Processing Model Research Articles

Edge–cloud collaborative estimation lithium-ion battery SOH based on MEWOA-VMD and Transformer

The State of Health (SOH) of lithium-ion batteries significantly impacts the performance, safety, and reliability of the battery, making it a crucial component of the battery management system. Addressing the issues of inadequate accuracy and lack of robustness in current SOH estimation methods, this study introduces a novel methodology for estimating SOH in lithium-ion batteries. It leverages the multi-population evolution whale optimization algorithm optimized variational mode decomposition (MEWOA-VMD) in conjunction with Transformer architecture. This framework enhances the efficiency and accuracy of SOH estimation by leveraging the computational capabilities of edge devices for real-time data processing, as well as the robust data processing power and model training advantages offered by cloud computing. Specifically, MEWOA is utilized to optimize VMD parameters, enabling MEWOA-VMD to fully decompose the capacity signal of lithium-ion batteries. This results in a component showing a global attenuation trend and a set of fluctuating components that represent capacity regeneration, thereby minimizing the impact of capacity regeneration on SOH estimation. Subsequently, all components are collectively input into the Transformer, marking the first application of this method for input. To enhance convergence speed and training efficiency, the layer normalization (LN) layer within the neural network architecture is proactively advanced. Finally, various artificial neural networks are compared and validated on three publicly available datasets. Furthermore, Gaussian noise is introduced into the original data to validate robustness. To confirm the practical applicability of the proposed method, real-world vehicle data is used for SOH estimation. The results indicate that the proposed method achieves a maximum MSE of no more than 0.009% across three publicly available datasets, showcasing improved accuracy and stability in SOH estimation. The practical applicability is further validated using real-world vehicle data, proving the proposed method’s potential for application in edge cloud-based battery management systems.

Self-Administered Interventions Based on Natural Language Processing Models for Reducing Depressive and Anxiety Symptoms: Systematic Review and Meta-Analysis.

The introduction of natural language processing (NLP) technologies has significantly enhanced the potential of self-administered interventions for treating anxiety and depression by improving human-computer interactions. Although these advances, particularly in complex models such as generative artificial intelligence (AI), are highly promising, robust evidence validating the effectiveness of the interventions remains sparse. The aim of this study was to determine whether self-administered interventions based on NLP models can reduce depressive and anxiety symptoms. We conducted a systematic review and meta-analysis. We searched Web of Science, Scopus, MEDLINE, PsycINFO, IEEE Xplore, Embase, and Cochrane Library from inception to November 3, 2023. We included studies with participants of any age diagnosed with depression or anxiety through professional consultation or validated psychometric instruments. Interventions had to be self-administered and based on NLP models, with passive or active comparators. Outcomes measured included depressive and anxiety symptom scores. We included randomized controlled trials and quasi-experimental studies but excluded narrative, systematic, and scoping reviews. Data extraction was performed independently by pairs of authors using a predefined form. Meta-analysis was conducted using standardized mean differences (SMDs) and random effects models to account for heterogeneity. In all, 21 articles were selected for review, of which 76% (16/21) were included in the meta-analysis for each outcome. Most of the studies (16/21, 76%) were recent (2020-2023), with interventions being mostly AI-based NLP models (11/21, 52%); most (19/21, 90%) delivered some form of therapy (primarily cognitive behavioral therapy: 16/19, 84%). The overall meta-analysis showed that self-administered interventions based on NLP models were significantly more effective in reducing both depressive (SMD 0.819, 95% CI 0.389-1.250; P<.001) and anxiety (SMD 0.272, 95% CI 0.116-0.428; P=.001) symptoms compared to various control conditions. Subgroup analysis indicated that AI-based NLP models were effective in reducing depressive symptoms (SMD 0.821, 95% CI 0.207-1.436; P<.001) compared to pooled control conditions. Rule-based NLP models showed effectiveness in reducing both depressive (SMD 0.854, 95% CI 0.172-1.537; P=.01) and anxiety (SMD 0.347, 95% CI 0.116-0.578; P=.003) symptoms. The meta-regression showed no significant association between participants' mean age and treatment outcomes (all P>.05). Although the findings were positive, the overall certainty of evidence was very low, mainly due to a high risk of bias, heterogeneity, and potential publication bias. Our findings support the effectiveness of self-administered NLP-based interventions in alleviating depressive and anxiety symptoms, highlighting their potential to increase accessibility to, and reduce costs in, mental health care. Although the results were encouraging, the certainty of evidence was low, underscoring the need for further high-quality randomized controlled trials and studies examining implementation and usability. These interventions could become valuable components of public health strategies to address mental health issues. PROSPERO International Prospective Register of Systematic Reviews CRD42023472120; https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42023472120.

