Abstract
Micro and Small Enterprises have assumed an important strategic role for the economic and social development of Brazil, representing about 99% of all business establishments and contributing to the generation of 55.3% of all formal jobs in the country. However, these ventures have faced a series of restrictions on entrepreneurial activity, such as low government support, high taxes, and excessive bureaucracy. In this adverse context, innovation assumes a significant mediation role for increasing business performance and, consequently, maintaining the permanence of these ventures in the market. Accordingly, the present investigation aimed to identify the degree of innovation, the main obstacles and facilitators of innovation, as well as the types of innovation prevalent in Brazilian Micro and Small Enterprises through a systematic literature review. The literature review included 21 articles extracted from the Web of Science, Scopus, Scielo, and Spell databases. In general, the analyzed literature points out that innovation in these ventures is the result of simple actions, such as imitation and the acquisition of technology from third parties, and that the degree of innovation is low, prevailing in the types of innovation in Products and Marketing. The main barriers highlighted were related to the lack of resources, precarious managerial and commercial capacities, and shortages of qualified labor, while the factors that facilitate innovation are related to the support of government programs and the organization of Micro and Small Enterprises in productive arrangements.

Abstract
This paper explores the concept of the local energy markets and, in particular, the need for trust and security in the negotiations necessary for this type of market. A multi-agent system is implemented to simulate the local energy market, and a trust model is proposed to evaluate the proposals sent by the participants, based on forecasting mechanisms that try to predict their expected behavior. A cyber-attack detection model is also implemented using several supervised classification techniques. Two case studies were carried out, one to evaluate the performance of the various classification methods using the IoT-23 cyber-attack dataset; and another one to evaluate the performance of the developed trust mode.

Abstract
In this work, we describe an innovative methodology based on combined surface plasmon resonance (SPR) and electrochemical responses (eSPR) in the same immunoassay for screening CA 15-3 cancer biomarker with high sensitivity (and selectivity), in a very simple, label-free, accurate, and fully automated manner. Detection was achieved by performing two simple steps. In the first step, direct SPR was used to monitor CA 15-3 interaction with surface immobilized antibody. Two linear response ranges were obtained and the detection limit achieved is poor (LOD of 21 U mL(-1)). However, in the second detection step, electrochemical measurements at the SPR gold surface were performed to measure the decrease of redox probe peak current upon antigen-antibody interaction, providing a suitable amplification strategy to lower detection levels of CA 15-3 (LOD of 0.0998 U mL(-1)), without the need of additional complex and/or expensive amplification steps to enhance the sensitivity. Moreover, selectivity studies were performed against other common cancer biomarkers and the results showed that the eSPR immunosensor is selective for the CA 15-3 protein. Finally, the clinical applicability of the developed eSPR biosensing methodology was successfully applied to detect CA 15-3 in human serum samples at clinically relevant levels due to the high sensitivity of electrochemical readout. The same concept may be further extended to other proteins of interest.

Abstract
Cervical cancer is the fourth leading cause of cancer-related deaths in women, especially in low to middle-income countries. Despite the outburst of recent scientific advances, there is no totally effective treatment, especially when diagnosed in an advanced stage. Screening tests, such as cytology or colposcopy, have been responsible for a substantial decrease in cervical cancer deaths. Cervical cancer automatic screening via Pap smear is a highly valuable cell imaging-based detection tool, where cells must be classified as being within one of a multitude of ordinal classes, ranging from abnormal to normal. Current approaches to ordinal inference for neural networks are found to not sufficiently take advantage of the ordinal problem or to be too uncompromising. A non-parametric ordinal loss for neuronal networks is proposed that promotes the output probabilities to follow a unimodal distribution. This is done by imposing a set of different constraints over all pairs of consecutive labels which allows for a more flexible decision boundary relative to approaches from the literature. Our proposed loss is contrasted against other methods from the literature by using a plethora of deep architectures. A first conclusion is the benefit of using non-parametric ordinal losses against parametric losses in cervical cancer risk prediction. Additionally, the proposed loss is found to be the top-performer in several cases. The best performing model scores an accuracy of 75.6% for seven classes and 81.3% for four classes.

Abstract
This papers proposes a multiple sensor-lens pair scheme to increase the misalignment tolerance and presents an accurate model of photon propagation based on the Monte Carlo simulation, that includes the photon refraction at the lenses interface and angular misalignment between emitter and receiver. The results show that the ideal divergence of the emitter's beam is around 15 degrees for a 1 metre Tx-Rx distance, increasing to 22 degrees for a shorter distance of 0.5 metres, but being independent of the water turbidity, showing that the geometry of the link is the dominant factor in such short-range links. Additionally, it was concluded that a 7 lenses scheme is approximately 3 times more offset tolerable than a single lens. Rotating the emitter increases the optimal divergence while rotating the plane of sensors decreases it.

Abstract
The increasing penetration of renewable energy sources in areas with wholesale energy markets may have significant impacts on the prices of electricity within these markets. These renewable energy sources typically have low or zero marginal prices and thus can bid into energy markets at prices which might be below plants using other generating technologies. This work seeks to understand the impact of these zero marginal cost plants in the Iberian Energy Market. This work makes use of an Artificial Neural Network (ANN) to evaluate the impact of growing renewable energy generation on the market-clearing price. Real data from the Iberian Energy Market is chosen and used to train the ANN. The scenarios used for renewable energy generation are taken from the newly published national energy and climate plans for both Spain and Portugal. Results show that increasing penetration of renewable energy leads to significant reductions in the forecasted energy price, showing a price decrease of about 23 (sic)/MWh in 2030 compared to the baseline. Increasing solar PV generation has the largest effect on market prices.