Labeled Training Data Research Articles

An intelligent system for determining the degree of tree bark beetle damage based on the use of generative‐adversarial neural networks

AbstractThe increasing prevalence of tree bark beetle infestations poses a significant threat to forest ecosystems, leading to detrimental impacts on biodiversity, carbon storage, and timber resources. This study addresses the pressing need for innovative solutions to enhance forest management practices through early detection and assessment of tree health. We developed an intelligent system leveraging multispectral images and generative‐adversarial networks (GANs) to accurately determine the extent of bark beetle damage. Recognizing the challenges posed by traditional neural networks, which require vast amounts of labeled training data, we proposed a novel approach that utilizes a GAN architecture. In this system, a discriminator functions as a classifier, effectively trained on both real and synthetically generated data. Our methodology not only reduces the dependency on extensive labeled datasets but also enhances the robustness of the classification process. The results indicate a classification accuracy of 87.5%, demonstrating significant promise for improving detection capabilities even with limited training resources. The implications of this research are profound, offering potential benefits for the forestry industry through optimized management strategies and economic gains. Furthermore, our findings contribute to the preservation of critical ecosystem services by providing a means to monitor forest health effectively. However, the responsible implementation of this technology is paramount. Continuous refinement of the model, integration with traditional ecological knowledge, and ensuring transparency and equitable access to the developed system are essential for maximizing societal benefits. Additionally, addressing risks such as misinterpretation of data, overreliance on technology, and privacy concerns is crucial to minimize unintended consequences. Overall, this research presents a significant advancement in the field of forest health monitoring and establishes a foundation for future developments in intelligent ecological management systems.

2D Echocardiographic aortic root dimensions and longitudinal analysis with artificial intelligence using an open dataset from a UK-wide collaborative

Abstract Background Surveillance of aortic dimensions requires reproducible measurements, and knowledge of the rate of progression in those with dilatation. Purpose To develop an open machine-learning method for measuring the aortic root and proximal ascending aorta on echocardiograms, validate it through a multi-expert panel from the Unity UK Echocardiography AI Collaborative, and applying the AI to derive the rate of progression in historical databases. Methods The neural network was trained on 1478 parasternal long-axis images. Each image was labelled with key points for the aortic annulus, sinus and sinotubular junction, and proximal ascending aorta dimensions. Labels accommodated inner-to-inner and leading-to-leading conventions. For validation, the end-diastolic and mid-systolic images were identified (accommodating different guidelines recommendations) from 100 studies, and 10 expert echocardiographers made the aortic measurements. The consensus of experts defined the reference standard, and the variation between individual experts defined the acceptable variation for the AI. The AI was then applied to 724 echocardiograms of 102 patients under surveillance spanning 12 years (ranging from 4 to 17 scans per patient) without intervention. Training data labels and networks are made freely available on our project website. Results The median absolute deviations between the AI and the expert consensus ranged from 0.06cm to 0.17cm across the 16 measurements. For the individual expert opinions this was from 0.08cm to 0.19cm. For 7/16 measurements, there was no statistically significant difference between the AI deviation and the individual-expert deviation. The AI’s confidence level was a useful indicator of the reliability of the AI measure. For the top 9 deciles of AI confidence level, the AI measurements had significantly smaller deviations than the individual experts. The rates of progression averaged across the three aortic measurements was 2.0 mm/decade (95% CI 1.0 to 3.0; p =0.007). Conclusion Expert consensus can be used to both define the reference standard and also the range of acceptable deviation from consensus. Overall, the AI performs similarly to experts and more importantly, provides an automated estimate of confidence. The validated AI is consistent over time, and when applied retrospectively to patients with dilated aortic roots, found the rate of progression was more rapid than in previously reported healthy patients. This approach could be applied generally in developing medical AI to not only make reproducible measurements but to automate large scale longitudinal studies with little input, and ultimately serves as a useful research and clinical tool.Figure: AI(red) vs Expert (blue)Table: AIDeviations from Consensus

