Applying Transfer Learning Research Articles

Transfer learning of species co-occurrence patterns between plant communities

AimThe use of neural networks (NNs) is spreading to all areas of life, and Ecology is no exception. However, the data-hungry nature of NNs can leave out many small, valuable datasets. Here we show how to apply transfer learning to rescue small datasets that can be invaluable in understanding patterns of species co-occurrence. LocationSemiarid plant communities in Spain and México. Time period2016–2022. Major taxa studiedAngiosperms. MethodsBased on a large sample of plant species co-occurrence in vegetation patches in a semi-arid area of eastern Spain, we fit a generative artificial intelligence (AI) model that correctly reproduces which species live with which in these patches. Subsequently, we train the same type of model on two communities for which we only have smaller datasets (another semi-arid community in eastern Spain, and a tropical community in Mexico). ResultsWhen we transfer the knowledge learnt from the large dataset directly to the other two, the predictions improve for the community more similar to our reference one. As for the more dissimilar community, improving the accuracy of the transfer requires a further tuning of the model to the local data. In particular, the knowledge transferred relates primarily to species frequency and, to a lesser extent, to their phylogenetic relationships, which are known to be determinants of species interaction patterns. Main conclusionsThis AI-based approach can be performed for communities similar or not so similar to the reference community, opening the door to systematic transfer learning for accurate predictions on small datasets. Interestingly, this transfer operates by matching unrelated species between the origin and target datasets, implying that arbitrary datasets can then be transferred to, or even combined in order to augment each other, irrespective of the species involved, potentially allowing such models to be applied to a wide range of plant communities in different climates.

465 Comparative analysis of semantic segmentation and deep regression models with supervised pre-training for accurate prediction of pig body weight from video data: Insights from industry-scale datasets

Abstract Accurate pig body weight (BW) measurement is essential for producers as it is related to pig growth, health, and marketing, yet conventional manual weighing methods are time-consuming and may cause potential stress to the animals. Although there is a growing trend towards the adoption of three-dimensional cameras coupled with computer vision techniques for pig BW estimation, their validation using industry-scale data are still limited. Among the prevailing methodologies, semantic segmentation and supervised pre-training regression are prominent models currently used. Therefore, the objectives of this study were: 1) to estimate pig BW from repeatedly measured video data obtained from a commercial setting, 2) to compare the performance of the two image analysis methods: thresholding segmentation and deep regression approaches, and 3) to evaluate the predictive ability of BW estimation models. An Intel RealSense D435 camera was installed in the commercial farm to collect top-view videos of 540 pigs biweekly at five different time points over three months. At the same time, manually measured BW records were collected using a digital weighing system. We used an automated video conversion pipeline and fine-tuned YOLOv8 to pre-process the raw depth videos. Subsequently, we acquired a total of 151,756 depth images and depth map files. Adaptive thresholding was applied to segment the pig body from the background. Four image-derived biometric features, including dorsal length, abdominal width, height, and volume, were estimated from the segmented images and fitted using ordinary least squares and random forest models. We applied transfer learning by initializing the weights of five deep learning models, including ResNet50, Xception, EfficientNetV2S, ConvNeXtBase, and Vision Transformer, with pre-trained weights from ImageNet. We then fine-tuned these models on the pig depth images. The last layer of each model was adapted to linear regression, enabling direct estimation of BW for each image without the need for additional image pre-processing steps. We employed random repeated subsampling cross-validation, dividing 80% of the pigs for training and 20% for testing, to evaluate prediction performance at each time point. The best prediction coefficients of determination and mean absolute percentage error for each time point were 0.76, 0.86, 0.90, 0.83, 0.90, and 4.57%, 3.76%, 3.01%, 3.41%, 4.84%, respectively. On average, the Xception model resulted in the best prediction coefficient of determination and mean absolute percentage error of 0.90 and 3.01%. Our results suggest that deep learning-based supervised learning models improve the prediction performance of pig BW from industry-scale depth video data.

