Search

• We developed a ML model to recommend first FSH dosage for all types of patients. • The model performance surpassed the clinicians' in both development and validation. • The model can serve as quality check, second opinion or learning tool for trainees. Is it possible to identify accurately the optimal first dose of FSH in ovarian stimulation by means of a machine learning model? Observational study (2011–2021) including first IVF cycles with own oocytes. A total of 2713 patients from five private reproductive centres were included in the development phase (2011–2019) and 774 in the validation phase (2020–2021). Predictor variables included age, BMI, AMH, AFC and previous live births. Performance was measured with a proposed score based on the number of MII oocytes retrieved and dose received, recommended, or both. The included cycles were from women aged 37.7 ± 4.4 years (18–45 years), with a BMI of 23.5 ± 4.2 kg/m2, AMH of 2.4 ± 2.3 ng/ml, AFC of 11.3 ± 7.6, and an average number of MII obtained 6.9 ± 5.4. The model reached a mean performance score of 0.87 (95% CI 0.86 to 0.88) in the development phase, significantly better than for doses prescribed by clinicians for the same patients (0.83, 95% CI 0.82 to 0.84; P = 2.44 e-10). Mean performance score of the model recommendations was 0.89 (95% CI 0.88 to 0.90) in the validation phase, also significantly better than clinicians (0.84, 95% CI 0.82 to 0.86; P = 3.81 e-05). The model was shown to surpass the performance of standard practice. This machine learning model could be used as a training and learning tool for new clinicians, and as quality control for experienced clinicians. [ABSTRACT FROM AUTHOR]

Foreign species can deteriorate the environment and the economy of a country. To automatically monitor biosecurity threats at country borders, this paper investigates compact deep networks for accurate and real-time object segmentation for hyperspectral images. To this end, knowledge distillation (KD) approaches compress the model by distilling the knowledge of a large teacher network to a compact student network. However, when the student is over-compressed, the performance of standard KD methods degrades significantly due to the large capacity gap between the teacher and the student. This gap can be addressed by adding medium-sized teacher assistants, but training them incurs significant computation and hence is impractical. To address this problem, this paper proposes a new framework called Knowledge Distillation from Multi-head Teacher (KDM), which distills the knowledge of a multi-head teacher to the student. By encapsulating multiple teachers in a single network, our proposed KDM assists the learning of a very compact student and significantly reduces the training time. We also introduce Bio-HSI, a new large benchmark hyperspectral image dataset of 3,125 high-resolution images with dense segmentation ground truth. This new, large dataset can be expected to advance research on deep models for hyperspectral image segmentation. Evaluated on this dataset, the student trained via our KDM has 762 times fewer parameters than the state-of-the-art segmentation model (i.e., HRNet), while achieving competitive accuracy. [ABSTRACT FROM AUTHOR]

For successful biomarker discovery, it is essential to develop computational frameworks that summarize high-dimensional neuroimaging data in terms of involved sub-systems of the brain, while also revealing underlying heterogeneous functional and structural changes covarying with specific cognitive and biological traits. However, unsupervised decompositions do not inculcate clinical assessment information, while supervised approaches extract only individual feature importance, thereby impeding qualitative interpretation at the level of subspaces. We present a novel framework to extract robust multimodal brain subspaces associated with changes in a given cognitive or biological trait. Our approach involves active subspace learning on the gradients of a trained machine learning model followed by clustering to extract and summarize the most salient and consistent subspaces associated with the target variable. Through a rigorous cross-validation procedure on an Alzheimer's disease (AD) dataset, our framework successfully extracts multimodal subspaces specific to a given clinical assessment (e.g., memory and other cognitive skills), and also retains predictive performance in standard machine learning algorithms. We also show that the salient active subspace directions occur consistently across randomly sub-sampled repetitions of the analysis. Compared to existing unsupervised decompositions based on principle component analysis, the subspace components in our framework retain higher predictive information. As an important step towards biomarker discovery, our framework not only uncovers AD-related brain regions in the associated brain subspaces, but also enables automated identification of multiple underlying structural and functional sub-systems of the brain that collectively characterize changes in memory and proficiency in cognitive skills related to brain disorders like AD. • A method to detect multiple underlying subspaces in the brain associated with a trait. • Multimodal active subspaces fuse structural and functional features of the brain. • Active subspaces retain predictive information for brain disorders. • Experiments with Alzheimer's related scores show successful subspace detection. [ABSTRACT FROM AUTHOR]

