Search

Background When using deep learning models, one of the most critical vulnerabilities is their exposure to adversarial inputs, which can cause wrong decisions (e.g., incorrect classification of an image) with minor perturbations. To address this vulnerability, it becomes necessary to retrain the affected model against adversarial inputs as part of the software testing process. In order to make this process energy efficient, data scientists need support on which are the best guidance metrics for reducing the adversarial inputs to create and use during testing, as well as optimal dataset configurations. Aim We examined six guidance metrics for retraining deep learning models, specifically with convolutional neural network architecture, and three retraining configurations. Our goal is to improve the convolutional neural networks against the attack of adversarial inputs with regard to the accuracy, resource utilization and execution time from the point of view of a data scientist in the context of image classification. Method We conducted an empirical study using five datasets for image classification. We explore: (a) the accuracy, resource utilization, and execution time of retraining convolutional neural networks with the guidance of six different guidance metrics (neuron coverage, likelihood-based surprise adequacy, distance-based surprise adequacy, DeepGini, softmax entropy and random), (b) the accuracy and resource utilization of retraining convolutional neural networks with three different configurations (one-step adversarial retraining, adversarial retraining and adversarial fine-tuning). Results We reveal that adversarial retraining from original model weights, and by ordering with uncertainty metrics, gives the best model w.r.t. accuracy, resource utilization, and execution time. Conclusions Although more studies are necessary, we recommend data scientists use the above configuration and metrics to deal with the vulnerability to adversarial inputs of deep learning models, as they can improve their models against adversarial inputs without using many inputs and without creating numerous adversarial inputs. We also show that dataset size has an important impact on the results.

This article describes the development of an optimal and predictive energy management system (EMS) for a microgrid with a high photovoltaic (PV) power contribution. The EMS utilizes a predictive long-short-term memory (LSTM) neural network trained on real PV power and consumption data. Optimal EMS decisions focus on managing the state of charge (SoC) of the battery energy storage system (BESS) within defined limits and determining the optimal power contributions from the microgrid components. The simulation utilizes MATLAB R2023a to solve a mixed-integer optimization problem and HOMER Pro 3.14 to simulate the microgrid. The EMS solves this optimization problem for the current sampling time (t) and the immediate sampling time (t+1), which implies a prediction of one hour in advance. An upper-layer decision algorithm determines the operating state of the BESS, that is, to charge or discharge the batteries. An economic and technical impact analysis of our approach compared to two EMSs based on a pure economic optimization approach and a peak-shaving algorithm reveals superior BESS integration, achieving 59% in demand satisfaction without compromising the life of the equipment, avoiding inexpedient power delivery, and preventing significant increases in operating costs.

Aim of study: To evaluate the effects of two different prescribed burning strategies on ectomycorrhizal (ECM) fungal species in Pinus cubensis-dominated natural forest. Area of study: Yateras Silvicultural Base Business Unit, Guantánamo, Cuba. Material and methods: In June 2015, six plots (20 × 50 m) were subjected to forward or back burning. Nine interval samplings (performed 1 week before and up to 120 days after prescribed burning) were undertaken to determine the total number of sporocarps and to evaluate the effect of fire on the soil. Main results: Eight ECM species were collected from the study plots. Suillus sp. and Amanita muscaria started fruiting 15 and 60 days after the fire, respectively. Boletus sp., Suillus brevipes, Suillus decipiens, Suillus sp., Amanita muscaria, Lactarius semisanguifluus, Scleroderma stellatum and Pisolithus arhizus were found before and after prescribed burning. Sporocarp numbers showed an increasing trend after fire and significantly recovered 75 days after forward or back burning and were significantly higher 120 days after forward burning compared to unburned plots. The ECM fungal community in the heading fire and the backfire plots did not differ significantly. However, non-metric multidimensional scaling confirmed that ECM composition differed over time. According to a Mantel test, the sampling time after prescribed burning accounted for 64% of the variation in ECM composition, followed by edaphic factors (26%) such as organic matter and Na. Research highlights: This preliminary study suggests that low-intensity prescribed burning does not have a negative effect on ECM fungal dynamics in humid tropical forests.

Este artículo reflexiona sobre la experiencia de la Universidad de Aysén como subversidad. Para ello, analizamos cómo la Universidad ha traducido las políticas nacionales y las expectativas de la comunidad local en prácticas institucionales orientadas a concretar el proyecto territorial. Así, en primer lugar, exploramos la imposibilidad de la Universidad de convertirse en subversidad en función de las tensiones que derivan de las políticas de mercado. En segundo lugar, indagamos en la posibilidad de la subversidad a partir de la constitución del Consejo Social. Este análisis nos permite concluir que la Universidad contiene una promesa que descansa en la capacidad del Consejo Social de repensar los principios fundantes de la Universidad.

En el presente artículo se analizan los usos del suelo existentes en las zonas de inundación frecuente de las áreas urbanas españolas. Estos se clasifican en grupos, distinguiendo las coberturas naturales de las antrópicas y, dentro de ellas, los usos compatibles y no compatibles con las inundaciones fluviales. El análisis realizado determina y compara el reparto de usos en el espacio fluvial y en las áreas urbanas, estudiando la ocupación de los ríos según la región biogeográfica a la que pertenecen y su temporalidad de flujo. Los resultados muestran que (1) los ríos en las ciudades son potenciales infraestructuras verdes donde abundan las coberturas naturales, espacios ajardinados, agrícolas y ganaderos; (2) se ha ocupado casi tres cuartas partes del espacio fluvial en las áreas urbanas; y (3) la presión antrópica es mayor en los ríos de la región mediterránea y, sobre todo, en los cursos de caudal intermitente. Este trabajo, pone de manifiesto la falta de una ordenación territorial coherente con los espacios fluviales y puede ser útil para contribuir a una sistematización metodológica que permita ampliar y contrastar el conocimiento de esta línea científica.

