Abstract
While the effects of innovation on employment have been a controversial issue in economic literature for several years, this economic puzzle is particularly relevant nowadays. We are witnessing tremendous technological developments which threaten to disrupt the labour market, due to their potential for significantly automating human labour. As such, this paper presents a qualitative study of the dynamics underlying the relationship between innovation and employment, using an agent-based model developed in Python. The model represents an economy populated by firms able to perform either Product Innovation (leading to the discovery of new tasks, which require human labour) or Process Innovation (leading to the automation of tasks previously performed by humans). The analysis led to three major conclusions, valid in this context. The first takeaway is that the Employment Rate in a given economy is dependent on the automation potential of the tasks in that economy and dependent on the type of innovation performed by firms in that economy (with Product Innovation having a positive effect on employment and Process Innovation having a negative effect). Second, in any given economy, if firms' propensity for product and process innovation, as well as the automation potential of their tasks are stable over time, the Employment Rate in that economy will tend towards stability over time. The third conclusion is that higher levels of Process Innovation and lower levels of Product Innovation, lead to a more intense decline of wage shares and to a wider gap between employee productivity growth and wage growth.

Abstract
The Robot Operating System (ROS) is one of the most popular open source robotic frameworks, and has contributed significantly to the fast development of robotics. Even though ROS provides many ready-made components, a robotic system is inherently complex, in particular regarding the architecture and orchestration of such components. Availability and analysis of a system's architecture at compile time is fundamental to ease comprehension and development of higher-quality software. However, ROS developers have to overcome this complexity relying mostly on testing and runtime visualisers. This work aims to enhance static-time support by proposing, firstly, a metamodel to describe the software architecture of ROS systems (the ROS Computation Graph) and, secondly, model extraction and visualisation tools for such architectural models. The provided tools allow users to specify custom-made queries over these models, enabling the static verification of relevant properties that had to be (manually) checked at runtime before.

Abstract
Each person carries digital devices that communicate with other person's devices and may transmit sensitive data. When the transmitted data is sensitive, it becomes necessary to implement secure mechanisms to hide the information and thus avoid it from being analyzed by third parties. This paper aims to present Ncryptr, a web-based application that allows its users to exchange encrypted instant messages in real time. Outside parties are unable to extract the contents of the messages exchanged between two or more users. Ncryptr allows its users to choose the type of encryption employed, being this a major distinction with others already existing instant messaging applications. The performance analysis is made by comparing the efficiency and latency time of different encryption methods, showing the pros and cons of using each one.

Abstract
Despite the high performance of current presentation attack detection (PAD) methods, the robustness to unseen attacks is still an under addressed challenge. This work approaches the problem by enforcing the learning of the bona fide presentations while making the model less dependent on the presentation attack instrument species (PAIS). The proposed model comprises an encoder, mapping from input features to latent representations, and two classifiers operating on these underlying representations: (i) the task-classifier, for predicting the class labels (as bona fide or attack); and (ii) the species-classifier, for predicting the PAIS. In the learning stage, the encoder is trained to help the task-classifier while trying to fool the species-classifier. Plus, an additional training objective enforcing the similarity of the latent distributions of different species is added leading to a 'PAI-species'-independent model. The experimental results demonstrated that the proposed regularisation strategies equipped the neural network with increased PAD robustness. The adversarial model obtained better loss and accuracy as well as improved error rates in the detection of attack and bona fide presentations. © 2019 Gesellschaft fuer Informatik.

Abstract
Lung cancer has an increasing preponderance in worldwide mortality, demanding for the development of efficient screening methods. With this in mind, a binary classification method using Lung-RADS (TM) guidelines to warn changes in the screening management is proposed. First, having into account the lack of public datasets for this task, the lung nodules in the LIDC-IDRI dataset were re-annotated to include a Lung-RADS (TM)-based referral label. Then, a wide residual network is used for automatically assessing lung nodules in 3D chest computed tomography exams. Unlike the standard malignancy prediction approaches, the proposed method avoids the need to segment and characterize lung nodules, and instead directly defines if a patient should be submitted for further lung cancer tests. The system achieves a nodule-wise accuracy of 0.87 +/- 0.02.

Abstract