Integrating ChatGPT in Orthopedic Education for Medical Undergraduates: Randomized Controlled Trial.

ChatGPT is a natural language processing model developed by OpenAI, which can be iteratively updated and optimized to accommodate the changing and complex requirements of human verbal communication. The study aimed to evaluate ChatGPT's accuracy in answering orthopedics-related multiple-choice questions (MCQs) and assess its short-term effects as a learning aid through a randomized controlled trial. In addition, long-term effects on student performance in other subjects were measured using final examination results. We first evaluated ChatGPT's accuracy in answering MCQs pertaining to orthopedics across various question formats. Then, 129 undergraduate medical students participated in a randomized controlled study in which the ChatGPT group used ChatGPT as a learning tool, while the control group was prohibited from using artificial intelligence software to support learning. Following a 2-week intervention, the 2 groups' understanding of orthopedics was assessed by an orthopedics test, and variations in the 2 groups' performance in other disciplines were noted through a follow-up at the end of the semester. ChatGPT-4.0 answered 1051 orthopedics-related MCQs with a 70.60% (742/1051) accuracy rate, including 71.8% (237/330) accuracy for A1 MCQs, 73.7% (330/448) accuracy for A2 MCQs, 70.2% (92/131) accuracy for A3/4 MCQs, and 58.5% (83/142) accuracy for case analysis MCQs. As of April 7, 2023, a total of 129 individuals participated in the experiment. However, 19 individuals withdrew from the experiment at various phases; thus, as of July 1, 2023, a total of 110 individuals accomplished the trial and completed all follow-up work. After we intervened in the learning style of the students in the short term, the ChatGPT group answered more questions correctly than the control group (ChatGPT group: mean 141.20, SD 26.68; control group: mean 130.80, SD 25.56; P=.04) in the orthopedics test, particularly on A1 (ChatGPT group: mean 46.57, SD 8.52; control group: mean 42.18, SD 9.43; P=.01), A2 (ChatGPT group: mean 60.59, SD 10.58; control group: mean 56.66, SD 9.91; P=.047), and A3/4 MCQs (ChatGPT group: mean 19.57, SD 5.48; control group: mean 16.46, SD 4.58; P=.002). At the end of the semester, we found that the ChatGPT group performed better on final examinations in surgery (ChatGPT group: mean 76.54, SD 9.79; control group: mean 72.54, SD 8.11; P=.02) and obstetrics and gynecology (ChatGPT group: mean 75.98, SD 8.94; control group: mean 72.54, SD 8.66; P=.04) than the control group. ChatGPT answers orthopedics-related MCQs accurately, and students using it excel in both short-term and long-term assessments. Our findings strongly support ChatGPT's integration into medical education, enhancing contemporary instructional methods. Chinese Clinical Trial Registry Chictr2300071774; https://www.chictr.org.cn/hvshowproject.html ?id=225740&v=1.0.

Enhancing ESL Students’ Writing Skills through Natural Language Processing Model Chat GPT

This research paper describes an experimental study to determine the effectiveness of using Chat GPT based natural language processing (NLP) model to improve the writing proficiency of ESL students. The study shows how such approach can be integrated into the teaching process to improve the quality of writing skills. The paper provides an overview of methods of teaching writing according to certain criteria for assessing writing skills. The effectiveness of the model is assessed in comparison with traditional teaching methods. Recommendations are given for optimizing the learning process using NLP technologies. The conclusion highlights the significance of the technology used in improving writing skills and discusses the prospects for using the Chat GPT model to optimize these skills in ESL students.

The BioRECIPE Knowledge Representation Format.

The BioRECIPE (Biological system Representation for Evaluation, Curation, Interoperability, Preserving, and Execution) knowledge representation format was introduced to standardize and facilitate human-machine interaction while creating, verifying, evaluating, curating, and expanding executable models of intra- and intercellular signaling. This format allows a human user to easily preview and modify any model component, while it is at the same time readable by machines and can be processed by a suite of model development and analysis tools. The BioRECIPE format is compatible with multiple representation formats, natural language processing tools, modeling tools, and databases that are used by the systems and synthetic biology communities.

