• Publicationsclose
• Research softwareclose
• 2022-2022close
• Digital Humanities and Cultural Heritage close
• Knowmad Institut close

Il processo che conduce alla degenerazione di un corpo, di un sistema politico o di qualsiasi altra forma di organizzazione culturale può apparire inevitabile. Eppure, secondo Francesco Patrizi da Cherso (1529-1597), esistono degli strumenti per minimizzarne le conseguenze. Nel presente contributo, a partire dallo strano caso dell’eterno raffreddore di Tarquinia Molza, ripercorreremo la storia del concetto di catastrofe nella filosofia patriziana, focalizzandoci in particolare sulle diverse forme di «corruzione» che ritmano tanto la vita di un singolo corpo quanto quella dell’intero universo. Basandoci sulla lettura di tre testi, composti nell’arco di più di quarant’anni – la Città felice (1553), i Dialoghi della historia (1560) e i Paralleli militari (1594-1595) – scopriremo come la dimensione concettuale della catastrofi, dal punto di vista di Patrizi, oscilli tra natura, storia e mito; e come, in un complicato intreccio di medicina, storia e filosofia, Patrizi ricerchi le condizioni affinché l’umanità possa prevedere (e quindi, in qualche modo, controllare) il decorso inalterabile della natura, adattandosi ad esso. Le processus qui conduit à la dégénérescence d’un corps, d’un système politique ou de toute autre forme d’organisation culturelle peut sembler inévitable. Pourtant, selon Francesco Patrizi da Cherso (1529-1597), il existe des moyens de minimiser ses conséquences. Dans cet article, en partant du cas étrange du rhume éternel de Tarquinia Molza, nous retracerons l’histoire du concept de catastrophe dans la philosophie de Patrizi, en nous concentrant en particulier sur les différentes formes de « corruption » qui rythment la vie d’un seul corps autant que celle de l’univers entier. À partir de la lecture de trois textes composés sur une période de plus de quarante ans – la Città felice (1553), les Dialoghi della historia (1560) et les Paralleli militari (1594-1595) – nous verrons comment la dimension conceptuelle de la catastrophe, du point de vue de Patrizi, oscille entre nature, histoire et mythe ; et comment, dans un entrelacement complexe de médecine, d’histoire et de philosophie, Patrizi recherche les conditions permettant à l’homme de prévoir (et donc, dans une certaine mesure, de contrôler) le cours inaltérable de la nature et de s’y adapter. The process that leads to the degeneration of a body, a political system or any other form of cultural organisation may seem inevitable. Yet, according to Francesco Patrizi da Cherso (1529-1597), there are ways to minimise its consequences. In this contribution, starting with the strange case of the eternal cold of Tarquinia Molza, we will trace the history of the concept of catastrophe in Patrizi’s philosophy, focusing in particular on the different forms of “corruption” that punctuate the life of a single body as much as that of the entire universe. Based on the reading of three texts composed over a period of more than forty years –the Città felice (1553), the Dialoghi della historia (1560) and the Paralleli militari (1594-1595)– we will discover how the conceptual dimension of catastrophe, from Patrizi’s point of view, oscillates between nature, history and myth, and how, in a complicated interweaving of medicine, history and philosophy, Patrizi seeks the conditions for mankind to foresee (and thus, to some extent, control) the unalterable course of nature and adapt to it.