Learning using privileged information (LUPI) has shown promise in improving supervised learning by embedding additional knowledge. However, its reliance on the assumption of readily available privileged information may not hold true in practical scenarios due to limitations in access or confidentiality. To address these challenges, this paper presents a novel weakly privileged learning (WPL) framework, integrating knowledge extraction methods within the LUPI context. An effective strategy is proposed to implement the WPL framework, where knowledge extraction techniques generate a weight matrix as weak privileged information. Extensive experiments employing various existing knowledge extraction techniques demonstrate that the proposed WPL outperforms traditional supervised learning and approaches the performance of standard privileged learning where privileged information is given in advance. This research establishes WPL as a promising learning paradigm, addressing limitations in privileged information availability and advancing the field of machine learning in practical settings. • Introduce a novel weakly privileged learning framework (WPL). • Knowledge extraction methods are used to generate weak privileged information. • Extensive experiments verify the superiority of the proposed methods. [ABSTRACT FROM AUTHOR]

We propose a novel spatial bagging workflow for predictive ensemble machine learning that improves on standard bagging models. Our proposed method integrates spatial bootstrap for bagging with the number of effective sample size, n e f f , for integration of the spatial context of the dataset. We benchmark the improved performance over standard machine learning bagging models with a large number of two-dimensional synthetic datasets with varying degrees of Gaussian noise. For noise free datasets, both methods demonstrate equivalent accuracy; however, spatial bagging achieves this with a significantly smaller sample size, showcasing its improved efficiency. As data noise increases, spatial bagging consistently outperforms standard bagging, displaying an improved Mean Squared Error (MSE) and robustness against overfitting. Our proposed spatial bagging method computes the optimal effective sample size for spatial data, reducing model overfitting. Furthermore, our proposed method requires only the additional step of variogram calculation and modeling, and can be implemented with any predictive machine learning bagging model with minimal code modification. i.e., specification of the number of bootstrap samples as the number of effective data. We recommend using spatial bagging for improved predictions for any spatial data setting across diverse scientific fields, e.g., atmospheric, agricultural, subsurface resources etc. • Ensemble learning methods employ standard bagging, which is prone to overfitting. • We introduce a novel spatial bagging technique utilizing effective sample size. • Spatial bagging mitigates overfitting, enhancing robustness against noise. • Spatial bagging provides a computational advantage by using smaller data. [ABSTRACT FROM AUTHOR]

Due to the enormous amount of user-generated content being generated on the web, labeling such data is a time-consuming and expensive endeavor. As a result, we have limited annotated data and the vast majority of data are unlabeled. Analysis reveals that extracting (external) knowledge from unlabeled data and integrating it with knowledge extracted from labeled data is a beneficial task for text information processing, in particular text classification. In this paper, we present a hybrid approach for classifying sentiments that employs external knowledge, which is categorized as either general-purpose sentiment knowledge or domain-related knowledge. General-purpose sentiment knowledge is extracted from sentiment lexicons, whereas domain-related knowledge is extracted from unlabeled data from the same or related domains. Similar domains for a given domain are identified based on their similarity score in terms of overlapping features. The proposed approach utilizes both forms of external knowledge and combines them with logistic regression to train an improved classification model. The classification model uses the conventional gradient descent algorithm for optimization, and its convergence analysis indicates that it is convex and converges to the global optimum. The proposed approach is empirically evaluated and compared to three baselines and one state-of-the-art method using standard performance evaluation metrics on a multi-domain sentiment dataset. The experiment results are encouraging, demonstrating that the proposed approach considerably outperforms the baseline approaches and outperforms the state-of-the-art approach by up to 2% in terms of both f-score and accuracy. • A hybrid approach for document-level sentiment classification that makes use of external knowledge. • A domain-specific knowledge extraction approach from unlabeled text documents. • A detailed convergence analysis of the proposed optimization model used in the proposed approach. [ABSTRACT FROM AUTHOR]