Transferable and data efficient metamodeling of storm water system nodal depths using auto-regressive graph neural networks

Storm water systems (SWSs) are essential infrastructure providing multiple services including environmental protection and flood prevention. Typically, utility companies rely on computer simulators to properly design, operate, and manage SWSs. However, multiple applications in SWSs are highly time-consuming. Researchers have resorted to cheaper-to-run models, i.e. metamodels, as alternatives of computationally expensive models. With the recent surge in artificial intelligence applications, machine learning has become a key approach for metamodelling urban water networks. Specifically, deep learning methods, such as feed-forward neural networks, have gained importance in this context. However, these methods require generating a sufficiently large database of examples and training their internal parameters. Both processes defeat the purpose of using a metamodel, i.e., saving time. To overcome this issue, this research focuses on the application of inductive biases and transfer learning for creating SWS metamodels which require less data and retain high performance when used elsewhere. In particular, this study proposes an auto-regressive graph neural network metamodel of the Storm Water Management Model (SWMM) from the Environmental Protection Agency (EPA) for estimating hydraulic heads. The results indicate that the proposed metamodel requires a smaller number of examples to reach high accuracy and speed-up, in comparison to fully connected neural networks. Furthermore, the metamodel shows transferability as it can be used to predict hydraulic heads with high accuracy on unseen parts of the network. This work presents a novel approach that benefits both urban drainage practitioners and water network modeling researchers. The proposed metamodel can help practitioners on the planning, operation, and maintenance of their systems by offering an efficient metamodel of SWMM for computationally intensive tasks like optimization and Monte Carlo analyses. Researchers can leverage the current metamodel’s structure for developing new surrogate model architectures tailored to their specific needs or start paving the way for more general foundation metamodels of urban drainage systems.

Enhancing Domain Diversity of Transfer Learning-Based SSVEP-BCIs by the Reconstruction of Channel Correlation.

The application of transfer learning, specifically pre-training and fine-tuning, in steady-state visual evoked potential (SSVEP)-based brain-computer interfaces (BCIs) has been demonstrated to effectively improve the classification performance of deep learning methods with limited calibration data. However, effectively learning task-related knowledge from source domains during the pre-training phase remains challenging. To address this issue, this study proposes an effective data augmentation method called Reconstruction of Channel Correlation (RCC) to optimize the utilization of the source domain data. Concretely, RCC reconstructs training samples using probabilistically mixed eigenvector matrices derived from covariance matrices across source domains. This process manipulates the channel correlation of training samples, implicitly creating novel synthesized domains. By increasing the diversity of source domains, RCC aims to enhance the domain generalizability of the pre-trained model. The effectiveness of RCC is validated through subject-independent and subject-adaptive classification experiments. The results of subject-independent classification demonstrate that RCC significantly improves the classification performance of the pre-trained model on unseen target subjects. Moreover, when compared to the fine-tuning process using the RCC-absent pre-trained model, the fine-tuning process using the RCC-enhanced pre-trained model yields significantly improved performance in the subject-adaptive classification. RCC proves to enhance the performance of transfer learning by optimizing the utilization of the source domain data. The RCC-enhanced transfer learning has the potential to facilitate the practical implementation of SSVEP-BCIs in real-world scenarios.

Improving 18F-FDG PET Quantification Through a Spatial Normalization Method.

