Training Examples Research Articles

Coverage Estimation of Droplets Sprayed on Water-Sensitive Papers Based on Domain-Adaptive Segmentation

Unmanned aerial spraying systems (UASSs) are widely used today for the effective control of pests affecting crops, and more advanced UASS techniques are now being developed. To evaluate such systems, artificial targets are typically used to assess droplet coverage through image processing. To evaluate performance accurately, high-quality binary image processing is necessary; however, this involves labor for sample collection, transportation, and storage, as well as the risk of potential contamination during the process. Therefore, rapid assessment in the field is essential. In the present study, we evaluated droplet coverage on water-sensitive papers (WSPs) under field conditions. A dataset was constructed consisting of paired training examples, each comprising source and target data. The source data were high-quality labeled images obtained from WSP samples through image processing, while the target data were aligned RoIs within field images captured in situ. Droplet coverage estimation was performed using an encoder–decoder model, trained on the labeled images, with features adapted to field images via self-supervised learning. The results indicate that the proposed method detected droplet coverage in field images with an error of less than 5%, demonstrating a strong correlation between measured and estimated values (R2 = 0.99). The method proposed in this paper enables immediate and accurate evaluation of the performance of UASSs in situ.

Kinematic matrix: One-shot human action recognition using kinematic data structure

One-shot action recognition, which refers to recognizing human-performed actions using only a single training example, holds significant promise in advancing video analysis, particularly in domains requiring rapid adaptation to new actions. However, existing algorithms for one-shot action recognition face multiple challenges, including high computational complexity, limited accuracy, and difficulties in generalization to unseen actions. To address these issues, we propose a novel kinematic-based skeleton representation that effectively reduces computational demands while enhancing recognition performance. This representation leverages skeleton locations, velocities, and accelerations to formulate the one-shot action recognition task as a metric learning problem, where a model projects kinematic data into an embedding space. In this space, actions are distinguished based on Euclidean distances, facilitating efficient nearest-neighbour searches among activity reference samples. Our approach not only reduces computational complexity but also achieves higher accuracy and better generalization compared to existing methods. Specifically, our model achieved a validation accuracy of 78.5%, outperforming state-of-the-art methods by 8.66% under comparable training conditions. These findings underscore the potential of our method for practical applications in real-time action recognition systems.

Generative adversarial networks (GANs) to produce synthetic 12-lead electrocardiogram signals for specific and rare diseases: a novel, powerful tool towards clinical advancements

Abstract Background Remarkable technological advancements in electrocardiogram (ECG) analysis face challenges due to insufficiently diverse and extensive real-world datasets. This limitation hampers both the development of sophisticated deep learning (DL) models for ECG analysis and the exposure of new physicians to substantial training examples. Purpose To leverage Generative Adversarial Networks (GANs), a novel AI algorithm type, to generate synthetic yet realistic ECG signals resembling specific diseases. Methods We initially employed a general-purpose GAN on more than 21,000 ECG samples from the PTB-XL database, encompassing normal and abnormal ECGs across various diseases. The GAN outputs served as inputs for training "DeceptionECG", a novel GAN model inspired by the InceptionTime architecture (code available online). DeceptionECG focused on generating disease-specific ECG signals, trained on annotated ECGs validated by two electrophysiologists. Post-training, we generated synthetic ECGs for common diseases like atrial fibrillation (AF) and anterior ST-segment elevation myocardial infarction (STEMI), as well as rarer conditions like Wolf-Parkinson White (WPW) syndrome. Authentication involved an independent discriminator DL model and two cardiologists who were asked to distinguish real from synthetic ECGs. Additionally, a DL model, a variant of the InceptionTime architecture, discerned specific diseases in the ECG signals to assess their realism. Finally, we trained a DL model for detecting abnormalities in ECGs with the augmented dataset to evaluate potential performance improvement. Results We generated sets of 5,000 synthetic ECGs for various abnormalities, at a rate of 2,500 12-lead ECG signals per minute. The generated ECGs were indistinguishable from real ones, evidenced by low accuracy (Accuracy, ACC: 48.9%) of the discriminator model and two clinicians (ACC: 56% and 51%), on test sets with anterior STEMI. A DL model trained on real ECGs accurately differentiated between synthetic normal ECGs and those hosting the target disease in nearly 9 out of 10 cases (ACC: 88.7%). Furthermore, the employed DL model detector trained on the augmented dataset, encompassing both real (n=1,589) and synthetic (n=5,000) ECGs, outperformed its counterpart trained solely on real ECGs, when tested on a 1,000-sample set of real signals (ACC: 94.6% vs. 89.3%). Conclusion Our GAN-based approach showcases the potential of generating high-quality synthetic ECGs tailored for specific diseases, overcoming data scarcity challenges, and enhancing the training and performance of machine learning models. Additionally, it holds promise in advancing medical professional training by providing diverse signals with distinct "disease stamps" at massive amounts, addressing privacy concerns associated with extensive data exposure. Further testing is warranted for the applicability of the approach, especially in exceedingly rare conditions with limited available training data.Project workflow and outcomesGAN-generated ECGs: Which is fake?

