Abstract

Abstract
Reinforcement Learning (RL) is a well-suited paradigm to train robots since it does not require any previous information or database to train an agent. This paper explores using Deep Reinforcement Learning (DRL) to train a robot to navigate in maps containing different sorts of obstacles and which emulate hallways. Training and testing were performed using the Flatland 2D simulator and a Deep Q-Network (DQN) provided by OpenAI gym. Different sets of maps were used for training and testing. The experiments illustrate how well the robot is able to navigate in maps distinct from the ones used for training by learning new behaviours (namely following walls) and highlight the key challenges when solving this task using DRL, including the appropriate definition of the state space and reward function, as well as of the stopping criteria during training.

Abstract
Disentangling innovation from growth is unrealistic in the present times. Also, anticipating the future behavior of innovative firms is relevant to the entire innovation ecosystem; and assessing the persistence of innovation and appraising the role of factors affecting ongoing innovation activities in firms is essential. This chapter discusses a very important subject related to the concept of innovation persistence in relation to structural innovation characteristics of firms, with a focus on technological regimes, to better understand if there is change in innivation continuity accordingly to the technological intensity embedded in the sector. The empirical research is based on data from CIS database, comprising 3237 firms which present in the 2014 and 2018 waves. We analyze the innovative persistence behavior of these firms regarding proxies like firm dimension, innovation activities, types of innovation, government funding, and more importantly, technological regimes. To do this, we applied binary logistic regression for developing a model which can forecast the drivers of innovation persistency propensity. The presented study shows that some very important results are achieved. Besides demonstrating innovative persistency in 75% of science-based firms, the findings confirm that firms in high-tech and science-based industries are more prone to continue innovating and, as a result, this consistency in innovation will generate virtuous cycles of innovation. Furthermore, our data shows that complex innovators are more likely to persist than single innovators, proving the existence of complementarities among the innovation types.

Abstract
Despite major advances in recent years, the field of Machine Learning continues to face research and technical challenges. Mostly, these stem from big data and streaming data, which require models to be frequently updated or re-trained, at the expense of significant computational resources. One solution is the use of distributed learning algorithms, which can learn in a distributed manner, from distributed datasets. In this paper, we describe CEDEs-a distributed learning system in which models are heterogeneous distributed Ensembles, i.e., complex models constituted by different base models, trained with different and distributed subsets of data. Specifically, we address the issue of predicting the training time of a given model, given its characteristics and the characteristics of the data. Given that the creation of an Ensemble may imply the training of hundreds of base models, information about the predicted duration of each of these individual tasks is paramount for an efficient management of the cluster's computational resources and for minimizing makespan, i.e., the time it takes to train the whole Ensemble. Results show that the proposed approach is able to predict the training time of Decision Trees with an average error of 0.103 s, and the training time of Neural Networks with an average error of 21.263 s. We also show how results depend significantly on the hyperparameters of the model and on the characteristics of the input data.

Abstract
Information and Communication Technologies are driving the improvement of industrial processes. According to the Industry 4.0 (I4.0) paradigm, digital systems provide real-time information to humans and machines, increasing flexibility and efficiency in production environments. Based on the I4.0 Design Principles concept, Virtual Assistants can play a vital role in processing production data and offer contextualized and real-time information to the workers in the production environment. This systematic review paper explored Virtual Assistant applications in the context of I4.0, discussing the Technical Assistance Design Principle and identifying the characteristics, services, and limitations regarding Virtual Assistant use in the production environments. The results showed that Virtual Assistants offer Physical and Virtual Assistance. Virtual Assistance provides real-time contextualized information mainly for support, while Physical Assistance is oriented toward task execution. Regarding services, the applications include integration with legacy systems and static information treatment. The limitations of the applications incorporate concerns about information security and adapting to noisy and unstable environments. It is possible to assume that the terminology of Virtual Assistants is not standardized and is mentioned as chatbots, robots, and others. Besides the worthy insights of this research, the small number of resulting papers did not allow for generalizations. Future research should focus on broadening the search scope to provide more-significant conclusions and research possibilities with new AI models and services, including the emergent Industry 5.0 concept.

Abstract
Recent developments in dexterous robotic manipulation technologies allowed for the design of very compact, yet capable, multi-fingered robotic hands. These can be designed to emulate the human touch and feel, reducing the aforementioned need for human expertise in highly detailed tasks. The presented work focused on the application of two simulation platforms Gazebo and MuJoCo - to a use-case of a Schunk Five Finger Robotic Hand, coupled to the UR5 collaborative manipulator. This allowed to assess the relative appropriateness of each of these platforms.