Practice abstracts and videos are often envisioned within the multi-actor approach (MAA), which is typical of Horizon 2020 projects and brings together a range of different stakeholders, spanning people from science, practice, or anyone who can help tackle the objectives of the project. All experience and knowledge are therefore taken into account and the partners co-create results together, to address real-life problems through innovative and intersectoral approaches. In this perspective, practice abstracts and videos are intended to communicate a short summary of information or recommendations that have been observed or developed during the lifetime of a project, as well as to provide good practices as role models for replication in other contexts. As part of SmartCulTour’s WP6, Task 6.4 included the production of practice abstracts and the recording of relevant videos on interventions that were identified in the SmartCulTour Labs for their relevance as interesting and insightful cultural tourism interventions, and were either taking place in the region and supported by the Lab’s activities, implemented within the SmartCulTour project or co-designed by Labs’ stakeholders for future implementation. These are innovative/promising direct and indirect interventions, or a combination thereof, for the sustainable development of cultural tourism, with potential/possible impacts on sustainability and resilience of the destination. Accordingly, each Lab Manager was requested to produce one or more practice abstracts and videos, by providing practical information and outlining the main objectives and findings related to the selected intervention(s). Deliverable 6.2 gathers the texts of the practice abstracts produced by each of the six SmartCulTour Living Labs, along with the links to access the videos on the SmartCulTour website and YouTube channel. Two typologies of videos were produced: presentation video(s) of each Living Lab, and videos showcasing the intervention(s) selected, designed and/or implemented by each of them. These materials were developed primarily with a communication purpose to facilitate information sharing with those who can directly apply the findings, namely cultural tourism practitioners, policy-makers, etc. Local stakeholders were also involved in the production of the practice videos, so as to let them shape the storytelling for their own destination, while leveraging local ownership of the interventions. In order to boost their dissemination and outreach, the practice abstracts are available both in English and in the Living Lab’s local language, while the videos are subtitled in English or in the local language, as needed.

[ES] Durante la última década, los medios de comunicación han experimentado una revolución, alejándose de la televisión convencional hacia las plataformas de contenido bajo demanda. Además, esta revolución no ha cambiado solamente la manera en la que nos entretenemos, si no también la manera en la que aprendemos. En este sentido, las plataformas de contenido educativo bajo demanda también han proliferado para proporcionar recursos educativos de diversos tipos. Estas nuevas vías de distribución de contenido han llegado con nuevos requisitos para mejorar la accesibilidad, en particular las relacionadas con las dificultades de audición y las barreras lingüísticas. Aquí radica la oportunidad para el reconocimiento automático del habla (RAH) para cumplir estos requisitos, proporcionando subtitulado automático de alta calidad. Este subtitulado proporciona una base sólida para reducir esta brecha de accesibilidad, especialmente para contenido en directo o streaming. Estos sistemas de streaming deben trabajar bajo estrictas condiciones de tiempo real, proporcionando la subtitulación tan rápido como sea posible, trabajando con un contexto limitado. Sin embargo, esta limitación puede conllevar una degradación de la calidad cuando se compara con los sistemas para contenido en diferido u offline. Esta tesis propone un sistema de RAH en streaming con baja latencia, con una calidad similar a un sistema offline. Concretamente, este trabajo describe el camino seguido desde el sistema offline híbrido inicial hasta el eficiente sistema final de reconocimiento en streaming. El primer paso es la adaptación del sistema para efectuar una sola iteración de reconocimiento haciendo uso de modelos de lenguaje estado del arte basados en redes neuronales. En los sistemas basados en múltiples iteraciones estos modelos son relegados a una segunda (o posterior) iteración por su gran coste computacional. Tras adaptar el modelo de lenguaje, el modelo acústico basado en redes neuronales también tiene que adaptarse para trabajar con un contexto limitado. La integración y la adaptación de estos modelos es ampliamente