Recommender systems are intrinsically tied to a reliability/coverage dilemma: The more reliable we desire the forecasts, the more conservative the decision will be and thus, the fewer items will be recommended. This causes a detriment to the predictive capability of the system, as it is only able to estimate potential interest in items for which there is a consensus in their evaluation, rather than being able to estimate potential interest in any item. In this paper, we propose the inclusion of a new term in the learning process of matrix factorization-based recommender systems, called recklessness, that takes into account the variance of the output probability distribution of the predicted ratings. In this way, gauging this recklessness measure we can force more spiky output distribution, enabling the control of the risk level desired when making decisions about the reliability of a prediction. Experimental results demonstrate that recklessness not only allows for risk regulation but also improves the quantity and quality of predictions provided by the recommender system.

The qualitative research on empirical software engineering that uses Grounded Theory is increasing (GT). The trustworthiness, rigor, and transparency of GT qualitative data analysis can benefit, among others, when multiple analysts juxtapose diverse perspectives and collaborate to develop a common code frame based on a consensual and consistent interpretation. Inter-Rater Reliability (IRR) and/or Inter-Rater Agreement (IRA) are commonly used techniques to measure consensus, and thus develop a shared interpretation. However, minimal guidance is available about how and when to measure IRR/IRA during the iterative process of GT, so researchers have been using ad hoc methods for years. This paper presents a process for systematically measuring IRR/IRA in GT studies, when appropriate, which is grounded in a previous systematic mapping study on collaborative GT in the field of software engineering. Meta-science guided us to analyze the issues and challenges of collaborative GT and formalize a process to measure IRR/IRA in GT. This process guides researchers to incrementally generate a theory while ensuring consensus on the constructs that support it, improving trustworthiness, rigor, and transparency, and promoting the communicability, reflexivity, and replicability of the research. The application of this process to a GT study seems to support its feasibility. In the absence of further confirmation, this would represent the first step in a de facto standard to be applied to those GT studies that may benefit from IRR/IRA techniques.

The research on empirical software engineering that uses qualitative data analysis is increasing. However, most of them do not deepen into the validity of the findings, specifically in the reliability of coding in which these methodologies rely on. This paper aims to establish a novel theoretical framework that enables a methodological approach for conducting this validity analysis through Inter-Coder Agreement (ICA), based on the use of coefficients to measure the degree of agreement in collaborative coding. We systematically review several existing variants of Krippendorff’s α coefficients and provide a novel common mathematical framework to unify them. Finally, this paper illustrates the use of this theoretical framework in a large case study on DevOps culture. We expect that this work will help researchers who are committed to measuring consensus with quantitative techniques in collaborative coding, conducted as part of a qualitative research, to improve the rigor of their findings.

Gender-based crime is one of the most concerning scourges of contemporary society, and governments worldwide have invested lots of economic and human resources to foretell their occurrence and anticipate the aggressions. In this work, we propose to apply Machine Learning (ML) techniques to create models that accurately predict the recidivism risk of a gender-violence offender. We feed the model with data extracted from the official Spanish VioGen system and comprising more than 40,000 reports of gender violence. To evaluate the performance, two new quality measures are proposed to assess the effective police protection that a model supplies and the overload in the invested resources that it generates. The empirical results show a clear outperformance of the ML-centered approach, with an improvement of up to a 25% with respect to the preexisting risk assessment system. Additionally, we propose a hybrid model that combines the statistical prediction methods with the ML method, permitting authorities to implement a smooth transition from the preexisting model to the ML-based model. To the best of our knowledge, this is the first work that achieves an effective ML-based prediction for this type of crimes against an official dataset.

Abstract Deep learning provides accurate collaborative filtering models to improve recommender system results. Deep matrix factorization and their related collaborative neural networks are the state of the art in the field; nevertheless, both models lack the necessary stochasticity to create the robust, continuous, and structured latent spaces that variational autoencoders exhibit. On the other hand, data augmentation through variational autoencoder does not provide accurate results in the collaborative filtering field due to the high sparsity of recommender systems. Our proposed models apply the variational concept to inject stochasticity in the latent space of the deep architecture, introducing the variational technique in the neural collaborative filtering field. This method does not depend on the particular model used to generate the latent representation. In this way, this approach can be applied as a plugin to any current and future specific models. The proposed models have been tested using four representative open datasets, three different quality measures, and state-of-the-art baselines. The results show the superiority of the proposed approach in scenarios where the variational enrichment exceeds the injected noise effect. Additionally, a framework is provided to enable the reproducibility of the conducted experiments.

Abstract With the latest advances in deep learning-based generative models, it has not taken long to take advantage of their remarkable performance in the area of time series. Deep neural networks used to work with time series heavily depend on the size and consistency of the datasets used in training. These features are not usually abundant in the real world, where they are usually limited and often have constraints that must be guaranteed. Therefore, an effective way to increase the amount of data is by using data augmentation techniques, either by adding noise or permutations and by generating new synthetic data. This work systematically reviews the current state of the art in the area to provide an overview of all available algorithms and proposes a taxonomy of the most relevant research. The efficiency of the different variants will be evaluated as a central part of the process, as well as the different metrics to evaluate the performance and the main problems concerning each model will be analysed. The ultimate aim of this study is to provide a summary of the evolution and performance of areas that produce better results to guide future researchers in this field.