Quantification of 18F-FDG PET images is useful for accurate diagnosis and evaluation of various brain diseases, including brain tumors, epilepsy, dementia, and Parkinson disease. However, accurate quantification of 18F-FDG PET images requires matched 3-dimensional T1 MRI scans of the same individuals to provide detailed information on brain anatomy. In this paper, we propose a transfer learning approach to adapt a pretrained deep neural network model from amyloid PET to spatially normalize 18F-FDG PET images without the need for 3-dimensional MRI. Methods: The proposed method is based on a deep learning model for automatic spatial normalization of 18F-FDG brain PET images, which was developed by fine-tuning a pretrained model for amyloid PET using only 103 18F-FDG PET and MR images. After training, the algorithm was tested on 65 internal and 78 external test sets. All T1 MR images with a 1-mm isotropic voxel size were processed with FreeSurfer software to provide cortical segmentation maps used to extract a ground-truth regional SUV ratio using cerebellar gray matter as a reference region. These values were compared with those from spatial normalization-based quantification methods using the proposed method and statistical parametric mapping software. Results: The proposed method showed superior spatial normalization compared with statistical parametric mapping, as evidenced by increased normalized mutual information and better size and shape matching in PET images. Quantitative evaluation revealed a consistently higher SUV ratio correlation and intraclass correlation coefficients for the proposed method across various brain regions in both internal and external datasets. The remarkably good correlation and intraclass correlation coefficient values of the proposed method for the external dataset are noteworthy, considering the dataset's different ethnic distribution and the use of different PET scanners and image reconstruction algorithms. Conclusion: This study successfully applied transfer learning to a deep neural network for 18F-FDG PET spatial normalization, demonstrating its resource efficiency and improved performance. This highlights the efficacy of transfer learning, which requires a smaller number of datasets than does the original network training, thus increasing the potential for broader use of deep learning-based brain PET spatial normalization techniques for various clinical and research radiotracers.

Stream salinity prediction in data-scarce regions: Application of transfer learning and uncertainty quantification

Scarcity of stream salinity data poses a challenge to understanding salinity dynamics and its implications for water supply management in water-scarce salt-prone regions around the world. This paper introduces a framework for generating continuous daily stream salinity estimates using instance-based transfer learning (TL) and assessing the reliability of the synthetic salinity data through uncertainty quantification via prediction intervals (PIs). The framework was developed using two temporally distinct specific conductance (SC) datasets from the Upper Red River Basin (URRB) located in southwestern Oklahoma and Texas Panhandle, United States. The instance-based TL approach was implemented by calibrating Feedforward Neural Networks (FFNNs) on a source SC dataset of around 1200 instantaneous grab samples collected by United States Geological Survey (USGS) from 1959 to 1993. The trained FFNNs were subsequently tested on a target dataset (1998-present) of 220 instantaneous grab samples collected by the Oklahoma Water Resources Board (OWRB). The framework's generalizability was assessed in the data-rich Bird Creek watershed in Oklahoma by manipulating continuous SC data to simulate data-scarce conditions for training the models and using the complete Bird Creek dataset for model evaluation. The Lower Upper Bound Estimation (LUBE) method was used with FFNNs to estimate PIs for uncertainty quantification. Autoregressive SC prediction methods via FFNN were found to be reliable with Nash Sutcliffe Efficiency (NSE) values of 0.65 and 0.45 on in-sample and out-of-sample test data, respectively. The same modeling scenario resulted in an NSE of 0.54 for the Bird Creek data using a similar missing data ratio, whereas a higher ratio of observed data increased the accuracy (NSE = 0.84). The relatively narrow estimated PIs for the North Fork Red River in the URRB indicated satisfactory stream salinity predictions, showing an average width equivalent to 25 % of the observed range and a confidence level of 70 %.

Physics-constrained deep learning approach for solving inverse problems in composite laminated plates

The applications of physics-informed neural networks (PINNs) in material parameters identification of composite laminates are currently research highlights. We present an efficient physics-constrained deep learning approach based on PINNs for solving material parameters of composite laminates. We explain the details of incorporating first-order shear deformation theory as physical information into designed loss functions. The performance of proposed approach is demonstrated through a numerical example of idealized composite laminated plates under uniform pressure conditions. The computational errors of material parameters are less than 0.1%. Moreover, the applicability of transfer learning (TL) has been illustrated for inversion of material parameters with few datasets. The ambition is to improve the identification efficiency of carbon and glass fiber reinforced plies material parameters across various loading conditions. Intensive studies are conducted to compare convergence time between with and without TL through a numerical example under thermal-force coupling conditions. The comparison shows PINNs with TL obtain average error of 0.144% and 0.077% in carbon and glass fiber reinforced plies, which is better than without TL of 0.472% and 0.284%, respectively. The proposed work emphasizes the effective application of PINNs-based approach that combines first-order deformation and data-driven solution features for inverse solving material parameters of composite laminates.

