Data Augmentation for Deep-Learning-Based Multiclass Structural Damage Detection Using Limited Information

The deterioration of infrastructure's health has become more predominant on a global scale during the 21st century. Aging infrastructure as well as those structures damaged by natural disasters have prompted the research community to improve state-of-the-art methodologies for conducting Structural Health Monitoring (SHM). The necessity for efficient SHM arises from the hazards damaged infrastructure imposes, often resulting in structural collapse, leading to economic loss and human fatalities. Furthermore, day-to-day operations in these affected areas are limited until an inspection is performed to assess the level of damage experienced by the structure and the required rehabilitation determined. However, human-based inspections are often labor-intensive, inefficient, subjective, and restricted to accessible site locations, which ultimately negatively impact our ability to collect large amounts of data from inspection sites. Though Deep-Learning (DL) methods have been heavily explored in the past decade to rectify the limitations of traditional methods and automate structural inspection, data scarcity continues to remain prevalent within the field of SHM. The absence of sufficiently large, balanced, and generalized databases to train DL-based models often results in inaccurate and biased damage predictions. Recently, Generative Adversarial Networks (GANs) have received attention from the SHM community as a data augmentation tool by which a training dataset can be expanded to improve the damage classification. However, there are no existing studies within the SHM field which investigate the performance of DL-based multiclass damage identification using synthetic data generated from GANs. Therefore, this paper investigates the performance of a convolutional neural network architecture using synthetic images generated from a GAN for multiclass damage detection of concrete surfaces. Through this study, it was determined the average classification performance of the proposed CNN on hybrid datasets decreased by 10.6% and 7.4% for validation and testing datasets when compared to the same model trained entirely on real samples. Moreover, each model's performance decreased on average by 1.6% when comparing a singular model trained with real samples and the same model trained with both real and synthetic samples for a given training configuration. The correlation between classification accuracy and the amount and diversity of synthetic data used for data augmentation is quantified and the effect of using limited data to train existing GAN architectures is investigated. It was observed that the diversity of the samples decreases and correlation increases with the increase in the number of synthetic samples.

Comparison of Machine Learning Techniques for Activities of Daily Living Classification with Electromyographic Data

Advances in data science and wearable robotic devices present an opportunity to improve rehabilitation outcomes. Some of these devices incorporate electromyography (EMG) electrodes that sense physiological patient activity, making it possible to develop rehabilitation systems able to assess the patient's progress when performing activities of daily living (ADLs). However, additional research is needed to improve the ability to interpret EMG signals. To address this issue, an off-line classification approach for the 26 upper-limb ADLs included in the KIN-MUS UJI dataset is presented in this paper. The ADLs were performed by 22 subjects, while seven EMG signals were recorded from their forearms. From variable-length EMG time windows, 18 features were computed, and 13 features more were extracted from frequency domain windows. The classification performance of five different machine learning techniques, including Support Vector Machines (SVM), Convolutional Neural Networks (CNN), Gated Recurrent Unit (GRU) network, XGBoost, and Random Forests, were compared. CNN performed best amongst individual models, with an accuracy above 80%, compared to SVM with 77%, GRU with 73.9%, and the tree-based models below 64%. Ensemble learning with four CNN models achieved an even higher accuracy of 86%. These results suggest that the CNN ensemble model is capable of classifying EMG signals for most ADLs, which could be used in off-line quantitative assessment of robotic rehabilitation outcomes.

(CF)2 architecture: contextual collaborative filtering

Recommender systems have dramatically changed the way we consume content. Internet applications rely on these systems to help users navigate among the ever-increasing number of choices available. However, most current systems ignore the fact that user preferences can change according to context, resulting in recommendations that do not fit user interests. This research addresses these issues by proposing the ( CF ) 2 architecture, which uses local learning techniques to embed contextual awareness into collaborative filtering models. The proposed architecture is demonstrated on two large-scale case studies involving over 130 million and over 7 million unique samples, respectively. Results show that contextual models trained with a small fraction of the data provided similar accuracy to collaborative filtering models trained with the complete dataset. Moreover, the impact of taking into account context in real-world datasets has been demonstrated by higher accuracy of context-based models in comparison to random selection models.

Data management in cloud environments: NoSQL and NewSQL data stores

Advances in Web technology and the proliferation of mobile devices and sensors connected to the Internet have resulted in immense processing and storage requirements. Cloud computing has emerged as a paradigm that promises to meet these requirements. This work focuses on the storage aspect of cloud computing, specifically on data management in cloud environments. Traditional relational databases were designed in a different hardware and software era and are facing challenges in meeting the performance and scale requirements of Big Data. NoSQL and NewSQL data stores present themselves as alternatives that can handle huge volume of data. Because of the large number and diversity of existing NoSQL and NewSQL solutions, it is difficult to comprehend the domain and even more challenging to choose an appropriate solution for a specific task. Therefore, this paper reviews NoSQL and NewSQL solutions with the objective of: (1) providing a perspective in the field, (2) providing guidance to practitioners and researchers to choose the appropriate data store, and (3) identifying challenges and opportunities in the field. Specifically, the most prominent solutions are compared focusing on data models, querying, scaling, and security related capabilities. Features driving the ability to scale read requests and write requests, or scaling data storage are investigated, in particular partitioning, replication, consistency, and concurrency control. Furthermore, use cases and scenarios in which NoSQL and NewSQL data stores have been used are discussed and the suitability of various solutions for different sets of applications is examined. Consequently, this study has identified challenges in the field, including the immense diversity and inconsistency of terminologies, limited documentation, sparse comparison and benchmarking criteria, and nonexistence of standardized query languages.