Solving the convergence issues of Generative Adversarial Networks (GANs) is one of the most outstanding problems in generative models. In this work, we propose a novel activation function to be used as output of the generator agent. This activation function is based on the Smirnov probabilistic transformation and it is specifically designed to improve the quality of the generated data. In sharp contrast to previous works, our activation function provides a more general approach that deals not only with the replication of categorical variables but with any type of data distribution (continuous or discrete). Moreover, our activation function is derivable and therefore, it can be seamlessly integrated in the backpropagation computations during the GAN training processes. To validate this approach, we firstly evaluate our proposal on two different data sets: a) an artificially rendered data set containing a mixture of discrete and continuous variables, and b) a real data set of flow-based network traffic data containing both normal connections and cryptomining attacks. In addition, three publicly available data sets were added to the evaluation to generalize the obtained results. To evaluate the fidelity of the generated data, we analyze their results both in terms of quality measures of statistical nature and regarding the use of these synthetic data to feed a nested machine learning-based classifier. The experimental results evince a clear outperformance of a Wasserstein GAN network (WGAN) tuned with this new activation function with respect to both a naïve mean-based generator and a standard WGAN. The quality of the generated data allows to fully substitute real data with synthetic data for training the nested classifier without a significant fall in the obtained accuracy.

Abstract In the context of recommender systems based on collaborative filtering (CF), obtaining accurate neighborhoods of the items of the datasets is relevant. Beyond particular individual recommendations, knowing these neighbors is fundamental for adding differentiating factors to recommendations, such as explainability, detecting shilling attacks, visualizing item relations, clustering, and providing reliabilities. This paper proposes a deep learning architecture to efficiently and accurately obtain CF neighborhoods. The proposed design makes use of a classification neural network to encode the dataset patterns of the items, followed by a generative process that obtains the neighborhood of each item by means of an iterative gradient localization algorithm. Experiments have been conducted using five popular open datasets and five representative baselines. The results show that the proposed method improves the quality of the neighborhoods compared to theK-Nearest Neighbors (KNN) algorithm for the five selected similarity measure baselines. The efficiency of the proposed method is also shown by comparing its computational requirements with that of KNN.

Context: DevOps can be defined as a cultural movement to improve and accelerate the delivery of business value by making the collaboration between development and operations effective.Objective: This paper aims to help practitioners and researchers to better understand the organizational structure and characteristics of teams adopting DevOps.Method: We conducted an exploratory study by leveraging in-depth, semi-structured interviews with relevant stakeholders of 31 multinational software-intensive companies, together with industrial workshops and observations at organizations’ facilities that supported triangulation. We used Grounded Theory as qualitative research method to explore the structure and characteristics of teams, and statistical analysis to discover their implications in software delivery performance.Results: We describe a taxonomy of team structures that shows emerging, stable and consolidated product teams that are classified according to six variables, such as collaboration frequency, product ownership sharing, and autonomy, among others, as well as their implications on software delivery performance. These teams are often supported by horizontal teams (DevOps platform teams, Centers of Excellence, and chapters) that provide them with platform technical capabilities, mentoring and evangelization, and even temporarily may facilitate human resources.Conclusion: This study aims to strengthen evidence and support practitioners in making better informed about organizational team structures by analyzing their main characteristics and implications in software delivery performance.

DevOps is becoming a main competency required by the software industry. However, academic institutions have been slow to provide DevOps training in software engineering (SE) curricula. One reason for this is the fact that the problems addressed by DevOps may be hard to understand to students who have not previously worked in the industry or on projects of meaningful size and complexity. This paper shows an experience that integrates DevOps in SE curricula through research-based teaching (RBT). We aim to expose students to the problems that have led companies to adopt DevOps by researching and analyzing real cases of companies, thereby placing students at the center of learning. The contribution of this work is to innovate the application of RBT in software engineering by showing that the RBT approach is, at least, as good as the traditional approach and that it also leads to some extra benefits. This innovative solution has been implemented by using (i) qualitative analysis, specifically coding techniques, to discover knowledge and (ii) inter-coder agreement (ICA), specifically Krippendorff’s$\alpha$coefficients, to measure the extent of students’ learning. These techniques allow teachers to determine whether students’ learning in the subject is homogeneous and to analyze disagreements among students during their analysis. This approach provides teachers with new tools (Krippendorff’s$\alpha$coefficients) to identify those concepts that are less understood by students and to evaluate whether improvements in the research instruments (e.g., the codebook used in the qualitative analysis) also generate improvements in the students’ agreement. This RBT experience shows evidence that can be used to assess whether a similar experience and the use of ICA could be applied in similar learning contexts with similar research contexts.

High efficiency and reliability are two desirable features for wind energy conversion systems. In this regard, multiphase machines provide inherent fault-tolerance and better power density than conventional three-phase systems. From the point of view of the reliability, multiphase machines can provide a natural/passive fault-tolerance without a mandatory control reconfiguration. For that purpose, the control of the x-y current must be realized in open-loop mode. A recent model predictive control based on virtual voltage vectors (VV-MPC) has satisfactory validated this fact. On the other hand, efficiency can be enhanced with the implementation of a variable-flux control that reduces the copper losses. In order to satisfy the requirements of new wind energy conversion systems, this work proposes an efficient model predictive control based on virtual voltage vectors (EVV-MPC) with a natural fault tolerance for six-phase induction machines (IM). Simulation results confirm the capability of the proposed system to minimize the cooper losses in pre- and post- fault situation without control reconfiguration.