Development and Validation of a Real-Time Service Model for Noise Removal and Arrhythmia Classification Using Electrocardiogram Signals.

Arrhythmias range from mild nuisances to potentially fatal conditions, detectable through electrocardiograms (ECGs). With advancements in wearable technology, ECGs can now be monitored on-the-go, although these devices often capture noisy data, complicating accurate arrhythmia detection. This study aims to create a new deep learning model that utilizes generative adversarial networks (GANs) for effective noise removal and ResNet for precise arrhythmia classification from wearable ECG data. We developed a deep learning model that cleans ECG measurements from wearable devices and detects arrhythmias using refined data. We pretrained our model using the MIT-BIH Arrhythmia and Noise databases. Least squares GANs were used for noise reduction, maintaining the integrity of the original ECG signal, while a residual network classified the type of arrhythmia. After initial training, we applied transfer learning with actual ECG data. Our noise removal model significantly enhanced data clarity, achieving over 30 dB in a signal-to-noise ratio. The arrhythmia detection model was highly accurate, with an F1-score of 99.10% for noise-free data. The developed model is capable of real-time, accurate arrhythmia detection using wearable ECG devices, allowing for immediate patient notification and facilitating timely medical response.

Applying transfer learning to address data scarcity: A case study on LWD gamma ray depth-lag remediation from Volve to another gasfield

In petroleum engineering, obtaining comprehensive subsurface data is challenging due to significant costs and protective regulations. This complexity often restricts the extensive use of data-driven approaches. This study investigates the applicability of deep learning with transfer learning, to tackle this prominent challenges. Grounded in a practical fusion of engineering and geology, we address a specific issue: the depth-lagged measurements of the Logging While Drilling (LWD) tool's Gamma Ray detector, owing to its position meters behind the drill bit. Our approach uniquely integrates real-time drilling engineering parameters with the delayed Gamma time-series features, aiming to predictively correct these lagged measurements.The research extends from the Volve oil field in Norway to a distinct gas field in China, covering a wide geographical range and diverse geological conditions. The results indicated that in environments with limited data, transfer learning could enhance accuracy by up to 13% compared to traditional methods. This transition tests transfer learning's adaptability to various gasfield settings, demonstrating its potential to enhance model accuracy in data-limited environments and advocating for data-driven approaches in petroleum engineering.

Transfer learning may explain pigeons’ ability to detect cancer in histopathology

Pigeons' unexpected competence in learning to categorize unseen histopathological images has remained an unexplained discovery for almost a decade (Levensonet al2015PLoS One10e0141357). Could it be that knowledge transferred from their bird's-eye views of the earth's surface gleaned during flight contributes to this ability? Employing a simulation-based verification strategy, we recapitulate this biological phenomenon with a machine-learning analog. We model pigeons' visual experience during flight with the self-supervised pre-training of a deep neural network on BirdsEyeViewNet; our large-scale aerial imagery dataset. As an analog of the differential food reinforcement performed in Levensonet al's study 2015PLoS One10e0141357), we apply transfer learning from this pre-trained model to the same Hematoxylin and Eosin (H&E) histopathology and radiology images and tasks that the pigeons were trained and tested on. The study demonstrates that pre-training neural networks with bird's-eye view data results in close agreement with pigeons' performance. These results support transfer learning as a reasonable computational model of pigeon representation learning. This is further validated with six large-scale downstream classification tasks using H&E stained whole slide image datasets representing diverse cancer types.