Review of physics-informed machine-learning inversion of geophysical data

We review five types of physics-informed machine-learning (PIML) algorithms for inversion and modeling of geophysical data. Such algorithms use the combination of a data-driven machine-learning (ML) method and the equations of physics to model or invert geophysical data (or both). By incorporating the constraints of physics, PIML algorithms can effectively reduce the size of the solution space for ML models, enabling them to be trained on smaller data sets. This is especially advantageous in scenarios in which data availability may be limited or expensive to obtain. In this review, we restrict the physics to be that from the governing wave equation, either as a constraint that must be satisfied or by using numerical solutions of the wave equation for modeling and inversion. This approach ensures that the resulting models adhere to physical principles while leveraging the power of ML to analyze and interpret complex geophysical data. There are several potential benefits of PIML compared to standard numerical modeling or inversion of seismic data computed by, for example, finite-difference solutions to the wave equation. 1) Empirical tests suggest that PIML algorithms constrained by the physics of wave propagation can sometimes resist getting stuck in a local minima compared with standard full-waveform inversion (FWI). 2) After the weights of the neural network are found by training, then the forward and inverse operations by PIML can be more than several orders of magnitude more efficient than FWI. However, the computational cost for general training can be enormous. 3) If the ML inversion operator [Formula: see text] is locally trained on a small portion of the recorded data [Formula: see text], then there is sometimes no need for millions of training examples that aim for global generalization of [Formula: see text]. The benefit is that the locally trained [Formula: see text] can be used to economically invert the remaining test data [Formula: see text] for the true velocity [Formula: see text], where [Formula: see text] can comprise more than 90% of the recorded data.

Enhancing Activity Recognition After Stroke: Generative Adversarial Networks for Kinematic Data Augmentation.

The generalizability of machine learning (ML) models for wearable monitoring in stroke rehabilitation is often constrained by the limited scale and heterogeneity of available data. Data augmentation addresses this challenge by adding computationally derived data to real data to enrich the variability represented in the training set. Traditional augmentation methods, such as rotation, permutation, and time-warping, have shown some benefits in improving classifier performance, but often fail to produce realistic training examples. This study employs Conditional Generative Adversarial Networks (cGANs) to create synthetic kinematic data from a publicly available dataset, closely mimicking the experimentally measured reaching movements of stroke survivors. This approach not only captures the complex temporal dynamics and common movement patterns after stroke, but also significantly enhances the training dataset. By training deep learning models on both synthetic and experimental data, we enhanced task classification accuracy: models incorporating synthetic data attained an overall accuracy of 80.0%, significantly higher than the 66.1% seen in models trained solely with real data. These improvements allow for more precise task classification, offering clinicians the potential to monitor patient progress more accurately and tailor rehabilitation interventions more effectively.