descrita en esta tesis, evaluando el sistema RAH resultante, completamente adaptado para streaming, en conjuntos de datos académicos extensamente utilizados y desafiantes tareas basadas en contenidos audiovisuales reales. Como resultado, el sistema proporciona bajas tasas de error con un reducido tiempo de respuesta, comparables al sistema offline. [CA] Durant l'última dècada, els mitjans de comunicació han experimentat una revolució, allunyant-se de la televisió convencional cap a les plataformes de contingut sota demanda. A més a més, aquesta revolució no ha canviat només la manera en la que ens entretenim, si no també la manera en la que aprenem. En aquest sentit, les plataformes de contingut educatiu sota demanda també han proliferat pera proporcionar recursos educatius de diversos tipus. Aquestes noves vies de distribució de contingut han arribat amb nous requisits per a millorar l'accessibilitat, en particular les relacionades amb les dificultats d'audició i les barreres lingüístiques. Aquí radica l'oportunitat per al reconeixement automàtic de la parla (RAH) per a complir aquests requisits, proporcionant subtitulat automàtic d'alta qualitat. Aquest subtitulat proporciona una base sòlida per a reduir aquesta bretxa d'accessibilitat, especialment per a contingut en directe o streaming. Aquests sistemes han de treballar sota estrictes condicions de temps real, proporcionant la subtitulació tan ràpid com sigui possible, treballant en un context limitat. Aquesta limitació, però, pot comportar una degradació de la qualitat quan es compara amb els sistemes per a contingut en diferit o offline. Aquesta tesi proposa un sistema de RAH en streaming amb baixa latència, amb una qualitat similar a un sistema offline. Concretament, aquest treball descriu el camí seguit des del sistema offline híbrid inicial fins l'eficient sistema final de reconeixement en streaming. El primer pas és l'adaptació del sistema per a efectuar una sola iteració de reconeixement fent servir els models de llenguatge de l'estat de l'art basat en xarxes neuronals. En els sistemes basats en múltiples iteracions aquests models son relegades a una segona (o posterior) iteració pel seu gran cost computacional. Un cop el model de llenguatge s'ha adaptat, el model acústic basat en xarxes neuronals també s'ha d'adaptar per a treballar amb un context limitat. La integració i l'adaptació d'aquests models és àmpliament descrita en aquesta tesi, avaluant el sistema RAH resultant, completament adaptat per streaming, en conjunts de dades acadèmiques àmpliament utilitzades i desafiants tasques basades en continguts audiovisuals reals. Com a resultat, el sistema proporciona baixes taxes d'error amb un reduït temps de resposta, comparables al sistema offline. [EN] Over the last decade, the media have experienced a revolution, turning away from the conventional TV in favor of on-demand platforms. In addition, this media revolution not only changed the way entertainment is conceived but also how learning is conducted. Indeed, on-demand educational platforms have also proliferated and are now providing educational resources on diverse topics. These new ways to distribute content have come along with requirements to improve accessibility, particularly related to hearing difficulties and language barriers. Here is the opportunity for automatic speech recognition (ASR) to comply with these requirements by providing high-quality automatic captioning. Automatic captioning provides a sound basis for diminishing the accessibility gap, especially for live or streaming content. To this end, streaming ASR must work under strict real-time conditions, providing captions as fast as possible, and working with limited context. However, this limited context usually leads to a quality degradation as compared to the pre-recorded or offline content. This thesis is aimed at developing low-latency streaming ASR with a quality similar to offline ASR. More precisely, it describes the path followed from an initial hybrid offline system to an efficient streaming-adapted system. The first step is to perform a single recognition pass using a state-of-the-art neural network-based language model. In conventional multi-pass systems, this model is often deferred to the second or later pass due to its computational complexity. As with the language model, the neural-based acoustic model is also properly adapted to work with limited context. The adaptation and integration of these models is thoroughly described and assessed using fully-fledged streaming systems on well-known academic and challenging real-world benchmarks. In brief, it is shown that the proposed adaptation of the language and acoustic models allows the streaming-adapted system to reach the accuracy of the initial offline system with low latency.