Transfer learning for probabilistic localization of hidden cracks in concrete structures

AbstractThe utility of discriminative supervised learning models built using multiple training-data sources is investigated for hidden crack localization in concrete. Feed-forward neural network (FFNN) is chosen as the model architecture, and transfer learning is used to assimilate the information obtained from different sources (computational physics simulations and laboratory experiments). The labeled training data consists of values of a damage index and the known locations of hidden cracks. The classification models need to learn how the presence of damage (hidden cracks) affects the damage index at different sensors for different test conditions. To this end, diagnostic FFNN models are built by sequentially adding and training new hidden layers to assimilate labeled information from computer models (different model geometries, test conditions, crack lengths, crack locations) and laboratory experiments on a plain cement slab. These transfer learning-based models are then used to localize damage in concrete specimens that reflect real-world conditions (i.e., specimens with steel reinforcement and randomly distributed aggregate). The actual damage state in these specimens is determined by extracting cores and performing petrographic studies on the extracted cores. The damage probability estimated by transfer learning-based models is compared with the petrographic damage rating index (DRI) to identify the most suitable approach to train the diagnostic models. The transfer learning-based diagnostic methodology shows promise and could be used in various structural health monitoring applications, where sufficient labeled data are typically not available from a single data source.

Noise attenuation in distributed acoustic sensing data using a guided unsupervised deep learning network

Distributed acoustic sensing (DAS) is a promising technology introducing a new paradigm in the acquisition of high-resolution seismic data. However, DAS data often show weak signals compared with the background noise, especially in challenging installation environments. In this study, we develop a new approach to denoise DAS data that leverages an unsupervised deep learning (DL) model, eliminating the need for labeled training data. The input DAS data undergo band-pass filtering to eliminate high-frequency content. Subsequently, a continuous wavelet transform (CWT) is performed, and the finest scale is used to guide the DL model in reconstructing the DAS signal. First, we extract 2D patches from the band-pass filtered data and the CWT scale of the data. Then, these patches are converted using an unrolling mechanism into 1D vectors to form the input of the DL model. A self-attention layer is included in each layer to extract the spatial relation between the band-pass filtered data and the CWT scale. Through an iterative process, the DL model tunes its parameters to suppress DAS noise, with the band-pass filtered data serving as the target for the network. The denoising performance of our framework is validated using field examples from the San Andreas Fault Observatory at Depth and Frontier Observatory for Research in Geothermal Energy data sets, where the data are recorded by a fiber-optic cable. Comparative analyses against three benchmark methods reveal the robust denoising performance of our framework.

GroupFormer for hyperspectral image classification through group attention

Hyperspectral image (HSI) data has a wide range of valuable spectral information for numerous tasks. HSI data encounters challenges such as small training samples, scarcity, and redundant information. Researchers have introduced various research works to address these challenges. Convolution Neural Network (CNN) has gained significant success in the field of HSI classification. CNN’s primary focus is to extract low-level features from HSI data, and it has a limited ability to detect long-range dependencies due to the confined filter size. In contrast, vision transformers exhibit great success in the HSI classification field due to the use of attention mechanisms to learn the long-range dependencies. As mentioned earlier, the primary issue with these models is that they require sufficient labeled training data. To address this challenge, we proposed a spectral-spatial feature extractor group attention transformer that consists of a multiscale feature extractor to extract low-level or shallow features. For high-level semantic feature extraction, we proposed a group attention mechanism. Our proposed model is evaluated using four publicly available HSI datasets, which are Indian Pines, Pavia University, Salinas, and the KSC dataset. Our proposed approach achieved the best classification results in terms of overall accuracy (OA), average accuracy (AA), and Kappa coefficient. As mentioned earlier, the proposed approach utilized only 5%, 1%, 1%, and 10% of the training samples from the publicly available four datasets.