LangchainIQ: Intelligent Content and Query Processing

Purpose: The purpose of this research is to introduce and evaluate a comprehensive framework called langchain(component of Large Language Model), designed to optimize data analysis and visualization processes across various business domains. The framework integrates advanced computational techniques with user-friendly interfaces to meet the growing demand for efficient information processing tools in research and industry settings. Design/Methodology/Approach: The framework consists of three primary components: PDF answering, CSV analytics, and data visualization using the LIDA library. Integration of advanced technologies such as the Mistral 7B model for language processing, Faiss for similarity search, and the LIDA library for data visualization. Detailed implementation steps include content processing, embedding using OpenAI embeddings, storage and retrieval using Faiss, and query handling using Mistral 7B. This involves breaking down PDF and CSV content into chunks, embedding them, and utilizing advanced algorithms for efficient data retrieval and visualization. Findings/Result: The fine-tuned Mistral 7B model significantly enhances data extraction speed compared to traditional models like Llama. Users can effectively query and extract specific information from PDFs and CSVs using natural language, facilitated by advanced AI models. The LIDA library automates the generation of insightful visualizations from processed data, enhancing data interpretation and decision-making. Originality/Value: Introducing langchain as a versatile framework that addresses the complexities of data analysis and visualization and it’s use in business analysis. Paper Type: Technical Research.

Cartographic Generalization of Islands Using Remote Sensing Images for Multiscale Representation

The multi-scale representation of remote sensing images provides various levels of image information crucial for decision-making in GIS applications and plays a significant role in information processing, data analysis, and geographic modeling. Traditional methods for multi-scale representation of remote sensing images often struggle to simplify local details of individual targets while preserving the overall characteristics of target groups. These methods also encounter issues such as transitional texture distortion and rough final boundaries. This paper proposes a novel multi-scale representation method for remote sensing images based on computer vision techniques, which effectively maintains the overall characteristics of target groups. Initially, the K-means algorithm is employed to distinguish between islands and oceans. Subsequently, a superpixel segmentation algorithm is used to aggregate island groups and simplify the generated boundaries. Finally, texture synthesis and transfer are applied based on the original image to produce the aggregated island images. Evaluation metrics demonstrate that this method can generate multi-scale aggregated images of islands, effectively eliminate redundant information, and produce smooth boundaries.

Utilizing Large Language Models to Illustrate Constraints for Construction Planning

Effective construction project planning relies on addressing constraints related to materials, labor, equipment, and others. Planning meetings are typical venues for stakeholders to identify, communicate, and remove constraints. However, a critical gap exists in lacking an automated approach to identify, classify, analyze, and track constraint discussions during onsite planning meetings. Therefore, this research aims to 1. develop a natural language processing model to classify constraints in meeting discussions; 2. uncover the discussion patterns of managers and foremen regarding various constraints; and 3. extract the root causes for constraints, evaluate their impacts, and prepare managers to develop practical solutions for constraint removal. This research collected meeting transcripts from 94 onsite planning meetings of a building project, spanning 263,836 words. Next, this research leveraged a general pretrained transformer (GPT) to segment discussion dialogs into topics. A Bidirectional Encoder Representations from Transformers (BERT)-based model was developed to categorize constraint types for each topic. The constraint patterns among meeting attendees were assessed. Furthermore, a GPT-based tool was devised to track root causes, impacts, and solutions for various constraints. Test results revealed an 8.8% improvement in constraint classification accuracy compared with the traditional classification model. An occupational characteristic in constraint discussion was observed in that the management team tended to balance their focus on various constraints, while foremen concentrated on more practical issues. This research contributes to the body of knowledge by leveraging language models to analyze construction planning meetings. The findings facilitate project managers in establishing constraint logs for diagnosing and prognosticating planning issues.

Determination of soluble solids content in tomatoes with different nitrogen levels based on hyperspectral imaging technique.

Tomato is sweet and sour with high nutritional value, and soluble solids content (SSC) is an important indicator of tomato flavor. Due to the different mechanisms of nitrogen uptake and assimilation in plants, exogenous supply of different forms of nitrogen will have different effects on the growth, development, and physiological metabolic processes of tomato, thus affecting the tomato flavor. In this paper, hyperspectral imaging (HSI) technique combined with neural network prediction model was used to predict SSC of tomato under different nitrogen treatments. Competitive adaptive reweighed sampling (CARS) and iterative retained information variable (IRIV) were used to extract the feature wavelengths. Based on the characteristic wavelength, the prediction models of tomato SSC are established by custom convolutional neural network (CNN) model that was constructed and optimized. The results showed that the SSC of tomato was negatively correlated with nitrogen fertilizer concentration. For tomatoes treated with different nitrogen concentrations, the residual predictive deviation (RPD) of CARS-CNN and IRIV-parallel convolutional neural networks (PCNN) reached 1.64 and 1.66, both more than 1.6, indicating good model prediction. This study provides technical support for future online nondestructive testing of tomato quality. PRACTICAL APPLICATION: The CARS-CNN and IRIV-PCNN were the best data processing model. Four customized convolutional neural networks were used for predictive modeling. The CNN model provides more accurate results than conventional methods.