A formal language for game design is an endeavor many academics and industry personalities have been tackling since the mid-nineties. One of the most renowned formal models for game design, the MDA Framework, includes steps to delimit and conceptualize the experience with a top-down approach. There is, however, a significant lack of high detail models for mechanic construction as well as difficulty balancing. In this paper we propose two formal models for novice designers. The 5-Part-Model (5PM) for building and diagnosing game mechanics, and the Dimensions of Challenge (DoC) formal model for balancing and fine-tuning difficulty in games.

El presente artículo muestra la reconciliación entre los ciudadanos de Arévalo (Ávila, España) y sus ríos, a través de la ejecución de un proyecto de mejora y acondicionamiento paisajístico. Las claves del éxito de la recuperación de este espacio fluvial, antes degradado y abocado al olvido, fueron tres, fundamentalmente. En primer lugar, se confeccionó un meditado proyecto de recuperación paisajística de escasa dotación presupuestaria, cuya ejecución se realizó satisfactoriamente. En segundo lugar, la idea de actuar en el paisaje fluvial de este pueblo recibió un importante impulso político. Por último, cabe destacar la sensibilidad ambiental con la que se desarrolló el proyecto en todas sus fases, así como la involucración de la propia ciudadanía, a quien se hizo partícipe en todo momento del proceso de ejecución. La transformación del paisaje y su revalorización, así como la recuperación del patrimonio natural y cultural ligado a los ríos de Arévalo, han mejorado la calidad de vida de sus habitantes. La metodología empleada, que es replicable para futuras intervenciones en esta o en otras localidades, así como los resultados más relevantes, se exponen en el presente artículo.

How to deploy and manage, in an efficient and adaptive way, complex applications across multiple heterogeneous cloud platforms is one of the problems that have emerged with the cloud revolution. In this paper we present context, motivations and objectives of the EU research project SeaClouds, which aims at enabling a seamless adaptive multi-cloud management of complex applications by supporting the distribution, monitoring and migration of application modules over multiple heterogeneous cloud platforms. After positioning SeaClouds with respect to related cloud initiatives, we present the SeaClouds architecture and discuss some of its aspect, such as the use of the OASIS standard TOSCA and the compatibility with the OASIS CAMP initiative.

El proyecto modernizador aplicado en Chile durante las últimas décadas ha consolidado una estructura productiva preponderantemente rentista, dependiente de los ciclos del precio de las materias primas y de mercados altamente concentrados. A partir del proceso de descentralización en curso, que modifica la interrelación entre los actores territoriales más preponderantes, esta investigación identifica las principales barreras que impiden o retrasan la transformación productiva a nivel territorial, especialmente al considerar las nuevas facetas de la descentralización en el ámbito político-institucional. Para esto, se consideran los casos de las Agencias Regionales de Desarrollo Productivo de las regiones de Valparaíso y Coquimbo. En virtud del análisis realizado, se evidencia que los factores económicos y sociales son fundamentales para entender el éxito o fracaso de una política de desarrollo productivo a nivel territorial, siendo el factor institucional insuficiente para explicar esto de forma integral. Se concluye que es esencial el proceso que fortalece las capacidades endógenas de los territorios, puesto que son el sustento de la nueva fase de la descentralización.

The Rufous Tree Rat, Diplomys labilis, is known to range geographically from Central Panamá to Colombia and probably Ecuador. It occurs in a variety of habitat types where it moves by the branches in the high canopy at night. We report the Rufous Tree Rat for Costa Rica, ranging outside its geographical distribution about 490 km west from its type locality.

This article aims to provide a regional view of poverty in Spain from the perspective of the Europe 2020 strategy emphasizing the changes resulting from the economic crisis. This analysis is performed under the objective of inclusive growth that secures the European 2020 Strategy (EU2020S), formally adopted by the European Union in 2010, to reduce the levels of poverty and social exclusion The survey study of 13 socio-economic indicators at the regional level and the identification of relationships between the socio-economic indicators of the EU2020S has allowed (i) to place Spain and its regions in the European context, (ii) to evidence the changes resulting from the crisis, (iii) to identify relationships between the stu- died variables and (iv) to propose clusters of regions according to their distance to the objectives of the EU2020S.

The starting point of this paper is a system described in form of a UML class diagram where system states are characterized by OCL invariants and system transitions are defined by OCL pre- and postconditions. The aim of our approach is to assist the developer in learning about the consequences of the described system states and transitions and about the formal implications of the properties that are explicitly given. We propose to draw conclusions about the stated constraints by translating the UML and OCL model into the algebraic specification language and system Maude, which is based on rewrite logic. We will concentrate in this paper on employing Maude's capabilities for state search. Maude's state search offers the possibility to describe a start configuration of the system and then explore all configurations reachable by rewriting. The search can be adjusted by formulating requirements for the allowed states and the allowed transitions.

e-Motions is an Eclipse-based visual timed model transformation framework with a Real-Time Maude semantics that supports the usual Maude formal analysis methods, including simulation, reachability analysis, and LTL model checking. e-Motions is characterized by a novel and powerful set of constructs for expressing timed behaviors. In this paper we illustrate the use of these constructs --- and thereby implicitly investigate their suitability to define real-time systems in an intuitive way --- to define and formally analyze two prototypical and very different real-time systems: (i) a simple round trip time protocol for computing the time it takes a message to travel from one node to another, and back; and (ii) the EDF scheduling algorithm.