Multi-dimensional classification (MDC) task can be considered the most inclusive description of all classifications tasks, as it joins multiple class spaces and their multiple class members into a single compound classification problem. The challenges in MDC arise from the possible class dependencies across different class spaces, as well as the imbalance of labels in training datasets due to lack of all possible combinations. In this paper, we propose a straightforward, yet efficient, MDC deep learning classifier, named Deep Self-Organizing Cube (DSOC) that can model dependencies among classes in multiple class spaces, while consolidating its ability to classify rare combinations of labels. DSOC is formed of two n -dimensional components, namely the Hypercube Classifier and the multiple DSOC Neural Networks connected to the hypercube. The multiple neural networks component is responsible for feature selection and segregation of classes, while the Hypercube classifier is responsible for creating the semantics among multiple class spaces and accommodate the model for rare sample classification. DSOC is a multiple-output learning algorithm that successfully classify samples across all class spaces simultaneously. To challenge the proposed DSOC model, we conducted an assessment on seventeen benchmark datasets in the four types of classification tasks, binary, multi-class, multi-label and multi-dimensional. The obtained results were compared to four standard classifiers and eight competitive state-of-the-art approaches reported in literature. The DSOC has achieved superior performance over standard classifiers as well as the state-of-the-art approaches in all the four classification tasks. Moreover, in terms of Exact Match accuracy metrics, DSOC has outperformed all state-of-the-art approaches in 77.8% of the cases, which reflects the superior ability of DSOC to model dependencies and successfully classify rare samples across all dimensions simultaneously. [ABSTRACT FROM AUTHOR]

The absorption and scattering properties of the water medium cause various types of distortion in underwater images, which seriously affects the accuracy and effectiveness of subsequent processing. The application of supervised learning algorithms in underwater image enhancement is limited by the difficulty of obtaining a large number of underwater paired images in practical applications. As a solution, we propose an unsupervised representation disentanglement based underwater image enhancement method (URD-UIE). URD-UIE disentangles content information (e.g., texture, semantics) and style information (e.g., chromatic aberration, blur, noise, and clarity) from underwater images and then employs the disentangled information to generate the target distortion-free image. Our proposed method URD-UIE adopts an unsupervised cycle-consistent adversarial translation architecture and combines multiple loss functions to impose specific constraints on the output results of each module to ensure the structural consistency of underwater images before and after enhancement. The experimental results demonstrate that the URD-UIE technique effectively enhances the quality of underwater images when training with unpaired data, resulting in a significant improvement in the performance of the standard model for underwater object detection and semantic segmentation. • URD-UIE accomplishes underwater image decomposition: domain-invariant content & domain-specific style. • We use disentangled cycle translation for structural consistency & diverse styles in recovered images. • Our method employs diverse loss functions ensuring realistic enhanced images via disentangled features. • URD-UIE trains unsupervised, providing valuable improvements for subsequent underwater tasks. [Display omitted] [ABSTRACT FROM AUTHOR]

Monitoring hydrogen refueling stations (HRS) ensures the safety of their operations as well as meeting the required performance standard for refueling. Traditionally, table-based or dynamic control of fueling process methods were used to guide the fueling process. We propose to use machine learning methods to predict the main performance target of the fueling process; the state of charge (SOC). Based on initial operating conditions in start-up fueling time, least absolute shrinkage and selection operator (LASSO) is used to predict non-dispensing fueling events, normal, and undershot dispensing events. The computational experiments were conducted on three hydrogen refueling stations with up to two-years of operational data. The classification accuracy reached over 85% on three stations. The SOC in future time steps were predicted using a multi-task regression model. The predicted future SOC within one minute is accurate on two data-rich stations. In both classification and regression tasks, the training models built on data-rich stations greatly improved the performance of a station with much less training data. These results show using machine learning methods on predicting the performance of fueling events is feasible and satisfactory on HRS. To optimize operation strategies, machine learning models can be integrated into the controller of HRS as an alternative way of controlling the fueling process. Image 1 • Machine learning methods for classification and prediction of the fueling process. • Classifying fueling events reached over 85% accuracy on three stations. • Multi-task machine learning models predicted state of charge accurately. • Models from data-rich stations improved the prediction on the data-poor station. [ABSTRACT FROM AUTHOR]