Hot Deformation Behavior and Microstructure Evolution of a Graphene/Copper Composite.

Graphene/copper composites are promising in electronic and energy fields due to their superior conductivity, but microstructure control during thermal mechanical processing (TMP) remains a crucial issue for the manufacturing of high-performance graphene/copper composites. In this study, the hot deformation behavior of graphene/copper composites was investigated by isothermal compression tests at deformation temperatures of 700~850 °C and strain rates of 0.01~10 s-1, and a constitutive equation based on the Arrhenius model and hot processing map was established. Results demonstrate that the deformation mechanism of the graphene/copper composites mainly involves dynamic recrystallization (DRX), and such DRX-mediated deformation behavior can be accurately described by the established Arrhenius model. In addition, it was found that the strain rate has a stronger impact on the DRX grain size than the deformation temperature. The optimum deformation temperature and strain rate were determined to be 800 °C and 1 s-1, respectively, with which a uniform microstructure with fine grains can be obtained.

A theoretical model of emotional processing in visual artmaking and art therapy

Current theoretical models of emotional processing rely mainly on detecting emotional processing through verbal, conscious, and cognitive processes. However, artmaking can potentially reveal embodied and implicit processes that may otherwise remain hidden in verbal expression. This paper attempts to close the scholarly gap by introducing a novel art-based emotional processing model that integrates emotional processing and art therapy literature, incorporating emotional meaning-making, awareness, acceptance, and memory consolidation. The art-based EP model explains the processes through which art creation may benefit emotional processing. It also elucidates the ways in which art therapy can be used to enhance therapeutic outcomes.

DiRecNetV2: A Transformer-Enhanced Network for Aerial Disaster Recognition

The integration of Unmanned Aerial Vehicles (UAVs) with artificial intelligence (AI) models for aerial imagery processing in disaster assessment, necessitates models that demonstrate exceptional accuracy, computational efficiency, and real-time processing capabilities. Traditionally Convolutional Neural Networks (CNNs), demonstrate efficiency in local feature extraction but are limited by their potential for global context interpretation. On the other hand, Vision Transformers (ViTs) show promise for improved global context interpretation through the use of attention mechanisms, although they still remain underinvestigated in UAV-based disaster response applications. Bridging this research gap, we introduce DiRecNetV2, an improved hybrid model that utilizes convolutional and transformer layers. It merges the inductive biases of CNNs for robust feature extraction with the global context understanding of Transformers, maintaining a low computational load ideal for UAV applications. Additionally, we introduce a new, compact multi-label dataset of disasters, to set an initial benchmark for future research, exploring how models trained on single-label data perform in a multi-label test set. The study assesses lightweight CNNs and ViTs on the AIDERSv2 dataset, based on the frames per second (FPS) for efficiency and the weighted F1 scores for classification performance. DiRecNetV2 not only achieves a weighted F1 score of 0.964 on a single-label test set but also demonstrates adaptability, with a score of 0.614 on a complex multi-label test set, while functioning at 176.13 FPS on the Nvidia Orin Jetson device.

Does gender impact the relationship between perceived value and intentions of use of natural processing models?

Purpose In the era of artificial intelligence, natural language processing (NLP) models are revolutionizing numerous sectors. This research aims to explore the perceived value of them among university students. In particular, it aims to investigate how gender may influence students’ intention to use these models in educational contexts, highlighting potentially significant differences that could inform the implementation and adoption of educational technologies. Design/methodology/approach This study investigates the relationship between perceived value and students' intention to adopt NLP models, considering gender as a moderator. The research involves 562 students from the University of Salerno, in Italy, and uses confirmatory factor analysis to evaluate the reliability and validity of the measurement scales. A regression model with robust errors is used to explore the moderating role of gender on the relationship between perceived value and intentions of use of NLP models. Findings The results reveal a significant positive association between perceived value and intention to use NLP models, confirming that students with higher perceived value are more likely to adopt these technologies. Furthermore, gender moderates this relationship, indicating that females are less prone to use NLP models than male counterparts. Originality/value Research takes on a significant role in the academic field, underlining the importance of adapting teaching practices to the increasingly widespread digitalization. The inclusion of NLP models in university programs emerges as a possible improvement of the learning experience, ensuring cutting-edge education in tune with the needs of the digital society.