Herding behavior in NFT Auction: The role of visual complexity and familiarity

Despite the growing popularity of non-fungible tokens (NFTs) in the market, they still lack established valuation guidelines. Invoking herding and auction theories, this study investigates how NFT collectors’ bidding tendencies are influenced by other collectors’ bidding decisions (herding behavior). This study also examines how NFT visual characteristics, such as complexity and familiarity, moderate herding behavior. We use a computer vision technique to quantify the visual complexity of an NFT and apply transfer learning to capture its visual familiarity. Examining data from a widely used NFT platform, we find evidence of herding behavior in such auctions. Furthermore, our results show that there are nonlinear moderating effects of visual characteristics on herding behavior. Specifically, we find that consumers exhibit a stronger herding tendency at the two extremes of complexity distribution and at the lower extreme of familiarity distribution. Further analysis of the NFT resale market suggests that auction winners consider herding in the primary sale of an NFT as a signal of high demand; thus, they are more willing to resell it and set a higher price premium in the resale. These findings can be taken as input by the NFT market platforms in their NFT display decisions to boost primary market sales and promote resale market participation. The findings can also help content creators formulate creative strategies that lead to higher bidding probability or sales performance.

Transfer learning across different chemical domains: virtual screening of organic materials with deep learning models pretrained on small molecule and chemical reaction data

Machine learning is becoming a preferred method for the virtual screening of organic materials due to its cost-effectiveness over traditional computationally demanding techniques. However, the scarcity of labeled data for organic materials poses a significant challenge for training advanced machine learning models. This study showcases the potential of utilizing databases of drug-like small molecules and chemical reactions to pretrain the BERT model, enhancing its performance in the virtual screening of organic materials. By fine-tuning the BERT models with data from five virtual screening tasks, the version pretrained with the USPTO–SMILES dataset achieved R2 scores exceeding 0.94 for three tasks and over 0.81 for two others. This performance surpasses that of models pretrained on the small molecule or organic materials databases and outperforms three traditional machine learning models trained directly on virtual screening data. The success of the USPTO–SMILES pretrained BERT model can be attributed to the diverse array of organic building blocks in the USPTO database, offering a broader exploration of the chemical space. The study further suggests that accessing a reaction database with a wider range of reactions than the USPTO could further enhance model performance. Overall, this research validates the feasibility of applying transfer learning across different chemical domains for the efficient virtual screening of organic materials.Scientific contributionThis study verifies the feasibility of applying transfer learning to large language models in different chemical fields to help organic materials perform virtual screening. Through the comparison of transfer learning from different chemical fields to a variety of organic material molecules, the high precision virtual screening of organic materials is realized.

Augmenting Knowledge for Individual NVR Prediction in Different Spatial and Temporal Cross-Building Environments

Natural ventilation is a critical method for reducing energy consumption for heating, cooling, and ventilating buildings. Recent research has focused on utilizing environmental IoT data from both inside and outside buildings for NVR prediction based on a deep learning model. To design an accurate NVR prediction model while considering individual building environments, various knowledge-sharing methods can be applied, such as transfer learning and ensemble models for cross-building prediction. However, the characteristics of learning data and model parameters should be considered when applying transfer learning and ensemble models to predict NVR with different spatial and temporal domains. In this paper, we propose a way to design an NVR prediction model for a cross-building environment by normalizing the training data, selecting transfer learning layers that are well-suited to the data environment, and augmenting NVR knowledge via ensemble methods. Based on the experimental results, we confirm that the proposed knowledge-sharing deep learning approach, while considering the normalizing of training data, the selecting transfer learning layers, and augmenting the NVR knowledge approach, can improve the accuracy up to 11.8% in the two different offices and seasons.