Deep neural networks are widely used to solve several difficult machine learning tasks due to their impressive performance on standard benchmarking datasets. Most of the state of the art neural architectures contain a staggering amount of parameters and have many layers. Thus, they are computationally intensive and memory demanding models. This fact prohibits their deployment in devices with limited computational resources such as smartphones and unmanned vehicles. Recently, a growing interest in developing methods that can compress and accelerate these networks emerged. In this paper, we propose the use of complete vector quantization for neural model compression and acceleration. More specifically, we show that it is possible to use product quantization with common subdictionaries to quantize both the parameters and the activations of the neural network without compromising significantly the network accuracy. The proposed method removes the need for multiplications in order to compute the neural preactivations and provides opportunities for acceleration using lookup tables. [ABSTRACT FROM AUTHOR]

[Display omitted] • Tensile behaviour of coir-based composites was analysed and modelled. • Coir fibers enhanced the properties of composites at higher aspect ratio. • Neural network regression model predicted the tensile properties with high accuracy. • Predicted and experimental tensile properties were in good agreement. • The developed framework provides insights into designing eco-friendly composites. Coconut husk-derived coir fiber composites, an eco-friendly alternative to conventional polymer composites, are being considered as a viable substitute for non-structural applications. The current research endeavor aims to develop a framework for analyzing and predicting the tensile properties of short coir-based composites (SCBCs) using neural networks. The effect of parameters such as fiber treatment, fiber content, fiber aspect ratio, manufacturing process, and matrix type on the tensile properties of SCBCs has been considered. The standard performance metrics assessed the feasibility of the developed model. The coefficient of determination (R2) of 0.9661 and 0.9816 was obtained for the tensile strength and tensile modulus prediction model, respectively. [ABSTRACT FROM AUTHOR]

The Internet of Things (IoT) has grown rapidly to become the core of many areas of application, leading to the integration of sensors, with IoT devices. However, the number of attacks against these types of devices has grown as fast as the paradigm itself. Certain inherent characteristics of the paradigm, as well as the limited computational capabilities of the devices involved, make it difficult to deploy security measures. This is why it is necessary to design, implement and study new solutions in the field of cybersecurity. In this paper, we propose an architecture that is capable of generating Complex Event Processing (CEP) rules automatically by integrating them with machine learning technologies. While the former is used to automatically detect attack patterns in real time, the latter, through the use of the Principal Component Analysis (PCA) algorithm, allows the characterization of events and the recognition of anomalies. This combination makes it possible to achieve efficient CEP rules at the computational level, with the results showing that the CEP rules obtained using our approach substantially improve upon the performance of the standard CEP rules, which are rules that are not generated by our proposal but can be defined independently by an expert in the field. Our proposal has achieved an F1-score of 0.98 on average, a 76 percent improvement in throughput over standard CEP rules, and a reduction in the network overhead of 86 percent over standard simple events, which are the simple events that are generated when our proposal is not used. [ABSTRACT FROM AUTHOR]

There have been research attempts to automate the process of lithology identification from the well-logs using machine learning algorithms. However, a single algorithm cannot be applied on a universal scale due to the nonstationary nature of geological data sets. The notion of changes in the well-log signal must be considered to resolve the issue. This paper presents a multistage method using multivariate change detection and statistical machine learning to (a) identify the significant changes in the well-logs, (b) map multivariate change in log signal with the lithofacies classes, (c) enhance the data set by removing the highly populated class, (d) using statistical machine learning tools, i.e., Support Vector Machine (SVM) to classify target classes. We apply the algorithm to the data collected from a mostly shale-contained subsurface environment in the Krishna-Godavari (KG) basin, India. The primary objective of the research is to locate thin layers of hydrocarbon-contained sand in the subsurface while dealing with challenges such as intra-well and inter-well heterogeneity and a severe class imbalance. The algorithm satisfactorily performs in lithofacies classification with significantly better performance than the standard machine learning algorithm. • Lithofacies identification using a multi-stage change detection approach • Lithofacies identification in a non-stationary subsurface • Lithofacies classification in an unbalanced condition [ABSTRACT FROM AUTHOR]