Building usage prediction in complex urban scenes by fusing text and facade features from street view images using deep learning

Building usage maps are inputs to many urban planning applications, however, the existing methods and the available data have limitations in generating instance-level high-resolution usage maps. In this study we tackle this problem by utilizing Street View Images (SVIs) and proposing a novel ensemble learning architecture that leverages building facade features and text extracted from hoardings, posters, etc. on buildings to predict the usage class. A pre-trained object detection model i.e., Grounding DINO, is implemented to efficiently identify buildings. A novel manually labeled training data of detected buildings corresponding to their usage is used to extract features from building facades across diverse Indian cities (Hyderabad, Mumbai, Bangalore, Delhi) using Vision Transformer (ViT) model. Following this, CLIPSeg a pre-trained segmentation model is used to recognizes text specifically on building elements like signs, posters, and banners. We then leverage GPT-3.5 Turbo, a Large Language Model (LLM), fine-tuned with a specifically designed few-shot prompting method, to infer building usage from the recognized text. To achieve optimal performance, the proposed ensemble linear metaclassifier combines predictions from ViT and LLM model. The predicted building usages are attributed to their corresponding locations to develop spatial maps. An analysis of our framework compared against ground truth data collected from various Indian cities reveals significantly accurate outcomes. Our findings highlight the utility of textual information in classifying utilities and commercial buildings, while features extracted from vision models prove more informative for residential buildings. Our approach can automate the generation of roadside building attributes and usage details on a larger scale.

Self-adaptive fault diagnosis for unseen working conditions based on digital twins and domain generalization

In recent years, intelligent fault diagnosis based on domain adaptation has been used to address domain shifts in cyber–physical systems; however, the need for acquiring target data sufficiently limits their applicability to unseen working conditions. To overcome such limitations, domain generalization techniques have been introduced to enhance the capacity of fault diagnostic models to operate under unseen working conditions. Nevertheless, existing approaches assume access to extensive labeled training data from various source domains, posing challenges in real-world engineering scenarios due to resource constraints. Moreover, the absence of a mechanism for updating diagnostic models over time calls for the exploration of self-adaptive generalized diagnosis models that are capable of autonomous reconfiguration in response to new unseen working conditions. In such a context, this paper proposes a self-adaptive fault diagnosis system that combines several paradigms, namely Monitor-Analyze-Plan-Execute over a shared Knowledge (MAPE-K), Domain Generalization Network Models (DGNMs), and Digital Twins (DT). The MAPE-K loop enables run-time adaptation to dynamic industrial environments without human intervention. To address the scarcity of labeled training data, digital twins are used to generate supplementary data and continuously tune parameters to reflect the dynamics of new unseen working conditions. DGNM incorporates adversarial learning and a domain-based discrepancy metric to enhance feature diversity and generalization. The introduction of multi-domain data augmentation enhances feature diversity and facilitates learning correlations among multiple domains, ultimately improving the generalization of feature representations. The proposed fault diagnosis system has been evaluated on three publicly available rotating machinery datasets to demonstrate its higher performance in cross-work operation and cross-machine tasks compared to other state-of-the-art methods.

MemberShield: A framework for federated learning with membership privacy

Federated Learning (FL) allows multiple data owners to build high-quality deep learning models collaboratively, by sharing only model updates and keeping data on their premises. Even though FL offers privacy-by-design, it is vulnerable to membership inference attacks (MIA), where an adversary tries to determine whether a sample was included in the training data. Existing defenses against MIA cannot offer meaningful privacy protection without significantly hampering the model’s utility and causing a non-negligible training overhead. In this paper we analyze the underlying causes of the differences in the model behavior for member and non-member samples, which arise from model overfitting and facilitate MIAs. Accordingly, we propose MemberShield, a generalization-based defense method for MIAs that consists of: (i) one-time preprocessing of each client’s training data labels that transforms one-hot encoded labels to soft labels and eventually exploits them in local training, and (ii) early stopping the training when the local model’s validation accuracy does not improve on that of the global model for a number of epochs. Extensive empirical evaluations on three widely used datasets and four model architectures demonstrate that MemberShield outperforms state-of-the-art defense methods by delivering substantially better practical privacy protection against all forms of MIAs, while better preserving the target model utility. On top of that, our proposal significantly reduces training time and is straightforward to implement, by just tuning a single hyperparameter.