Classification of Breast Cancer with Transfer Learning on Convolutional Neural Network Models

Breast cancer is a serious medical condition and a leading cause of death among women. Early and accurate diagnosis is crucial for improving patient outcomes. This study explores the use of Convolutional Neural Networks (CNNs) with Transfer Learning using DenseNet121 and ResNet50 models to enhance breast cancer classification via mammography. Transfer Learning enables CNN models to leverage knowledge learned from larger datasets such as ImageNet to improve performance on specific breast cancer datasets. The dataset comprised medical images with three breast variations: benign, malignant, and normal, totaling 531 data points. Data was split with a 70% training and 30% validation ratio. Two CNN models, AlexNet and ResNet50, were evaluated to compare their performance in classifying these breast cancer types. The experimental results show that AlexNet achieved a training accuracy of 98.01%, while ResNet50 achieved 64.07%. AlexNet demonstrated superior performance in identifying complex patterns in mammography images, resulting in more accurate classification of different breast cancer types. These findings highlight the potential of deep learning applications to support more precise and effective medical diagnostics for breast cancer. This research contributes significantly to the development of AI technologies in healthcare aimed at improving early detection of breast cancer. The implications of this study could expand our understanding of Transfer Learning applications in medical contexts, driving further advancements in this field to enhance patient care and prognosis

Innovative Approaches to Clinical Diagnosis: Transfer Learning in Facial Image Classification for Celiac Disease Identification

Background: Celiac disease arises from gluten consumption and shares symptoms with other conditions, leading to delayed diagnoses. Untreated celiac disease heightens the risk of autoimmune disorders, neurological issues, and certain cancers like lymphoma while also impacting skin health due to intestinal disruptions. This study uses facial photos to distinguish individuals with celiac disease from those without. Surprisingly, there is a lack of research involving transfer learning for this purpose despite its benefits such as faster training, enhanced performance, and reduced overfitting. While numerous studies exist on endoscopic intestinal photo classification and a few have explored the link between facial morphology measurements and celiac disease, none have concentrated on diagnosing celiac disease through facial photo classification. Methods: This study sought to apply transfer learning techniques with VGG16 to address a gap in research by identifying distinct facial features that differentiate patients with celiac disease from healthy individuals. A dataset containing a total of 200 facial images of adult individuals with and without celiac condition was utilized. Half of the dataset had a ratio of 70% females to 30% males with celiac condition, and the rest had a ratio of 60% females to 40% males without celiac condition. Among those with celiac condition, 28 were newly diagnosed and 72 had been previously diagnosed, with 25 not adhering to a gluten-free diet and 47 partially adhering to such a diet. Results: Utilizing transfer learning, the model achieved a 73% accuracy in classifying the facial images of the patients during testing, with corresponding precision, recall, and F1 score values of 0.54, 0.56, and 0.52, respectively. The training process involved 50,178 parameters, showcasing the model’s efficacy in diagnostic image analysis. Conclusions: The model correctly classified approximately three-quarters of the test images. While this is a reasonable level of accuracy, it also suggests that there is room for improvement as the dataset contains images that are inherently difficult to classify even for humans. Increasing the proportion of newly diagnosed patients in the dataset and expanding the dataset size could notably improve the model’s efficacy. Despite being the first study in this field, further refinement holds promise for the development of a diagnostic tool for celiac disease using transfer learning in medical image analysis, addressing the lack of prior studies in this area.