Autoencoder-Driven Training Data Selection Based on Hidden Features for Improved Accuracy of ANN Short-Term Load Forecasting in ADMS

This paper presents a novel methodology for short-term load forecasting in the context of significant shifts in the daily load curve due to the rapid and extensive adoption of Distributed Energy Resources (DERs). The proposed solution, built upon the Similar Days Method (SDM) and Artificial Neural Network (ANN), introduces several novelties: (1) selection of similar days based on hidden representations of day data using Autoencoder (AE); (2) enhancement of model generalization by utilizing a broader set of training examples; (3) incorporating the relative importance of training examples derived from the similarity measure during training; and (4) mitigation of the influence of outliers by applying an ensemble of ANN models trained with different data splits. The presented AE configuration and procedure for selecting similar days generated a higher-quality training dataset, which led to more robust predictions by the ANN model for days with unexpected deviations. Experiments were conducted on actual load data from a Serbian electrical power system, and the results were compared to predictions obtained by the field-proven STLF tool. The experiments demonstrated an improved performance of the presented solution on test days when the existing STLF tool had poor predictions over the past year.

Classification and Comparison of Data Augmentation Techniques

Data augmentation is a technical method that generates additional data samples by applying different transformations and processes to the original training data. It aims to increase the size and diversity of the dataset, thus improving the models' generalization ability. This paper provides a detailed introduction to geometric transformations, data mixing, automated data augmentation methods, and generative adversarial methods. Geometric transformations manipulate the spatial configuration of data samples to create new variants that retain the original content. Data mixing techniques blend multiple data samples to produce novel training examples, promoting robustness and reducing overfitting. Automated data augmentation methods leverage search algorithms to discover optimal augmentation policies that improve model performance without manual intervention. Generative adversarial methods employ generative adversarial networks to synthesize realistic data samples that enrich the training dataset. These methods find broad application in domains including object detection, image classification, natural language processing, and speech recognition, significantly improving model performance. The paper explores the advantages and limitations of data augmentation techniques in practical applications and identifies future research directions and challenges. By systematically classifying and selecting appropriate data augmentation methods for specific application scenarios, this review provides a theoretical foundation and practical guidance for real-world applications.

EEG-based emotional valence and emotion regulation classification: a data-centric and explainable approach

Emotion classification using electroencephalographic (EEG) data is a challenging task in the field of Artificial Intelligence. While many researchers have focused on finding the best model or feature extraction technique to achieve optimal results, few have attempted to select the best methodological steps for working with the dataset. In this study, we applied two different theoretical approaches based on the noise of the dataset: curriculum learning and confident learning. Curriculum learning involves presenting training examples to the model in a specific order, starting with easier examples and gradually increasing in difficulty. This approach has been shown to improve model performance. Confident learning is a method for identifying and correcting label errors in datasets. By identifying and correcting these errors, confident learning can improve the performance of machine learning models trained on noisy datasets. We then applied the Integrated Gradient technique in order to assess the explainability of each model. Our aim was to explore the impact of different models and methods on emotion classification performance using EEG data. We collected and used an EEG dataset in which participants rated the emotional valence of positive and negative pictures while performing an emotion regulation (ER) task, comparing a control condition (Look) with two ER strategies: cognitive reappraisal and expressive suppression. We performed a multilabel classification to identify emotional neutrality or polarization of emotional valence (both positive and negative) rated by participants and the emotion regulation strategy adopted during the task. We compared the performance of models trained on three datasets selected based on label noise and evaluated their suitability for this task. Our results suggest different patterns based on the architecture used for feature importance, highlighting both advantages and criticisms.

ML-GAP: machine learning-enhanced genomic analysis pipeline using autoencoders and data augmentation.