Prompt-guided and multimodal landscape scenicness assessments with vision-language models.

Recent advances in deep learning and Vision-Language Models (VLM) have enabled efficient transfer to downstream tasks even when limited labelled training data is available, as well as for text to be directly compared to image content. These properties of VLMs enable new opportunities for the annotation and analysis of images. We test the potential of VLMs for landscape scenicness prediction, i.e., the aesthetic quality of a landscape, using zero- and few-shot methods. We experiment with few-shot learning by fine-tuning a single linear layer on a pre-trained VLM representation. We find that a model fitted to just a few hundred samples performs favourably compared to a model trained on hundreds of thousands of examples in a fully supervised way. We also explore the zero-shot prediction potential of contrastive prompting using positive and negative landscape aesthetic concepts. Our results show that this method outperforms a linear probe with few-shot learning when using a small number of samples to tune the prompt configuration. We introduce Landscape Prompt Ensembling (LPE), which is an annotation method for acquiring landscape scenicness ratings through rated text descriptions without needing an image dataset during annotation. We demonstrate that LPE can provide landscape scenicness assessments that are concordant with a dataset of image ratings. The success of zero- and few-shot methods combined with their ability to use text-based annotations highlights the potential for VLMs to provide efficient landscape scenicness assessments with greater flexibility.

Self-Supervised Machine Learning Framework for Online Container Security Attack Detection

Container security has received much research attention recently. Previous work has proposed to apply various machine learning techniques to detect security attacks in containerized applications. On one hand, supervised machine learning schemes require sufficient labeled training data to achieve good attack detection accuracy. On the other hand, unsupervised machine learning methods are more practical by avoiding training data labeling requirements, but they often suffer from high false alarm rates. In this article, we present a generic self-supervised hybrid learning (SHIL) framework for achieving efficient online security attack detection in containerized systems. SHIL can effectively combine both unsupervised and supervised learning algorithms but does not require any manual data labeling. We have implemented a prototype of SHIL and conducted experiments over 46 real-world security attacks in 29 commonly used server applications. Our experimental results show that SHIL can reduce false alarms by 33%–93% compared to existing supervised, unsupervised, or semi-supervised machine learning schemes while achieving a higher or similar detection rate.

Geometry of Textual Data Augmentation: Insights from Large Language Models

Data augmentation is crucial for enhancing the performance of text classification models when labelled training data are scarce. For natural language processing (NLP) tasks, large language models (LLMs) are able to generate high-quality augmented data. But a fundamental understanding of the reasons for their effectiveness remains limited. This paper presents a geometric and topological perspective on textual data augmentation using LLMs. We compare the augmentation data generated by GPT-J with those generated through cosine similarity from Word2Vec and GloVe embeddings. Topological data analysis reveals that GPT-J generated data maintains label coherence. Convex hull analysis of such data represented by their two principal components shows that they lie within the spatial boundaries of the original training data. Delaunay triangulation reveals that increasing the number of augmented data points that are connected within these boundaries correlates with improved classification accuracy. These findings provide insights into the superior performance of LLMs in data augmentation. A framework for predicting the usefulness of augmentation data based on geometric properties could be formed based on these techniques.