Enhancing Asian Indigenous Language Processing through Deep Learning-based Handwriting Recognition and Optimization Techniques

Asian indigenous language or autochthonous language is a language which is native to a region and spoken by indigenous people in Asia. This language is a linguistically different community created in the region. Recently, researchers in handwriting detection studies comparing with indigenous languages have attained important internet amongst the research community. A new development of artificial intelligence (AI), natural language processing (NLP), cognitive analytics, and computational linguistics (CL) find it helpful in the analysis of regional low-resource languages. It can be obvious in the obtainability of effectual machine detection methods and open access handwritten databases. Tamil is the most ancient Indian language that is mostly exploited in the Southern part of India, Sri Lanka, and Malaysia. Tamil handwritten Character Recognition (HCR) is a critical procedure in optical character detection. Therefore, this study designs a Henry Gas Solubility Optimization with Deep Learning-based Handwriting Recognition Model (HGSODL-HRM) for Asian Indigenous Language Processing. The proposed HGSODL-HRM technique relies on computer vision and DL concepts for automated handwriting recognition in the Tamil language, which is one of the popular indigenous languages in Asia. To accomplish this, the HGSODL-HRM technique employs a capsule network (CapsNet) model for feature vector generation with the HGSO algorithm as a hyperparameter optimizer. For the recognition of handwritten characters, wavelet neural network (WNN) model is exploited. Finally, the WNN parameters can be optimally chosen by sail fish optimizer (SFO) algorithm. To demonstrate the promising results of the HGSODL-HRM system, an extensive range of simulations can be implemented. The simulation outcomes stated the betterment of the HGSODL-HRM system compared to recent DL models.

Implementation and Efficient Analysis of Preprocessing Techniques in Deep Learning for Image Classification

Deep learning models have recently been preferred to perform certain image-processing tasks. Recently, with the increasing radiation, heat, and poor lighting conditions, the raw image samples may contain noisy and ambiguous information. To process these images, the deep learning model requires a large number of data samples to learn the missing information from other clear data samples. This necessitates training the neural network with a huge dataset. The researchers are now attempting to filter and improve such noisy images via preprocessing in order to provide valid and accurate feature information to the neural network layers. However, certain research studies claim that some useful information may be lost when the image is not preprocessed with an appropriate filter or enhancement technique. The MSA (meta-synthesis and analysis) approach is utilized in this work to present the impact of the image processing applications done with and without preprocessing steps. Also, this work summarizes the existing deep learning-based image processing models utilizing or not preprocessing steps in their implementation. This work has also found that 85% of the existing techniques involve a preprocessing step while developing a deep learning model. However, a maximum accuracy of 96.89% is observed on Sine-Net when it is implemented without a preprocessing and the same model gave 96.85% when implemented with preprocessing. This research provides various research insights on the requirement and non-requirement of preprocessing steps in a deep learning-based implementation.

Narratives of and for survival in family firms: family influence on narrative processing

PurposeResilience of long-lived family businesses has been widely acknowledged but the mechanisms enabling longevity need to be further investigated. This can be done by examining how narratives about crisis situations are processed in family firms.Design/methodology/approachBased on narrative interviews, this qualitative study examines how crisis situations have been dealt with and narrated by long-lived German family firms.FindingsNarratives of survival can have a strengthening effect so they become also narratives for survival. The analysis reveals how the constructive management of crises contributes to emotional attachment, identification and commitment and also strengthens resilience and longevity.Originality/valueThe study contributes to narrative identity theory by answering calls for narrative analysis and capturing the influence of narratives on family firms and business families. The discussion of the findings leads to the narrative processing model showing how past crises are processes to gain identity and resilience in the future and how the business family can influence this process.