• Face expression analysis suffers from insufficiently annotated data. • We introduce the timid semi-supervised strategy based on diversity for face expression analysis. • The solution produces superior results in both expression recognition and action unit estimation. • Simple solutions based on diversity perform better when biased unlabelled data is available. [Display omitted] In the last years, semi–supervised learning has been proposed as a strategy with high potential for improving machine learning capabilities. Face expression recognition may highly benefit from such a technique, as accurate labeling is both difficult and costly, whereas millions of unlabeled images with human faces are available on the Internet, but without annotations. In this paper we evaluate the benefits of semi–supervised learning in the practical scenarios of face expression analysis. Our conclusion is that better performance is indeed achievable, but by methods that put a distinct emphasis on the diversity of exploring patterns in the unlabeled data domain. The evaluation is carried on multiple tasks such as detecting Action Units on EmotioNet, assessing Action Units intensity on the spontaneous DISFA database and, respectively, recognizing expressions on static images acquired in the wild, from the RAF-DB and FER+ databases. We show that, in these scenarios, a so–called timid semi–supervised learner is more robust and achieves higher performance than standard, confident semi–supervised learners. [ABSTRACT FROM AUTHOR]

Obesity, termed as excessive body fat, is a major public health problem. Being able to accurately predict body fat percentage provides an opportunity for making appropriate prevention and diagnosis. Although various assessment techniques have been applied for body fat estimation, high-cost tests are required using special equipment, and the performance of standard machine learning models are typically impacted by noise data. In this paper, we propose a novel fuzzy-weighted approach for improving body fat prediction using easily accessed body measurements in a cost-effective manner. Given the noise data problem, we design a novel fuzzy weighted operation by extending the minimum and maximum values for calculating the fuzzy weights to cover more possibilities. The fuzzy weights are incorporated into the error constraints of a relative error support vector machine model to improve its robustness to noise data. In addition, the Whale Optimization Algorithm (WOA) is applied in our proposed model for parameter optimization. Our proposed fuzzy-weighted approach delivers excellent results for body fat prediction by outperforming other well-known machine learning models. Further, the statistical test based on the 20 runs of experiments not only confirms that our proposed model could significantly outperform the models under comparison, but also show the fuzzy weighted operation and WOA are effective to improve the proposed approach. This offers a viable alternative for medical decision makers to make appropriate prevention, early diagnosis, and early treatment of body fat related diseases when high-cost measurement techniques are not available. • A novel fuzzy-weighted approach is proposed for body fat prediction. • The whale optimization algorithm is applied for parameter optimization. • Evaluation metrics indicate the effectiveness of our proposed approach. • Our approach can be a useful tool for medical decision makers. [ABSTRACT FROM AUTHOR]

The relationship between performance and experience is non-linear, thus planning models that seek to manage workforce development through task assignment are difficult to solve. This gets even more complicated when taking into account multi-skilled workers that are capable of performing a variety of tasks. In this paper we develop a competences-based analytical model of the performance of multi-skilled workers undertaking repetitive tasks, taking into account learning and forgetting. A learning curve can be used to estimate improvement when repeating the same operation. Inverse phenomenon is forgetting, which can occur due to interruption in the production process. The Performance Evaluation Algorithm (PEA) was developed for two cases: fixed shift duration and fixed production output. The aim was to build a tool that better describes the capabilities of workers to perform repetitive tasks by binding together hierarchical competences modeled as a weighted digraph together with a learning and forgetting curve model (LFCM) to express individual learning rates. [ABSTRACT FROM AUTHOR]