Self-Supervised Learning across the Spectrum

Satellite image time series (SITS) segmentation is crucial for many applications, like environmental monitoring, land cover mapping, and agricultural crop type classification. However, training models for SITS segmentation remains a challenging task due to the lack of abundant training data, which requires fine-grained annotation. We propose S4, a new self-supervised pretraining approach that significantly reduces the requirement for labeled training data by utilizing two key insights of satellite imagery: (a) Satellites capture images in different parts of the spectrum, such as radio frequencies and visible frequencies. (b) Satellite imagery is geo-registered, allowing for fine-grained spatial alignment. We use these insights to formulate pretraining tasks in S4. To the best of our knowledge, S4 is the first multimodal and temporal approach for SITS segmentation. S4’s novelty stems from leveraging multiple properties required for SITS self-supervision: (1) multiple modalities, (2) temporal information, and (3) pixel-level feature extraction. We also curate m2s2-SITS, a large-scale dataset of unlabeled, spatially aligned, multimodal, and geographic-specific SITS that serves as representative pretraining data for S4. Finally, we evaluate S4 on multiple SITS segmentation datasets and demonstrate its efficacy against competing baselines while using limited labeled data. Through a series of extensive comparisons and ablation studies, we demonstrate S4’s ability as an effective feature extractor for downstream semantic segmentation.

TPLAD: Template-Parsed Log Anomaly Detection for Electrical Database Systems

In recent years, with the increasing complexity of software systems, logs have become crucial for system maintenance. Log-based anomaly detection plays a vital role in automatically detecting system anomalies through log analysis. However, current log-based anomaly detection approaches encounter significant practical challenges. Supervised methods often require a large amount of manually labeled training data, which can be time-consuming and costly to obtain. On the other hand, unsupervised and semi-supervised approaches may suffer from subpar performance, as they do not leverage historical anomalies to improve their detection capabilities. These challenges underscore the necessity for the development of more efficient and effective log-based anomaly detection methods. Database anomaly access detection is critical for ensuring the stability and security of database systems. We present a survey of existing log anomaly detection models and propose a novel approach, Template-Parsed Log Anomaly Detection (TPLAD) model, for automated anomaly detection in massive database log files. The proposed model combines the original log template with template parsing using code and text semantic representation. Experimental results demonstrate the effectiveness of the proposed approach in detecting abnormal database access patterns, including runtime errors, unauthorized access, and data leaks. The findings indicate that TPLAD model shows promise in enhancing database security and stability in business systems.

Localization and recognition of human action in 3D using transformers

Understanding a person’s behavior from their 3D motion sequence is a fundamental problem in computer vision with many applications. An important component of this problem is 3D action localization, which involves recognizing what actions a person is performing, and when the actions occur in the sequence. To promote the progress of the 3D action localization community, we introduce a new, challenging, and more complex benchmark dataset, BABEL-TAL (BT), for 3D action localization. Important baselines and evaluating metrics, as well as human evaluations, are carefully established on this benchmark. We also propose a strong baseline model, i.e., Localizing Actions with Transformers (LocATe), that jointly localizes and recognizes actions in a 3D sequence. The proposed LocATe shows superior performance on BABEL-TAL as well as on the large-scale PKU-MMD dataset, achieving state-of-the-art performance by using only 10% of the labeled training data. Our research could advance the development of more accurate and efficient systems for human behavior analysis, with potential applications in areas such as human-computer interaction and healthcare.

Multiscale Conditional Adversarial Networks based domain-adaptive method for rotating machinery fault diagnosis under variable working conditions