Resumen El artículo es un estudio detallado del panfleto An Essay in Defense of the Female Sex (1696). Ofrece una reconstrucción de su estructura argumentativa y un análisis de sus estrategias retóricas. El marco teórico y metodológico empleado es histórico y sistemático: buscamos contextualizar el escrito en el campo intelectual de la ilustración británica en los albores del siglo XVIII, así como también interponer una perspectiva de género contemporánea para el examen de los argumentos empleados, del despliegue retórico del discurso y de las conclusiones filosóficamente más robustas. En especial, el artículo defiende la emergencia de una dicotomía naturaleza-historia que permite a la autora, por un lado, desenmascarar una situación de dominio masculino (críticamente reconocible hoy día como antecedente directo de las discusiones feministas contemporáneas) y, por el otro, ofrecer una vía retórica para la conformación y reconfiguración de un ethos y un pathos femenino. Abstract The article is a detailed study of the pamphlet An Essay in Defense of the Female Sex (1696). It offers a reconstruction of its argumentative structure and an analysis of its rhetorical strategies. The theoretical and methodological framework used is both historical and systematic: we seek to contextualize the work in the intellectual field of British Enlightenment at the dawn of the 18th century, as well as to interpose a contemporary gender perspective for the examination of its arguments, of the rhetorical deployment of its discourse and its more philosophically robust conclusions. In particular, the article defends the emergence of a nature-history dichotomy that allows the author, on the one hand, to unmask a situation of male dominance (that we can today critically recognize as a direct antecedent of contemporary feminist discussions) and, on the other, to offer a rhetorical pathway for the conformation and reconfiguration of a feminine ethos and pathos.

Research background: In a modern economy, full of complexities, ensuring a business' financial stability, and increasing its financial performance and competitiveness, has become especially difficult. Then, monitoring the company's financial situation and predicting its future develop-ment becomes important. Assessing the financial health of business entities using various models is an important area in not only scientific research, but also business practice.Purpose of the article: This study aims to predict the bankruptcy of companies in the engineer-ing and automotive industries of the Slovak Republic using a multilayer neural network and logistic regression. Importantly, we develop a novel an early warning model for the Slovak engi-neering and automotive industries, which can be applied in countries with undeveloped capital markets. Methods: Data on the financial ratios of 2,384 companies were used. We used a logistic regres-sion to analyse the data for the year 2019 and designed a logistic model. Meanwhile, the data for the years 2018 and 2019 were analysed using the neural network. In the prediction model, we analysed the predictive performance of several combinations of factors based on the industry sector, use of the scaling technique, activation function, and ratio of the sample distribution to the test and training parts. Findings & value added: The financial indicators ROS, QR, NWC/A, and PC/S reduce the likelihood of bankruptcy. Regarding the value of this work, we constructed an optimal network for the automotive and engineering industries using nine financial indicators on the input layer in combination with one hidden layer. Moreover, we developed a novel prediction model for bank-ruptcy using six of these indicators. Almost all sampled industries are privatised, and most com-panies are foreign owned. Hence, international companies as well as researchers can apply our models to understand their financial health and sustainability. Moreover, they can conduct com-parative analyses of their own model with ours to reveal areas of model improvements. KEGA [001PU-4/2022]; Scientific Grant Agency of the Ministry of Education, Science, Research, and Sport of the Slovak Republic; Slovak Academy Sciences [1/0590/22] 1/0590/22; Kultúrna a Edukacná Grantová Agentúra MŠVVaŠ SR, KEGA: 001PU-4/2022

Automated Essay Scoring (AES) automatically allocates scores to essays at scale and may help teachers reduce the heavy burden during grading activities. Recently, researchers have deployed neural-based AES approaches to improve upon the state-of-the-art AES performance. These neural-based AES methods mainly take student essays as the sole input and focus on learning the relationship between student essays and essay scores through deep neural networks. However, their only product, the predicted holistic score, is far from providing adequate pedagogical information, such as automated writing evaluation (AWE). In this work, we propose Topic-aware BERT, a new method of learning relations among scores, student essays, as well as topical information in essay instructions. Beyond improving the AES benchmark performance, Topic-aware BERT can automatically retrieve key topical sentences in student essays by probing self-attention maps in intermediate layers. We evaluate the performance of Topic-aware BERT of different variants to (i) perform AES and (ii) retrieve key topical sentences using the open dataset Automated Student Assessment Prize and a manually annotated dataset. Our experiments show that Topic-aware BERT achieves a strong AES performance compared with the previous best neural-based AES methods and demonstrates effectiveness in identifying key topical sentences in argumentative essays. This article belongs to the Special Issue Artificial Intelligence Solutions and Applications for Distributed Systems in Smart Spaces