COVID-19 Identification System from X-Ray Images of Chest using Deep Neural Network with Transfer Learning

Recently, the impact of COVID-19 has significantly diminished; however, it has not been completely eradicated. There are still instances where individuals are experiencing suffering due to this life-threatening virus which has a significant impact on health care as well as lifestyles throughout the world. So, early discovery is important to controlling case extension and the death rate. The RT-PCR is known as the true leading diagnosis test; nevertheless, the expense and result times of these tests are long, thus additional quick and accessible diagnostic techniques are required. However, most countries are suffering due to limited testing resources and kits. The unavailability of testing resources, kits, and a rising amount of regular occurrences, caused us to develop a model on Deep Learning which may benefit radiologists as well as doctors for detecting COVID-19 instances using images of chest X-rays. For developing a representation of modality-specific features, a convolutions neural network and a variety of ImageNet pre-trained models are trained and evaluated at the patient level by using different available CXR datasets. We choose 5000 images in total from the dataset collected from Kaggle where we kept 4000 images in case of training and validation, and the remaining 1000 in case of testing. We use four Pre-train Deep CNN Models which are very popular for image calcification. VVG16, VGG19, InceptionV3, and Resnet50 CNN Models we choose to analyze the performance and find the best one among them. In our testing, we get 88.5% testing accuracy on ResNet and 95.10% on InceptionV3 models while VGG19 gives 90.22% accuracy and VGG16 gives the highest 96.10% accuracy. To increase performance accuracy, Transfer Learning knowledge is transmitted and fine-tuned. After applying Transfer Learning in the modified VGG16 we got an accuracy of 97% which is clearly an improvement over the previous VGG16 model. GUB JOURNAL OF SCIENCE AND ENGINEERING, Vol 10(1), 2023 P 53-67

Patient-specific deep learning for 3D protoacoustic image reconstruction and dose verification in proton therapy.

Protoacoustic (PA) imaging has the potential to provide real-time 3D dose verification of proton therapy. However, PA images are susceptible to severe distortion due to limited angle acquisition. Our previous studies showed the potential of using deep learning to enhance PA images. As the model was trained using a limited number of patients' data, its efficacy was limited when applied to individualpatients. In this study, we developed a patient-specific deep learning method for protoacoustic imaging to improve the reconstruction quality of protoacoustic imaging and the accuracy of dose verification for individualpatients. Our method consists of two stages: in the first stage, a group model is trained from a diverse training set containing all patients, where a novel deep learning network is employed to directly reconstruct the initial pressure maps from the radiofrequency (RF) signals; in the second stage, we apply transfer learning on the pre-trained group model using patient-specific dataset derived from a novel data augmentation method to tune it into a patient-specific model. Raw PA signals were simulated based on computed tomography (CT) images and the pressure map derived from the planned dose. The reconstructed PA images were evaluated against the ground truth by using the root mean squared errors (RMSE), structural similarity index measure (SSIM) and gamma index on 10 specific prostate cancer patients. The significance level was evaluated by t-test with the p-value threshold of 0.05 compared with the results from the groupmodel. The patient-specific model achieved an average RMSE of 0.014 ( ), and an average SSIM of 0.981 ( ), out-performing the group model. Qualitative results also demonstrated that our patient-specific approach acquired better imaging quality with more details reconstructed when comparing with the group model. Dose verification achieved an average RMSE of 0.011 ( ), and an average SSIM of 0.995 ( ). Gamma index evaluation demonstrated a high agreement (97.4% [ ] and 97.9% [ ] for 1%/3 and 1%/5mm) between the predicted and the ground truth dose maps. Our approach approximately took 6 s to reconstruct PA images for each patient, demonstrating its feasibility for online 3D dose verification for prostate protontherapy. Our method demonstrated the feasibility of achieving 3D high-precision PA-based dose verification using patient-specific deep-learning approaches, which can potentially be used to guide the treatment to mitigate the impact of range uncertainty and improve the precision. Further studies are needed to validate the clinical impact of thetechnique.