The advent of RNA sequencing (RNA-Seq) has significantly advanced our understanding of the transcriptomic landscape, revealing intricate gene expression patterns across biological states and conditions. However, the complexity and volume of RNA-Seq data pose challenges in identifying differentially expressed genes (DEGs), critical for understanding the molecular basis of diseases like cancer. We introduce a novel Machine Learning-Enhanced Genomic Data Analysis Pipeline (ML-GAP) that incorporates autoencoders and innovative data augmentation strategies, notably the MixUp method, to overcome these challenges. By creating synthetic training examples through a linear combination of input pairs and their labels, MixUp significantly enhances the model's ability to generalize from the training data to unseen examples. Our results demonstrate the ML-GAP's superiority in accuracy, efficiency, and insights, particularly crediting the MixUp method for its substantial contribution to the pipeline's effectiveness, advancing greatly genomic data analysis and setting a new standard in the field. This, in turn, suggests that ML-GAP has the potential to perform more accurate detection of DEGs but also offers new avenues for therapeutic intervention and research. By integrating explainable artificial intelligence (XAI) techniques, ML-GAP ensures a transparent and interpretable analysis, highlighting the significance of identified genetic markers.

AI-accelerated Nazca survey nearly doubles the number of known figurative geoglyphs and sheds light on their purpose

It took nearly a century to discover a total of 430 figurative Nazca geoglyphs, which offer significant insights into the ancient cultures at the Nazca Pampa. Here, we report the deployment of an AI system to the entire Nazca region, a UNESCO World Heritage site, leading to the discovery of 303 new figurative geoglyphs within only 6 mo of field survey, nearly doubling the number of known figurative geoglyphs. Even with limited training examples, the developed AI approach is demonstrated to be effective in detecting the smaller relief-type geoglyphs, which unlike the giant line-type geoglyphs are very difficult to discern. The improved account of figurative geoglyphs enables us to analyze their motifs and distribution across the Nazca Pampa. We find that relief-type geoglyphs depict mainly human motifs or motifs of things modified by humans, such as domesticated animals and decapitated heads (81.6%). They are typically located within viewing distance (on average 43 m) of ancient trails that crisscross the Nazca Pampa and were most likely built and viewed at the individual or small-group level. On the other hand, the giant line-type figurative geoglyphs mainly depict wild animals (64%). They are found an average of 34 m from the elaborate linear/trapezoidal network of geoglyphs, which suggests that they were probably built and used on a community level for ritual activities.

Public Health Security Systems Empowered by Artificial Intelligence for Early Monitoring and Prevention of Epidemics

Artificial intelligence (AI) methods have been extensively used for detecting and predicting Infectious Disease (ID) outbreaks as time series and modeling and evaluating Public Health responses. The significant tasks of PH monitoring and intervention present distinct technical difficulties, including limited data availability, absence of sufficient positive training examples, challenges in establishing benchmarks, measuring the effectiveness of management policies, complex relationships between spatial and time series elements, and more detailed risk assessments involving interaction and social networks. Conventional PH monitoring mainly depends on statistical methods. In recent years, there has been a significant expansion of approaches that AI enables. This research presents an AI method called Early Monitoring and Prevention of Epidemics (AI-EMPE) for enhancing PH security systems. The suggested approach converts a substantial amount of collected PH security incidents into separate incident characteristics and utilizes a Deep Learning (DL)--based detection technique to enhance EMPE. AI-EMPE incorporates sophisticated AI methods, including integrated Convolutional Neural Networks (CNN) and Backpropagation Neural Networks (BP-NN). These findings indicate that the integrated method is the most efficient in improving PH security systems.

