Abstract
Due to its pervasive nature, the Internet of Things (IoT) is demanding for Low Power Wide Area Networks (LPWAN) since wirelessly connected devices need battery-efficient and long-range communications. Due to its low-cost and high availability (regional/city level scale), this type of network has been widely used in several IoT applications, such as Smart Metering, Smart Grids, Smart Buildings, Intelligent Transportation Systems (ITS), SCADA Systems. By using LPWAN technologies, the IoT devices are less dependent on common and existing infrastructure, can operate using small, inexpensive, and long-lasting batteries (up to 10 years), and can be easily deployed within wide areas, typically above 2 km in urban zones. The starting point of this work was an overview of the security vulnerabilities that exist in LPWANs, followed by a literature review with the main goal of substantiating an attack vector analysis specifically designed for the IoT ecosystem. This methodological approach resulted in three main contributions: (i) a systematic review regarding cybersecurity in LPWANs with a focus on vulnerabilities, threats, and typical defense strategies; (ii) a state-of-the-art review on the most prominent results that have been found in the systematic review, with focus on the last three years; (iii) a security analysis on the recent attack vectors regarding IoT applications using LPWANs. Results have shown that LPWANs communication technologies contain security vulnerabilities that can lead to irreversible harm in critical and non-critical IoT application domains. Also, the conception and implementation of up-to-date defenses are relevant to protect systems, networks, and data.

Abstract
Understanding the intellectual landscape of scientific communities and their collaborations has become an indispensable part of research per se. In this regard, measuring similarities among scientific documents can help researchers to identify groups with similar interests as a basis for strengthening collaboration and university-industry linkages. To this end, we intend to evaluate the performance of hybrid crowd-computing methods in measuring the similarity between document pairs by comparing the results achieved by crowds and artificial intelligence (AI) algorithms. That said, in this paper we designed two types of experiments to illustrate some issues in calculating how similar an automatic solution is to a given ground truth. In the first type of experiments, we created a crowdsourcing campaign consisting of four human intelligence tasks (HITs) in which the participants had to indicate whether or not a set of papers belonged to the same author. The second type involves a set of natural language processing (NLP) processes in which we used the TF-IDF measure and the Bidirectional Encoder Representation from Transformers (BERT) model. The results of the two types of experiments carried out in this study provide preliminary insight into detecting major contributions from human-AI cooperation at similarity calculation in order to achieve better decision support. We believe that in this case decision makers can be better informed about potential collaborators based on content-based insights enhanced by hybrid human-AI mechanisms.

Abstract
In this paper, we present and discuss an innovative approach to solve Job Shop scheduling problems based on machine learning techniques. Traditionally, when choosing how to solve Job Shop scheduling problems, there are two main options: either use an efficient heuristic that provides a solution quickly, or use classic optimization approaches (e.g., metaheuristics) that take more time but will output better solutions, closer to their optimal value. In this work, we aim to create a novel architecture that incorporates reinforcement learning into scheduling systems in order to improve their overall performance and overcome the limitations that current approaches present. It is also intended to investigate the development of a learning environment for reinforcement learning agents to be able to solve the Job Shop scheduling problem. The reported experimental results and the conducted statistical analysis conclude about the benefits of using an intelligent agent created with reinforcement learning techniques. The main contribution of this work is proving that reinforcement learning has the potential to become the standard method whenever a solution is necessary quickly, since it solves any problem in very few seconds with high quality, approximate to the optimal methods.

Abstract

Abstract

Abstract
In recent years, deep neural networks have shown significant progress in computer vision due to their large generalization capacity; however, the overfitting problem ubiquitously threatens the learning process of these highly nonlinear architectures. Dropout is a recent solution to mitigate overfitting that has witnessed significant success in various classification applications. Recently, many efforts have been made to improve the Standard dropout using an unsupervised merit-based semantic selection of neurons in the latent space. However, these studies do not consider the task-relevant information quality and quantity and the diversity of the latent kernels. To solve the challenge of dropping less informative neurons in deep learning, we propose an efficient end-to-end dropout algorithm that selects the most informative neurons with the highest correlation with the target output considering the sparsity in its selection procedure. First, to promote activation diversity, we devise an approach to select the most diverse set of neurons by making use of determinantal point process (DPP) sampling. Furthermore, to incorporate task specificity into deep latent features, a mutual information (MI)-based merit function is developed. Leveraging the proposed MI with DPP sampling, we introduce the novel DPPMI dropout that adaptively adjusts the retention rate of neurons based on their contribution to the neural network task. Empirical studies on real-world classification benchmarks including, MNIST, SVHN, CIFAR10, CIFAR100, demonstrate the superiority of our proposed method over recent state-of-the-art dropout algorithms in the literature.