Deep learning has been increasingly used in health management and maintenance decision-making for rotating machinery. However, some challenges must be addressed to make this technology more effective. For example, the collected data is assumed to follow the same feature distribution, and sufficient labeled training data are available. Unfortunately, domain shifts occur inevitably in real-world scenarios due to different working conditions, and acquiring sufficient labeled samples is time-consuming and expensive in complex environments. This study proposes a novel domain adaptive framework called deep Multiscale Conditional Adversarial Networks (MCAN) for machinery fault diagnosis to address these shortcomings. The MCAN model comprises two key components. Constructed by a novel multiscale module with an attention mechanism, the first component is a shared feature generator that captures rich features at different internal perceptual scales, and the attention mechanism determines the weights assigned to each scale, enhancing the model’s dynamic adjustment and self-adaptation capabilities. The second component consists of two domain classifiers based on Bidirectional Long Short-Term Memory (BiLSTM) leveraging spatiotemporal features at various levels to achieve domain adaptation in the output space. The deep domain classifier also captures the cross-covariance dependencies between feature representations and classifier predictions, thereby improving the predictions’ discriminability. The proposed method has been evaluated using two publicly available fault diagnosis datasets and one condition monitoring experiment. The results of cross-domain transfer tasks demonstrated that the proposed method outperformed several state-of-the-art methods in terms of transferability and stability. This result is a significant step forward in deep learning for health management and maintenance decision-making for rotating machinery, and it has the potential to revolutionize its future application.

Remote sensing framework for geological mapping via stacked autoencoders and clustering

Supervised machine learning methods for geological mapping via remote sensing face limitations due to the scarcity of accurately labelled training data that can be addressed by unsupervised learning, such as dimensionality reduction and clustering. Dimensionality reduction methods have the potential to play a crucial role in improving the accuracy of geological maps. Although conventional dimensionality reduction methods may struggle with nonlinear data, unsupervised deep learning models such as autoencoders can model non-linear relationships. Stacked autoencoders feature multiple interconnected layers to capture hierarchical data representations useful for remote sensing data. We present an unsupervised machine learning-based framework for processing remote sensing data using stacked autoencoders for dimensionality reduction and k-means clustering for mapping geological units. We use Landsat 8, ASTER, and Sentinel-2 datasets to evaluate the framework for geological mapping of the Mutawintji region in Western New South Wales, Australia. We also compare stacked autoencoders with principal component analysis (PCA) and canonical autoencoders. Our results reveal that the framework produces accurate and interpretable geological maps, efficiently discriminating rock units. The results reveal that the combination of stacked autoencoders with Sentinel-2 data yields the best performance accuracy when compared to other combinations. We find that stacked autoencoders enable better extraction of complex and hierarchical representation of the input data when compared to canonical autoencoders and PCA. We also find that the generated maps align with prior geological knowledge of the study area while providing novel insights into geological structures.

Predicting Signal Reception Information from GNSS Satellites in Indoor Environments without Site Survey: Towards Opportunistic Indoor Positioning Based on GNSS Fingerprinting

While traditional GPS localization is impractical in certain indoor environments, GPS satellites offer a valuable source of indoor contextual information, particularly around window-side areas where signals are more accessible. This study introduces a new method that predicts GPS signal reception information in a target indoor environment without the need for actual data collection in the target environment. By predicting the strengths of GPS signals that are expected to be measured at various indoor positions, we can construct a GPS signal reception map of a target environment, which functions similarly to a Wi-Fi fingerprinting map surrounding window-side areas, offering a new dimension in indoor positioning systems. For instance, it aids in the correction of accumulated errors in pedestrian dead reckoning (PDR) systems. Our proposed method leverages easily accessible inputs, including satellite location data, indoor floorplan image of the target floor, and 3D mapping of the surrounding buildings, to predict signal reception at each position in the target environment. We developed a specialized neural network, named multi-scale branch fusion network (MSBF-Net), designed to process and integrate data of varying scales, such as the floorplan image of the target environment and 3D map of the surrounding area obtained from Google Earth, with a specific focus on understanding the line-of-sight (LOS) and multi-path effects caused by internal obstacles and surrounding buildings. This advanced capability enables the network to effectively interpret complex signal interactions within urban environments, enhancing its predictive accuracy for GPS signal reception. The effectiveness of our method was rigorously evaluated using real-world environments. In addition, we employed our method to implement opportunistic GPS fingerprint-based indoor positioning, where position estimates are provided when a strong signal is observed from at least one satellite. Surprisingly, the positioning method achieved a positioning error of only 2.8 meters when a smartphone is close to window-side areas even though the method does not rely on labeled training data collected in a target environment and signal infrastructures installed in the target environment.