AbstractThis paper discusses how the use of AI (artificial intelligence) detected later prehistoric field systems provides a more reliable base for reconstructing palaeodemographic trends, using the Netherlands as a case study. Despite its long tradition of settlement excavations, models that could be used to reconstruct (changes in) prehistoric land use have been few and often relied on (insufficiently mapped) nodal data points such as settlements and barrows. We argue that prehistoric field systems of field plots beset on all sides by earthen banks—known as Celtic fields—are a more suitable (i.e. less nodal) proxy for reconstructing later prehistoric land use.For four 32.25 km2 case study areas in different geogenetic regions of the Netherlands, prehistoric land use surface areas are modelled based on conventional methods and the results are compared to the results we obtained by using AI‐assisted detection of prehistoric field systems. The nationally available LiDAR data were used for automated detection. Geotiff DTM images were fed into an object detection algorithm (based on the YOLOv4 framework and trained with known Dutch sites), and resultant geospatial vectors were imported into GIS.Our analysis shows that AI‐assisted detection of prehistoric embanked field systems on average leads to a factor 1.84 increase in known surface areas of Celtic fields. Modelling the numbers of occupants from this spatial coverage, yields population sizes of 37–135 persons for the case study regions (i.e. 1.15 to 4.19 p/km2). This range aligns well with previous estimates and offers a more robust and representative proxy for palaeodemographic reconstructions. Variations in land use coverage between the regions could be explained by differences in present‐day land use and research intensity. Particularly the regionally different extent of forestlands and heathlands (ideal for the (a) preservation and (b) automated LiDAR detection of embanked field systems) explains minor variations between the four case study regions.

In order to prevent the illegal export of paintings abroad, a museum examination using various methods for studying a work of art is carried out. At the same time, an analysis is also made of historical, art history, financial and other information and documents confirming the painting’s authenticity — provenance. Automation of such examination is hampered by the need to take into account numerical values of visual features, quality indicators, and verbal descriptions from provenance. In this paper, we consider the problem of automatic multi-task classification of paintings for museum expertise. A system architecture is proposed that checks provenance, implements a fine-grained image analysis (FGIA) of visual image features, and automatically classifies a painting by authorship, genre, and time of creation. Provenance is contained in a knowledge graph; for its vectorization, it is proposed to use a graph2vec type encoder with an attention mechanism. Fine-grained image analysis is proposed to be performed using searching discriminative regions (SDR) and learning discriminative regions (LDR) allocated by convolutional neural networks. To train the classifier, a generalized loss function is proposed. A data set is also proposed, including provenance and images of paintings by European and Ukrainian artists. Для запобігання незаконному вивезенню картин за кордон прово- диться музейна експертиза з використанням різних методів дослідження твору мистецтва, зокрема аналіз історичних, мистецтвознавчих, фінансових та інших відомостей і документів, що підтверджують справжність картин – провенансу. Автоматизація такої експертизи ускладнюється необхідністю враховувати числові значення візуальних ознак, показників якості та словесні описи з провенансу. Розглянуто завдання автоматичної багатозадачної класифікації картин під час музейної експертизи. Запропоновано архітектуру системи, яка перевіряє провенанс, реалізує детальний аналіз (FGIA) візуальних ознак зображення та виконує автоматичну класифікацію картини за авторством, жанром та часом створення. Провенанс міститься у графі знань, для векторизації якого запропоновано використовувати енкодер типу graph2vec з механізмом уваги, а детальний аналіз пропонується виконувати за допомогою пошукових відмітних регіонів (SDR) та навчальних відмітних регіонів (LDR), що виділяються згортковими нейронними мережами. Для навчання класифікатора запропоновано узагальнену функцію втрат, а також набір даних, що включає провенанс та зображення картин європейських та українських художників.