Learning how to automatically solve optimization problems has the potential to provide the next big leap in optimization technology. The performance of automatically learned heuristics on routing problems has been steadily improving in recent years, but approaches based purely on machine learning are still outperformed by state-of-the-art optimization methods. To close this performance gap, we propose a novel large neighborhood search (LNS) framework for vehicle routing that integrates learned heuristics for generating new solutions. The learning mechanism is based on a deep neural network with an attention mechanism and has been especially designed to be integrated into an LNS search setting. We evaluate our approach on the capacitated vehicle routing problem (CVRP), the split delivery vehicle routing problem (SDVRP), and the capacitated team orienteering problem (CTOP). We show that the NLNS approach is able to outperform a handcrafted LNS on the CVRP and SDVRP and match the performance of a standard LNS on the CTOP. NLNS is thus able to quickly and effectively learn high performance heuristics to maneuver through the search space of difficult routing problems, coming close to the performance of state-of-the-art optimization approaches. [ABSTRACT FROM AUTHOR]

This paper presents a novel hierarchical correlated Q-learning (HCEQ) algorithm to solve the dynamic optimization of generation command dispatch (GCD) in the Automatic Generation Control (AGC). The GCD problem is to dynamically allocate the total AGC generation command from the central to each individual AGC generator. The proposed HCEQ is a novel multi-agent Q-learning algorithm based on the concept of correlated equilibrium point, and each AGC generator with an agent is to optimize its regulation participation factor and coordinate its decision with others for the overall GCD performance enhancement. In order to cope with the curse of dimensionality in the GCD problem with the increased number of AGC plants involved, a multi-layer optimum GCD framework is developed in this paper. In this hierarchical framework, the multiobjective design and a time-varying coordination factor have been formulated into the reward functions to improve the optimization efficiency and convergence of HCEQ. The application of the proposed approach has been fully verified on the China southern power grid (CSG) model to demonstrate its superior performance and dynamic optimization capability in various power system scenarios. [ABSTRACT FROM AUTHOR]

For an efficient detection of intrusion in the network, the system requires sensitive information. This information contains various features depending upon the network. The selection of useful features and relevant information from the dataset has always been difficult for the user. These features bound the performance of the system and also restrict them in making accurate decisions. This irrelevant feature plays a major role in making confusion between the normal data pattern against the attack data pattern. This article proposes a novel Maximum correlation-based mutual information technique for efficient feature selection (MCMIFS) in a data network. The proposed method is utilized with Kernel Extreme Learning Machine (KELM) based multiclass classifier for effective intrusion detection. The performance of the proposed scheme is evaluated using standard intrusion detection datasets with feature ranging from 42 to 155. The hybrid of MCMIFS with KELM reveals the effectiveness of the proposed scheme by improving the accuracy of detection, i.e., 99.97%, decreasing the FPR, i.e., 0.19, by decreasing the computational complexity. The proposed Intrusion Detection Scheme (IDS) results are analyzed and compared with the other existing techniques. • An optimized MCMIFS based feature selection scheme is proposed. • The proposed MCMIFS is utilized with Kernel ELM for multiclass intrusion detection. • The proposed IDS is tested using standard intrusion detection datasets. • Standard performance metrics i.e. Accuracy, Detection Rate and FPR are utilized. • Proposed technique outperforms the existing ones for all used cases. [ABSTRACT FROM AUTHOR]

Bayesian computation plays an important role in modern machine learning and statistics to reason about uncertainty. A key computational challenge in Bayesian inference is to develop efficient techniques to approximate, or draw samples from posterior distributions. Stein variational gradient decent (SVGD) has been shown to be a powerful approximate inference algorithm for this issue. However, the vanilla SVGD requires calculating the gradient of the target density and cannot be applied when the gradient is unavailable or too expensive to evaluate. In this paper we explore one way to address this challenge by the construction of a local surrogate for the target distribution in which the gradient can be obtained in a much more computationally feasible manner. More specifically, we approximate the forward model using a deep neural network (DNN) which is trained on a carefully chosen training set, which also determines the quality of the surrogate. To this end, we propose a general adaptation procedure to refine the local approximation online without destroying the convergence of the resulting SVGD. This significantly reduces the computational cost of SVGD and leads to a suite of algorithms that are straightforward to implement. The new algorithm is illustrated on a set of challenging Bayesian inverse problems, and numerical experiments demonstrate a clear improvement in performance and applicability of standard SVGD. [ABSTRACT FROM AUTHOR]

