Abstract
There has been a growing interest in academia regarding the term 'social innovation', including in disciplines such as sociology, administration, history, management, psychology, and economics. The literature highlights the lack of scientific clarity in the use of the term, and some scholars argue that the term is no more than a 'buzzword' or a 'fad'. This article focusses on the analysis of the conceptualizations of social innovation, contrasting sociological and economical approaches, and adopts an integrative approach to propose a categorization scheme of social innovation projects based on three distinct variables, namely the level of policy support, the profit orientation and the geographical scale. We argue that government support and the scalability of social innovations should be carefully pondered depending on the characteristics of the social innovation initiatives. We conclude that policy support should privilege social innovation initiatives that, while having the potential to deliver social good, are constrained by market failures. In addition, we also argue in favour of policy support for small bottom-up initiatives that have a profit-logic but are not sufficiently robust to survive on their own due to the liabilities of smallness and newness. Finally, we advise caution in public policies supporting scale-up strategies and highlighted the inherent challenges.

Abstract

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
In recent years, there have been notable commitments and obligations by the electricity sector for more sustainable generation and delivery processes to reduce the environmental footprint. However, there is still a long way to go to achieve necessary sustainability goals while ensuring standards of robustness and the quality of power grids. One of the main challenges hindering this progress are uncertainties and stochasticity associated with the electricity sector and especially renewable generation. In this paradigm shift, forecasting tools are indispensable, and their utilization can significantly improve system operation and minimize costs associated with all related activities. Thus, forecasting tools have an essential key role in all decision-making stages. In this work, a hybrid probabilistic forecasting model (HPFM) was developed for short-term electricity market prices (EMP) combining wavelet transforms (WT), hybrid particle swarm optimization (DEEPSO), adaptive neuro-fuzzy inference system (ANFIS), and Monte Carlo simulation (MCS). The proposed hybrid probabilistic forecasting model (HPFM) was tested and validated with real data from the Spanish and Pennsylvania-New Jersey-Maryland (PJM) markets. The proposed model exhibited favorable results and performance in comparison with previously published work considering electricity market prices (EMP) data, which is notable.

Abstract
Virtual Reality is becoming more popular over the years because it allows the user to be the main actor in another environment and interact with it in real time. New interaction methods are being studied, like tangible interfaces, but there is little work done related to small distances when grabbing objects through a virtual environment. This study is important because, in our perspective, interaction in virtual reality will be at arms reach, meaning that the user will interact within very close distances (under 1 m). In this paper, the research team further evaluate distance perception using gender, the presence of avatar and height (fixed or personalised). The sample consisted of 64 participants (32 females and 32 males) evenly distributed between all four conditions (8 males and 8 females for each condition). Results revealed that gender does have an impact on small distance estimation; height does not have an impact on distance estimation; and avatar does make a difference when trying to grab a real object through the virtual environment. © Springer Nature Switzerland AG 2019.

Abstract
Additive Manufacturing (AM) is one of the most trending production technologies, with a growing number of companies looking forward to implementing it in their processes. Producing through AM not only means that there are no supplier lead times needed to account for, but also enables production closer to the end customer, reducing then the delivery time. This is especially true for companies with a wide range of low and variable demand products. This paper proposes a mixed integer linear programming (MILP) model for the optimal design of supply chains facing the introduction of AM processes. In the addressed problem, the 3D printers allocation to distribution centers (DC), that will make or customize parts, and the Suppliers-DC-Customers connections for each product need to be defined. The model aims at minimizing the supply chain costs, exploring the trade-offs between safety stock and stockout costs, and between buying and 3D printing a part. The main relevant characteristics of this model are the introduction of stock service levels as decision variables and the use of a linearization of the cumulative distribution function to account for demand uncertainty. A real-world problem from a maintenance provider is solved, showing the applicability of the model. © 2019, IEOM Society International.