LoGAN: local generative adversarial network for novel structure prediction

Abstract The efficient generation and filtering of candidate structures for new materials is becoming increasingly important as starting points for computational studies. In this work, we introduce an approach to Wasserstein generative adversarial networks for predicting unique crystal and molecular structures. Leveraging translation- and rotation-invariant atom-centered local descriptors addresses some of the major challenges faced by similar methods. Our models require only small sets of known structures as training data. Furthermore, the approach is able to generate both non-periodic and periodic structures based on local coordination. We showcase the data efficiency and versatility of the approach by recovering all stable C5H12O isomers using only 39 C4H10O and C6H14O training examples, as well as a few randomly selected known low-energy SiO2 crystal structures utilizing only 167 training examples of other SiO2 crystal structures. We also introduce a filtration technique to reduce the computational cost of subsequent characterization steps by selecting samples from unique basins on the potential energy surface, which allows to minimize the number of geometry relaxations needed after structure generation. The present method thus represents a new, versatile approach to generative modeling of crystal and molecular structures in the low-data regime, and is available as open-source.

Distribution Matching for Machine Teaching.

Machine teaching is an inverse problem of machine learning that aims at steering the student toward its target hypothesis, in which the teacher has already known the student's learning parameters. Previous studies on machine teaching focused on balancing the teaching risk and cost to find the best teaching examples deriving from the student model. This optimization solver is in general ineffective when the student does not disclose any cue of the learning parameters. To supervise such a teaching scenario, this article presents a distribution matching-based machine teaching strategy via iteratively shrinking the teaching cost in a smooth surrogate, which eliminates boundary perturbations from the version space. Technically, our strategy could be redefined as a cost-controlled optimization process that finds the optimal teaching examples without further exploring the parameter distribution of the student. Then, given any limited teaching cost, the training examples would have a closed-form expression. Theoretical analysis and experiment results demonstrate the effectiveness of this strategy.

Predicting peritumoral glioblastoma infiltration and subsequent recurrence using deep-learning-based analysis of multi-parametric magnetic resonance imaging.

Glioblastoma (GBM) is the most common and aggressive primary adult brain tumor. The standard treatment approach is surgical resection to target the enhancing tumor mass, followed by adjuvant chemoradiotherapy. However, malignant cells often extend beyond the enhancing tumor boundaries and infiltrate the peritumoral edema. Traditional supervised machine learning techniques hold potential in predicting tumor infiltration extent but are hindered by the extensive resources needed to generate expertly delineated regions of interest (ROIs) for training models on tissue most and least likely to be infiltrated. We developed a method combining expert knowledge and training-based data augmentation to automatically generate numerous training examples, enhancing the accuracy of our model for predicting tumor infiltration through predictive maps. Such maps can be used for targeted supra-total surgical resection and other therapies that might benefit from intensive yet well-targeted treatment of infiltrated tissue. We apply our method to preoperative multi-parametric magnetic resonance imaging (mpMRI) scans from a subset of 229 patients of a multi-institutional consortium (Radiomics Signatures for Precision Diagnostics) and test the model on subsequent scans with pathology-proven recurrence. Leave-one-site-out cross-validation was used to train and evaluate the tumor infiltration prediction model using initial pre-surgical scans, comparing the generated prediction maps with follow-up mpMRI scans confirming recurrence through post-resection tissue analysis. Performance was measured by voxel-wised odds ratios (ORs) across six institutions: University of Pennsylvania (OR: 9.97), Ohio State University (OR: 14.03), Case Western Reserve University (OR: 8.13), New York University (OR: 16.43), Thomas Jefferson University (OR: 8.22), and Rio Hortega (OR: 19.48). The proposed model demonstrates that mpMRI analysis using deep learning can predict infiltration in the peri-tumoral brain region for GBM patients without needing to train a model using expert ROI drawings. Results for each institution demonstrate the model's generalizability and reproducibility.