[Display omitted] • Strain-sensing ability of a nanoengineered self-sensing cementitious composite is assessed by integrating multiscale finite element analysis simulations and machine learning. • Hyperparameter-optimized feed-forward multilayer perceptron-based neural network showed excellent efficacy for the prediction of strain-sensing ability. • Shapley Additive Explanations algorithm reveals that coating thickness shows higher influence on the sensing ability than fraction of coated sand. • Comprehensive approach presented here can be used as starting point toward development of reliable performance standards for self-sensing cementitious composites. Prediction of in-situ strain sensing efficiency of self-sensing cementitious composites using machine learning (ML) requires a large, representative, consistent, and accurate dataset. However, such large experimental dataset is not readily available. Moreover, the success of the ML approach depends on its ability to abide by the fundamental laws of physics. To address these challenges this paper synergistically integrates a validated finite element analysis (FEA)-based multiscale simulation framework with ML to predict the strain-sensing ability of self-sensing cementitious composites enabled by incorporating nano-engineered interfaces. The multiscale simulation framework is leveraged to develop a balanced, representative, complete, and consistent dataset containing 3000 combinations of strain-dependent electromechanical responses. This large dataset is used to predict the strain-sensing ability of the nanoengineered cementitious composites using a feed-forward multilayer perceptron-based neural network (NN) approach which shows excellent prediction efficacy. This paper also applies a Shapley Additive Explanations (SHAP) algorithm to interpret the NN predictions in light of the relative importance of different design parameters on the strain-sensing ability of the composite. Overall, the synergistic and comprehensive approach presented here can be used as a starting point toward the development of reliable performance standards to accelerate the acceptance of these self-sensing cementitious composites for large-scale applications. [ABSTRACT FROM AUTHOR]

• The proposed framework can learn rotation equivariant representations of images. • Models equiped with group structure achieve better performances than standard models. • The roto-translation group structure removes the need for rotation data augmentation. Rotation-invariance is a desired property of machine-learning models for medical image analysis and in particular for computational pathology applications. We propose a framework to encode the geometric structure of the special Euclidean motion group SE(2) in convolutional networks to yield translation and rotation equivariance via the introduction of SE(2) -group convolution layers. This structure enables models to learn feature representations with a discretized orientation dimension that guarantees that their outputs are invariant under a discrete set of rotations. Conventional approaches for rotation invariance rely mostly on data augmentation, but this does not guarantee the robustness of the output when the input is rotated. At that, trained conventional CNNs may require test-time rotation augmentation to reach their full capability. This study is focused on histopathology image analysis applications for which it is desirable that the arbitrary global orientation information of the imaged tissues is not captured by the machine learning models. The proposed framework is evaluated on three different histopathology image analysis tasks (mitosis detection, nuclei segmentation and tumor detection). We present a comparative analysis for each problem and show that consistent increase of performances can be achieved when using the proposed framework. [ABSTRACT FROM AUTHOR]

We incorporate future information in the form of the estimated value of future gradients in online convex optimization. This is motivated by demand response in power systems, where forecasts about the current round, e.g., the weather or the loads' behavior, can be used to improve on predictions made with only past observations. Specifically, we introduce an additional predictive step that follows the standard online convex optimization step when certain conditions on the estimated gradient and descent direction are met. We show that under these conditions and without any assumptions on the predictability of the environment, the predictive update strictly improves on the performance of the standard update. We give two types of predictive update for various family of loss functions. We provide a regret bound for each of our predictive online convex optimization algorithms. Finally, we apply our framework to an example based on demand response which demonstrates its superior performance to a standard online convex optimization algorithm. [ABSTRACT FROM AUTHOR]