in FL and APR tasks, proving the effectiveness of our approach. Compared with our previously proposed FL method TRANSFER-FL (which is also the state-of-the-art statement-level FL method), MTL-TRANSFER increases the faults hit by 8/11/12 on Top-1/3/5 metrics (92/159/183 in total). And on APR tasks, the number of successfully repaired bugs of MTL-TRANSFER under the perfect localization setting reaches 75, which is 8 more than our previous APR method TRANSFER-PR. Furthermore, another experiment to simulate the actual repair scenarios shows that MTL-TRANSFER can successfully repair 15 and 9 more bugs (56 in total) compared with TBar and TRANSFER, which demonstrates the effectiveness of the combination of our optimized FL and APR components.

Predicting changes in brain metabolism and progression from mild cognitive impairment to dementia using multitask Deep Learning models and explainable AI

Background:The prediction of Alzheimer’s disease (AD) progression from its early stages is a research priority. In this context, the use of Artificial Intelligence (AI) in AD has experienced a notable surge in recent years. However, existing investigations predominantly concentrate on distinguishing clinical phenotypes through cross-sectional approaches. This study aims to investigate the potential of modeling additional dimensions of the disease, such as variations in brain metabolism assessed via [18F]-fluorodeoxyglucose positron emission tomography (FDG-PET), and utilize this information to identify patients with mild cognitive impairment (MCI) who will progress to dementia (pMCI). Methods:We analyzed data from 1,617 participants from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) who had undergone at least one FDG-PET scan. We identified the brain regions with the most significant hypometabolism in AD and used Deep Learning (DL) models to predict future changes in brain metabolism. The best-performing model was then adapted under a multi-task learning framework to identify pMCI individuals. Finally, this model underwent further analysis using eXplainable AI (XAI) techniques. Results:Our results confirm a strong association between hypometabolism, disease progression, and cognitive decline. Furthermore, we demonstrated that integrating data on changes in brain metabolism during training enhanced the models’ ability to detect pMCI individuals (sensitivity=88.4%, specificity=86.9%). Lastly, the application of XAI techniques enabled us to delve into the brain regions with the most significant impact on model predictions, highlighting the importance of the hippocampus, cingulate cortex, and some subcortical structures. Conclusion:This study introduces a novel dimension to predictive modeling in AD, emphasizing the importance of projecting variations in brain metabolism under a multi-task learning paradigm.

ROCOv2: Radiology Objects in COntext Version 2, an Updated Multimodal Image Dataset

Automated medical image analysis systems often require large amounts of training data with high quality labels, which are difficult and time consuming to generate. This paper introduces Radiology Object in COntext version 2 (ROCOv2), a multimodal dataset consisting of radiological images and associated medical concepts and captions extracted from the PMC Open Access subset. It is an updated version of the ROCO dataset published in 2018, and adds 35,705 new images added to PMC since 2018. It further provides manually curated concepts for imaging modalities with additional anatomical and directional concepts for X-rays. The dataset consists of 79,789 images and has been used, with minor modifications, in the concept detection and caption prediction tasks of ImageCLEFmedical Caption 2023. The dataset is suitable for training image annotation models based on image-caption pairs, or for multi-label image classification using Unified Medical Language System (UMLS) concepts provided with each image. In addition, it can serve for pre-training of medical domain models, and evaluation of deep learning models for multi-task learning.

A Multi-Task Joint Learning Model Based on Transformer and Customized Gate Control for Predicting Remaining Useful Life and Health Status of Tools.

Previous studies have primarily focused on predicting the remaining useful life (RUL) of tools as an independent process. However, the RUL of a tool is closely related to its wear stage. In light of this, a multi-task joint learning model based on a transformer encoder and customized gate control (TECGC) is proposed for simultaneous prediction of tool RUL and tool wear stages. Specifically, the transformer encoder is employed as the backbone of the TECGC model for extracting shared features from the original data. The customized gate control (CGC) is utilized to extract task-specific features relevant to tool RUL prediction and tool wear stage and shared features. Finally, by integrating these components, the tool RUL and the tool wear stage can be predicted simultaneously by the TECGC model. In addition, a dynamic adaptive multi-task learning loss function is proposed for the model's training to enhance its calculation efficiency. This approach avoids unsatisfactory prediction performance of the model caused by unreasonable selection of trade-off parameters of the loss function. The effectiveness of the TECGC model is evaluated using the PHM2010 dataset. The results demonstrate its capability to accurately predict tool RUL and tool wear stages.

Objective and neutral summarization of customer reviews

Opinion mining aims to detect and extract relevant information from large quantity of customer reviews. Automatic opinion summarization then seeks to create a consensual point of view often oriented toward the main sentiment of clients to render their experience. Although factual information is valuable for companies to understand what works or not in their products, summarization approaches that convey objectivity, and constructive feedback from customer reviews have yet to be explored. We propose an adversarial multi-task learning model for document summarization to address this new issue. Our algorithm combines an autoencoder for document summarization with a gradient reversal layer to learn independent representations of subjective and sentiment-based material. We assess and compare our method on the Amazon product review dataset where we introduce an original evaluation dataset for objective summarization. We further completed the analysis with neutrality and objectivity metrics. This study demonstrates that the generated summaries carry out relevant and objective content but also emphasize the importance of various processes and layers in multi-task learners to control the information effectively.

TMO-Net: an explainable pretrained multi-omics model for multi-task learning in oncology

Cancer is a complex disease composing systemic alterations in multiple scales. In this study, we develop the Tumor Multi-Omics pre-trained Network (TMO-Net) that integrates multi-omics pan-cancer datasets for model pre-training, facilitating cross-omics interactions and enabling joint representation learning and incomplete omics inference. This model enhances multi-omics sample representation and empowers various downstream oncology tasks with incomplete multi-omics datasets. By employing interpretable learning, we characterize the contributions of distinct omics features to clinical outcomes. The TMO-Net model serves as a versatile framework for cross-modal multi-omics learning in oncology, paving the way for tumor omics-specific foundation models.