Abstract
Ensemble methods have been successfully used in the past to improve recommender systems; however, they have never been studied with incremental recommendation algorithms. Many online recommender systems deal with continuous, potentially fast, and unbounded flows of databig data streamsand often need to be responsive to fresh user feedback, adjusting recommendations accordingly. This is clear in tasks such as social network feeds, news recommender systems, automatic playlist completion, and other similar applications. Batch ensemble approaches are not suitable to perform continuous learning, given the complexity of retraining new models on demand. In this paper, we adapt a general purpose online bagging algorithm for top-N recommendation tasks and propose two novel online bagging methods specifically tailored for recommender systems. We evaluate the three approaches, using an incremental matrix factorization algorithm for top-N recommendation with positive-only user feedback data as the base model. Our results show that online bagging is able to improve accuracy up to 55% over the baseline, with manageable computational overhead.

Abstract
A Community of Practice (CoP) allows practitioners of a clearly defined domain to share knowledge, experience, and best practices. It provides a social context for practitioners, often distributed across multiple organizations, and emerged over the last few decades as a fundamental mechanism for knowledge sharing, management, and generation within organizations. Best practices, innovations, and solutions to shared problems first emerge within CoPs. These are, and must be perceived as, an investment in organizations' future and competitiveness. Establishing a CoP is a straightforward process, the most challenging factor being the recruitment of members to attain critical mass. The challenge is to maintain the CoP active, with members contributing with high quality, innovative content. Increasing a CoP's medium / long-term survival probabilities requires careful planning to avoid incurring in some well-known pitfalls. This paper proposes and discusses a set of nine guidelines for establishing and maintaining a community of practice within the context of Electronic Governance (EGOV) and Government Chief Information Officers (GCIO). This research was motivated by the initiative of the government of a developing country. Results are based on a review of the relevant literature, together with the detailed analysis of interviews to members or coordinators of large communities of practice. This analysis was further validated against the opinions of public servants directly involved on EGOV-GCIO-related functions during two focus groups meetings.

Abstract
Many real-world phenomena exhibit both relational structure and uncertainty. Probabilistic Inductive Logic Programming (PILP) uses Inductive Logic Programming (ILP) extended with probabilistic facts to produce meaningful and interpretable models for real-world phenomena. This merge between First Order Logic (FOL) theories and uncertainty makes PILP a very adequate tool for knowledge representation and extraction. However, this flexibility is coupled with a problem (inherited from ILP) of exponential search space growth and so, often, only a subset of all possible models is explored due to limited resources. Furthermore, the probabilistic evaluation of FOL theories, coming from the underlying probabilistic logic language and its solver, is also computationally demanding. This work introduces a prediction-based pruning strategy, which can reduce the search space based on the probabilistic evaluation of models, and a safe pruning criterion, which guarantees that the optimal model is not pruned away, as well as two alternative more aggressive criteria that do not provide this guarantee. Experiments performed using three benchmarks from different areas show that prediction pruning is effective in (i) maintaining predictive accuracy for all criteria and experimental settings; (ii) reducing the execution time when using some of the more aggressive criteria, compared to using no pruning; and (iii) selecting better candidate models in limited resource settings, also when compared to using no pruning.

Abstract
Cloud software continues to expand globally, highly motivated by how widespread the Internet is and the possibilities it unlocks with cloud computing. Still, cloud development has some intrinsic properties to it, making it complex to unexperienced developers. This research is capturing those intricacies in the form of a pattern language that gathers ten patterns for engineering software for the cloud. This paper elaborates on that research by contributing with two new patterns: Automated Recovery, which checks if a container is working properly, automatically recovering it in case of failure and Scheduler, which periodically executes actions within the infrastructure. The described patterns are useful for anyone designing software for the cloud, either to bootstrap or to validate their design decisions with the end goal of enabling them to create better software for the cloud.

Abstract
Line-following robots have the ability to recognize and follow a line drawn on a surface. Elements of their operating principles could be used in the evelopment of numerous autonomous technologies, with applications in education and industry. A simulator has been developed to aide in performing several trials in order to validate a project. By taking the Pololu 3pi Robot as the model, the proposed solution simulates its physical structure, behavior, and operations-being able to read lines on surfaces-enabling the user to observe the robot following the line according to the code used. This paper aims to validate the developed simulator as an alternative to ease the process of learning to use the 3pi platform applied in both educational and competitive environments.

Abstract
This chapter provides an overview of the book theme, motivating the need for high-performance and time-predictable embedded computing. It describes the challenges introduced by the need for time-predictability on the one hand, and high-performance on the other, discussing on a high level how these contradictory requirements can be simultaneously supported.