LVPocket: integrated 3D global-local information to protein binding pockets prediction with transfer learning of protein structure classification

BackgroundPrevious deep learning methods for predicting protein binding pockets mainly employed 3D convolution, yet an abundance of convolution operations may lead the model to excessively prioritize local information, thus overlooking global information. Moreover, it is essential for us to account for the influence of diverse protein folding structural classes. Because proteins classified differently structurally exhibit varying biological functions, whereas those within the same structural class share similar functional attributes.ResultsWe proposed LVPocket, a novel method that synergistically captures both local and global information of protein structure through the integration of Transformer encoders, which help the model achieve better performance in binding pockets prediction. And then we tailored prediction models for data of four distinct structural classes of proteins using the transfer learning. The four fine-tuned models were trained on the baseline LVPocket model which was trained on the sc-PDB dataset. LVPocket exhibits superior performance on three independent datasets compared to current state-of-the-art methods. Additionally, the fine-tuned model outperforms the baseline model in terms of performance.Scientific contributionWe present a novel model structure for predicting protein binding pockets that provides a solution for relying on extensive convolutional computation while neglecting global information about protein structures. Furthermore, we tackle the impact of different protein folding structures on binding pocket prediction tasks through the application of transfer learning methods.Graphical

Data Imbalance, Uncertainty Quantification, and Transfer Learning in Data‐Driven Parameterizations: Lessons From the Emulation of Gravity Wave Momentum Transport in WACCM

AbstractNeural networks (NNs) are increasingly used for data‐driven subgrid‐scale parameterizations in weather and climate models. While NNs are powerful tools for learning complex non‐linear relationships from data, there are several challenges in using them for parameterizations. Three of these challenges are (a) data imbalance related to learning rare, often large‐amplitude, samples; (b) uncertainty quantification (UQ) of the predictions to provide an accuracy indicator; and (c) generalization to other climates, for example, those with different radiative forcings. Here, we examine the performance of methods for addressing these challenges using NN‐based emulators of the Whole Atmosphere Community Climate Model (WACCM) physics‐based gravity wave (GW) parameterizations as a test case. WACCM has complex, state‐of‐the‐art parameterizations for orography‐, convection‐, and front‐driven GWs. Convection‐ and orography‐driven GWs have significant data imbalance due to the absence of convection or orography in most grid points. We address data imbalance using resampling and/or weighted loss functions, enabling the successful emulation of parameterizations for all three sources. We demonstrate that three UQ methods (Bayesian NNs, variational auto‐encoders, and dropouts) provide ensemble spreads that correspond to accuracy during testing, offering criteria for identifying when an NN gives inaccurate predictions. Finally, we show that the accuracy of these NNs decreases for a warmer climate (4 × CO2). However, their performance is significantly improved by applying transfer learning, for example, re‐training only one layer using ∼1% new data from the warmer climate. The findings of this study offer insights for developing reliable and generalizable data‐driven parameterizations for various processes, including (but not limited to) GWs.

Prediction of LncRNA-protein Interactions Using Auto-Encoder, SE-ResNet Models and Transfer Learning.

Long non-coding RNA (lncRNA) plays a crucial role in various biological processes, and mutations or imbalances of lncRNAs can lead to several diseases, including cancer, Prader-Willi syndrome, autism, Alzheimer's disease, cartilage-hair hypoplasia, and hearing loss. Understanding lncRNA-protein interactions (LPIs) is vital for elucidating basic cellular processes, human diseases, viral replication, transcription, and plant pathogen resistance. Despite the development of several LPI calculation methods, predicting LPI remains challenging, with the selection of variables and deep learning structure being the focus of LPI research. We propose a deep learning framework called AR-LPI, which extracts sequence and secondary structure features of proteins and lncRNAs. The framework utilizes an auto-encoder for feature extraction and employs SE-ResNet for prediction. Additionally, we apply transfer learning to the deep neural network SE-ResNet for predicting small-sample datasets. Through comprehensive experimental comparison, we demonstrate that the AR-LPI architecture performs better in LPI prediction. Specifically, the accuracy of AR-LPI increases by 2.86% to 94.52%, while the F-value of AR-LPI increases by 2.71% to 94.73%. Our experimental results show that the overall performance of AR-LPI is better than that of other LPI prediction tools.