Good results from sensor data: Performance of machine learning algorithms for regression problems in chemical sensors

Accurately predicting unseen data, instead of mere memorization of training examples, is a critical goal of machine learning. This generalization is particularly important in the field of chemical sensors, where the ability to accurately predict the chemical properties or concentration levels of unknown samples is crucial. The paper presents a comprehensive yet accessible introduction to various machine learning concepts, highlighting the importance of model interpretability and generalization in ensuring reliable and accurate results in this context. Nonlinear sensor array data are utilized to introduce key concepts (e.g., bias-variance tradeoff) and techniques (linear models, partial least squares regression, support vector machines, k-nearest neighbors, decision trees, ensemble methods, automated machine learning, symbolic regression, and artificial neural networks), providing a solid foundation to make informed decisions when selecting machine learning techniques for sensor-specific regression applications. The results clearly indicate a number of conclusions. First, overparameterized deep feedforward neural networks show great accuracy and generalization when trained on a sufficiently large dataset. Second, symbolic regression models proved to be more accurate than deep feedforward neural networks and classical machine learning techniques on smaller datasets. Third, the performance of various machine learning models was dataset-dependent, showing the importance of comparative studies to determine the most suitable approach. It is clear that the optimal model cannot be known a priori. This paper aims to provide a starting point for investigations on the performance of different machine learning techniques in chemical sensor applications.

Automated peripheral nerve segmentation for MR-neurography

BackgroundMagnetic resonance neurography (MRN) is increasingly used as a diagnostic tool for peripheral neuropathies. Quantitative measures enhance MRN interpretation but require nerve segmentation which is time-consuming and error-prone and has not become clinical routine. In this study, we applied neural networks for the automated segmentation of peripheral nerves.MethodsA neural segmentation network was trained to segment the sciatic nerve and its proximal branches on the MRN scans of the right and left upper leg of 35 healthy individuals, resulting in 70 training examples, via 5-fold cross-validation (CV). The model performance was evaluated on an independent test set of one-sided MRN scans of 60 healthy individuals.ResultsMean Dice similarity coefficient (DSC) in CV was 0.892 (95% confidence interval [CI]: 0.888–0.897) with a mean Jaccard index (JI) of 0.806 (95% CI: 0.799–0.814) and mean Hausdorff distance (HD) of 2.146 (95% CI: 2.184–2.208). For the independent test set, DSC and JI were lower while HD was higher, with a mean DSC of 0.789 (95% CI: 0.760–0.815), mean JI of 0.672 (95% CI: 0.642–0.699), and mean HD of 2.118 (95% CI: 2.047–2.190).ConclusionThe deep learning-based segmentation model showed a good performance for the task of nerve segmentation. Future work will focus on extending training data and including individuals with peripheral neuropathies in training to enable advanced peripheral nerve disease characterization.Relevance statementThe results will serve as a baseline to build upon while developing an automated quantitative MRN feature analysis framework for application in routine reading of MRN examinations.Key PointsQuantitative measures enhance MRN interpretation, requiring complex and challenging nerve segmentation.We present a deep learning-based segmentation model with good performance.Our results may serve as a baseline for clinical automated quantitative MRN segmentation.Graphical

Comparing a Large Language Model with Previous Deep Learning Models on Named Entity Recognition of Adverse Drug Events.

The ability to fine-tune pre-trained deep learning models to learn how to process a downstream task using a large training set allow to significantly improve performances of named entity recognition. Large language models are recent models based on the Transformers architecture that may be conditioned on a new task with in-context learning, by providing a series of instructions or prompt. These models only require few examples and such approach is defined as few shot learning. Our objective was to compare performances of named entity recognition of adverse drug events between state of the art deep learning models fine-tuned on Pubmed abstracts and a large language model using few-shot learning. Hussain et al's state of the art model (PMID: 34422092) significantly outperformed the ChatGPT-3.5 model (F1-Score: 97.6% vs 86.0%). Few-shot learning is a convenient way to perform named entity recognition when training examples are rare, but performances are still inferior to those of a deep learning model fine-tuned with several training examples. Perspectives are to evaluate few-shot prompting with GPT-4 and perform fine-tuning on GPT-3.5.

Predicting 30-day mortality in severely injured elderly patients with trauma in Korea using machine learning algorithms: a retrospective study.

