Abstract
The structural changes in the energy sector caused by renewable sources and digitization have resulted in an increased use of Artificial Intelligence (AI), including Machine Learning (ML) models. However, these models' black-box nature and complexity can create issues with transparency and trust, thereby hindering their interpretability. The use of Explainable AI (XAI) can offer a solution to these challenges. This paper explores the application of an XAI-based framework to analyze and evaluate a photovoltaic energy generation forecasting problem and contribute to the trustworthiness of ML solutions.

Abstract
This study presents a new tridimensional perspective on Castelo Velho de Safara (Moura), one of the great walled settlements of the Iron Age/Roman Republic by the Guadiana River, obtained through a high-resolution survey using a drone integrated with a LiDAR sensor. The outline of the walls was defined in more detail, which meant revising the occupation area, now estimated at circa 1.36 hectares. Several unknown elements were detected, such as the entrance area and other possible defensive structures. The data obtained for the Castelo Velho de Safara demonstrate the potential of LiDAR for understanding the topographical characteristics of this type of fortified enclosure, whose structural remains are not always clear to the naked eye. © 2023, APEQ - Associacao Portuguesa para o Estudo do Quaternario. All rights reserved.

Abstract
Purpose Technological advances in the last decade have caused both business and economic sectors to seek for new ways to adapt their business models to a connected data-centric era. Family businesses have also been forced to leave behind traditional strategies rooted in family stimuli and ties and to adapt their actions in digital environments. In this context, this study aims to identify major online marketing strategies, business models and technology applications developed to date by family firms. Methodology: Upon a systematic literature review, we develop a multiple correspondence analysis (MCA) under the homogeneity analysis of variance by means of alternating least squares (HOMALS) framework programmed in the R language. Based on the results, the analyzed contributions are visually analyzed in clusters. Design/methodology/approach Upon a systematic literature review, we develop an MCA under the HOMALS framework programmed in the R language. Based on the results, the analyzed contributions are visually analyzed in clusters. Findings Relevant indicators are identified for the successful development of digital family businesses classified in the following three categories: (1) digital business models, (2) digital marketing techniques and (3) technology applications. The first category consists of four digital business models: mobile marketing, e-commerce, cost per click, cost per mile and cost per acquisition. The second category includes six digital marketing techniques: search marketing (search engine optimization and search engine marketing (SEM) strategies), social media marketing, social ads, social selling, websites and online reputation optimization. Finally, the third category consists of the following aspects: digital innovation, digital tools, innovative marketing, knowledge discovery and online decision making. In addition, five research propositions are developed for further discussion and future research. Originality/value To the best of our knowledge, this study is the first to cover this research topic applying the emerging programming language R for the development of an MCA under the HOMALS framework.

Abstract
Background: In peripheral arterial disease (PAD) patients with intermittent claudication (IC), the combination of aerobic and resistance exercises could counteract muscle loss and attenuate disease progression. This study analyzed the effects of six months of a combined exercise program on walking ability, lower limb body composition, cardiovascular risk factors, and Ankle-Brachial Index (ABI). Methods: Twenty-three patients (age 63.2 +/- 1.5 years and ABI 0.58 +/- 0.07) with PAD and IC were allocated to a control group (CG) or a supervised exercise group (SUP). Ten patients underwent six months of treadmill walking combined with resistance exercises, three times a week. The CG (N.=13) received a recommendation for walking. All patients were measured at baseline (M0), after three months (M3), and six months (M6). Results: During constant treadmill protocol, the claudication onset time/distance (COT/COD), absolute claudication time/distance (ACT/ACD), and number of pauses of overall patients significantly improved at M3 and M6. Between groups were found significant differences in COT and COD at M6 (P=0.005 and P=0.007, respectively); and in ACT and ACD at M3 (P=0.003 for both) and at M6 (P=0.005 and P=0.005, respectively), with major improvements in the SUP. Over the six months, a significant group effect was found in fat-free mass (P=0.041) and predicted muscle mass (P=0.039) of the lower ABI leg, with greater improvements in the SUP. Conclusions: A supervised exercise program that combines aerobic and resistance training improves PAD symptoms and has additional benefits for patients. Patients in the program showed improvements in walking ability, lower-limb body composition, perceived exertion, and heart rate during treadmill walking.

Abstract
Advancements in semiconductor technology no longer occur at the pace the industry had been accustomed to. We have entered what is considered by many to be the post-Moore era. In order to continue scaling performance, increasingly heterogeneous architectures are being developed and the use of special purpose accelerators is on the rise. One notable example are Field-Programmable-Gate-Arrays (FPGAs), both in the data-center and embedded spaces. Advances in FPGA features and tools is allowing for critical kernels to be accelerated on specialized hardware without fabrication costs. However, re-targeting code to such heterogeneous platforms still requires significant refactoring of the compute intensive kernels, as well as knowledge of parallel compute and hardware design concepts for maximization of performance. We present Tribble, a source-to-source framework under active development, capable of transforming regular C/C++ programs for execution on heterogeneous architectures. This includes transforming the target kernel source code so that it is amenable for circuit generation while keeping the original version for software execution, inserting code for task and memory management and injecting a scheduler algorithm.

Abstract
In the last years, the number of machine learning algorithms and their parameters has increased significantly. On the one hand, this increases the chances of finding better models. On the other hand, it increases the complexity of the task of training a model, as the search space expands significantly. As the size of datasets also grows, traditional approaches based on extensive search start to become prohibitively expensive in terms of computational resources and time, especially in data streaming scenarios. This paper describes an approach based on meta-learning that tackles two main challenges. The first is to predict key performance indicators of machine learning models. The second is to recommend the best algorithm/configuration for training a model for a given machine learning problem. When compared to a state-of-the-art method (AutoML), the proposed approach is up to 130x faster and only 4% worse in terms of average model quality. Hence, it is especially suited for scenarios in which models need to be updated regularly, such as in streaming scenarios with big data, in which some accuracy can be traded for a much shorter model training time.