Operando study of the dynamic evolution of multiple Fe-rich intermetallics of an Al recycled alloy in solidification by synchrotron X-ray and machine learning

Using synchrotron X-ray diffraction, tomography and machine-learning enabled phase segmentation strategy, we have studied under operando conditions the nucleation, co-growth and dynamic interplays among the dendritic and multiple intermetallic phases of a typical recycled Al alloy (Al5Cu1.5Fe1Si, wt.%) in solidification with and without ultrasound. The research has revealed and elucidated the underlying mechanisms that drive the formation of the very complex and convoluted Fe-rich phases with rhombic dodecahedron and 3D skeleton networks (the so-called Chinese-script type morphology). Through statistical microstructural analyses and numerical modelling of the ultrasound melt processing, the research has demonstrated that a short period of ultrasound processing of just 7 s in the liquid state is able to reduce the average size of the α-Al dendrites and the Fe-containing intermetallic phases by ∼5 times compared to the cases without ultrasound. For the first time, this work has revealed fully the nucleation and growth dynamics of the convoluted morphology of the Fe-containing intermetallic phases in 4D domain. The research has also demonstrated clearly the beneficial effects of applying ultrasound to control the Fe phases' morphology in recycled Al alloys and it is one of the most effective and green processing strategies.

Pseudo-ISP: Learning pseudo in-camera signal processing pipeline from a color image denoiser

Existing deep denoisers usually ignore the problem of noise discrepancy between training and test images caused by the different noise modeling of sensor and in-camera signal processing (ISP) pipeline, which inevitably degrades the denoising performance. In this paper, we present an unpaired learning scheme to adapt the color image denoisers for handling test images with noise discrepancy. To this end, we consider a practical training setting equipped with a pre-trained color denoiser, a set of test noisy sRGB images, and an unpaired set of clean sRGB images. Then, we propose to alternate between two modules, i.e., Pseudo-ISP training for learning noise modeling of sensor and in-camera signal processing (ISP) pipeline and denoiser adaption. Instead of modeling the complex noise in sRGB images, we assume the signal-dependent and spatially independent noise in rawRGB space, and specifically design the Pseudo-ISP to learn the pseudo ISP pipeline and pseudo rawRGB noise model jointly. Applying the learned Pseudo-ISP to the unpaired set of clean sRGB images, the corresponding realistic noisy sRGB observations can be easily obtained. Conversely, using the generated paired data for model fine-tuning, the color denoiser can be well adapted for test images. Through iterating between Pseudo-ISP training and denoiser adaption, denoising performance of deep denoisers can be further improved. Experiments show that our Pseudo-ISP not only can boost simple Gaussian blurring-based denoiser to achieve competitive performance against CBDNet, but also is effective in improving state-of-the-art deep denoisers, e.g., CBDNet and RIDNet.

Research status and application of artificial intelligence large models in the oil and gas industry

This article elucidates the concept of large model technology, summarizes the research status of large model technology both domestically and internationally, provides an overview of the application status of large models in vertical industries, outlines the challenges and issues confronted in applying large models in the oil and gas sector, and offers prospects for the application of large models in the oil and gas industry. The existing large models can be briefly divided into three categories: large language models, visual large models, and multimodal large models. The application of large models in the oil and gas industry is still in its infancy. Based on open-source large language models, some oil and gas enterprises have released large language model products using methods like fine-tuning and retrieval augmented generation. Scholars have attempted to develop scenario-specific models for oil and gas operations by using visual/multimodal foundation models. A few researchers have constructed pre-trained foundation models for seismic data processing and interpretation, as well as core analysis. The application of large models in the oil and gas industry faces challenges such as current data quantity and quality being difficult to support the training of large models, high research and development costs, and poor algorithm autonomy and control. The application of large models should be guided by the needs of oil and gas business, taking the application of large models as an opportunity to improve data lifecycle management, enhance data governance capabilities, promote the construction of computing power, strengthen the construction of “artificial intelligence + energy” composite teams, and boost the autonomy and control of large model technology.

Self-Supervised Medical Image Segmentation Using Deep Reinforced Adaptive Masking.

Self-supervised learning aims to learn transferable representations from unlabeled data for downstream tasks. Inspired by masked language modeling in natural language processing, masked image modeling (MIM) has achieved certain success in the field of computer vision, but its effectiveness in medical images remains unsatisfactory. This is mainly due to the high redundancy and small discriminative regions in medical images compared to natural images. Therefore, this paper proposes an adaptive hard masking (AHM) approach based on deep reinforcement learning to expand the application of MIM in medical images. Unlike predefined random masks, AHM uses an asynchronous advantage actor-critic (A3C) model to predict reconstruction loss for each patch, enabling the model to learn where masking is valuable. By optimizing the non-differentiable sampling process using reinforcement learning, AHM enhances the understanding of key regions, thereby improving downstream task performance. Experimental results on two medical image datasets demonstrate that AHM outperforms state-of-the-art methods. Additional experiments under various settings validate the effectiveness of AHM in constructing masked images.