A training-free non-intrusive air conditioning load monitoring method based on fuzzy comprehensive evaluation

Air conditioner (AC) is a grid friendly load, monitoring its operating status plays a key role in evaluating user side potential for participation in demand response. With continuous innovations in AC technology, the AC operation characteristics have shown increasing complexity, which poses significant challenges for non-intrusive AC load monitoring. Furthermore, obtaining labeled training data for target scenarios is often challenging, which hinders deployment of many Non-Intrusive Load Monitoring (NILM) methods in practice. In this regard, this article proposes a training-free non-intrusive AC load monitoring method based on fuzzy comprehensive evaluation (FCE), which is very easy to deploy and has good robustness. Firstly, considering various influence factors, a FCE model is constructed based on common knowledge about AC. This model simulates the manual labeling process to locate AC operation periods. Secondly, an AC power estimation method based on weighted filtering is established, in which weighted filter is used to eliminate the non-AC load components that may overlap during AC operation periods. Specifically, an optimization model is constructed to make the cumulative power of the AC events in its working cycle approach zero, with the filter weight vector as the objective variable, in which a heuristic greedy algorithm is used to find the optimal filter weight vector. When multiple feasible solutions with similar objective function values are derived in the solving process, the probability of load events appearing in AC and non-AC operation periods are compared. The final solution is selected, taking into account the likelihood of each feasible solution corresponding to a load event sequence belonging to AC load. The testing results on the PecanStreet dataset and multiple real-world measured datasets prove the effectiveness of the proposed method in monitoring complex ACs with poor waveform consistency.

Real-time tracking of structural evolution in 2D MXenes using theory-enhanced machine learning

In situ Electron Energy Loss Spectroscopy (EELS) combined with Transmission Electron Microscopy (TEM) has traditionally been pivotal for understanding how material processing choices affect local structure and composition. However, the ability to monitor and respond to ultrafast transient changes, now achievable with EELS and TEM, necessitates innovative analytical frameworks. Here, we introduce a machine learning (ML) framework tailored for the real-time assessment and characterization of in operando EELS Spectrum Images (EELS-SI). We focus on 2D MXenes as the sample material system, specifically targeting the understanding and control of their atomic-scale structural transformations that critically influence their electronic and optical properties. This approach requires fewer labeled training data points than typical deep learning classification methods. By integrating computationally generated structures of MXenes and experimental datasets into a unified latent space using Variational Autoencoders (VAE) in a unique training method, our framework accurately predicts structural evolutions at latencies pertinent to closed-loop processing within the TEM. This study presents a critical advancement in enabling automated, on-the-fly synthesis and characterization, significantly enhancing capabilities for materials discovery and the precision engineering of functional materials at the atomic scale.

Structural damage identification by using physics-guided residual neural networks

In recent years, vibration-based structural damage identification has made significant progress by exploiting data-driven deep learning techniques, which can efficiently extract damage-sensitive features from a large amount of data. However, in some practical engineering applications, large volumes of measurement data are not readily available. This paper proposes a novel physics-guided residual neural network (PhyResNet) framework to improve the robustness and accuracy of structural damage identification under data-scarce conditions. In contrast to the state-of-the-art purely data-driven ResNet, the proposed method embedded available physics knowledge (e.g., governing equations of dynamics) of structures into the feature learning process via a novel physics-based loss function. The input-output relationship of the network is constrained to retain its physical meaning implicitly while the demand for large amounts of labeled training data is reduced. Notably, even with only 5 % of the dataset used for training, PhyResNet achieves a 13.1 % improvement in R-Value. The performance of the proposed approach is evaluated through both numerical and experimental verifications. Results demonstrate that damage localization and quantification are achieved with high accuracies and good robustness.