The number of elderly patients with trauma is increasing; therefore, precise models are necessary to estimate the mortality risk of elderly patients with trauma for informed clinical decision-making. This study aimed to develop machine learning based predictive models that predict 30-day mortality in severely injured elderly patients with trauma and to compare the predictive performance of various machine learning models. This study targeted patients aged ≥65 years with an Injury Severity Score of ≥15 who visited the regional trauma center at Chungbuk National University Hospital between 2016 and 2022. Four machine learning models-logistic regression, decision tree, random forest, and eXtreme Gradient Boosting (XGBoost)-were developed to predict 30-day mortality. The models' performance was compared using metrics such as area under the receiver operating characteristic curve (AUC), accuracy, precision, recall, specificity, F1 score, as well as Shapley Additive Explanations (SHAP) values and learning curves. The performance evaluation of the machine learning models for predicting mortality in severely injured elderly patients with trauma showed AUC values for logistic regression, decision tree, random forest, and XGBoost of 0.938, 0.863, 0.919, and 0.934, respectively. Among the four models, XGBoost demonstrated superior accuracy, precision, recall, specificity, and F1 score of 0.91, 0.72, 0.86, 0.92, and 0.78, respectively. Analysis of important features of XGBoost using SHAP revealed associations such as a high Glasgow Coma Scale negatively impacting mortality probability, while higher counts of transfused red blood cells were positively correlated with mortality probability. The learning curves indicated increased generalization and robustness as training examples increased. We showed that machine learning models, especially XGBoost, can be used to predict 30-day mortality in severely injured elderly patients with trauma. Prognostic tools utilizing these models are helpful for physicians to evaluate the risk of mortality in elderly patients with severe trauma.

Optimizing SEM parameters for segmentation with AI – Part 1: Generating a training set

Extracting significant quantitative results from SEM images requires feature segmentation with image processing software. The efficiency of segmentation algorithms depends on the image quality, determined by the parameters set on the microscope during acquisitions. By integrating AI within SEM acquisition workflows, it is possible to suggest microscope parameters that will produce images where the features to quantify will be easily segmented. Specifically, a model is trained to automatically suggest the beam energy and probe current to set on the microscope during acquisitions. This paper is the first of two parts, describing workflows for generating a complete training set. The training set is carefully designed, consisting of both simulated data and real data acquired on the SEM by varying the energy and current. Separate workflows are required for generating simulated and acquired training examples. Simulated data generation is accomplished with the MC X-ray simulator in Dragonfly, where multiple virtual samples are created to intentionally diversify the training set. Acquiring data on the SEM for training is a time-consuming process when compared to generating simulations and would ideally be avoided but is included here to determine the degree to which it is required. Using only simulated data for adequate training, we show that our data generation workflow can be fully automated and produces a considerable amount of high quality data rapidly and with minimal effort.

Partial label learning via identifying outlier features

Partial label learning (PLL) is a classification problem where each training instance is ambiguously annotated with a set of candidate labels, among which only one is the ground truth label. The topological structure in the feature space is frequently used by existing PLL approaches to clarify if a candidate label is the ground truth label for a training example. However, these techniques frequently experience cumulative errors brought on by the error-prone labeling confidence estimate throughout the topological structure because of the redundant and noisy features occurring naturally in the feature space. In this paper, we propose a novel approach through Identifying Outlier Features (PL-IOF) to address this challenge. The feature space is decomposed into class prototype features and outlier features, where the former captures the discriminative features of each class prototype, and the latter captures the outlier features in each training instance. A unified framework is specifically suggested to simultaneously optimize class prototypes and outlier features, as well as to estimate labeling confidences over incomplete label training instances. This framework ensures the high quality of the extracted class prototypes by recognizing the outliers. Experiments on both synthetic and real-world datasets demonstrate the out-performance of PL-IOF over the state-of-the-art.