• Digital Humanities and Cultural Heritage
• Publications close
• Research data close
• Other research products close
• FR close
• arXiv.org e-Print Archive close
• Climate of the Past (CP) close

The unprecedented success of image reconstruction approaches based on deep neural networks has revolutionised both the processing and the analysis paradigms in several applied disciplines. In the field of digital humanities, the task of digital reconstruction of ancient frescoes is particularly challenging due to the scarce amount of available training data caused by ageing, wear, tear and retouching over time. To overcome these difficulties, we consider the Deep Image Prior (DIP) inpainting approach which computes appropriate reconstructions by relying on the progressive updating of an untrained convolutional neural network so as to match the reliable piece of information in the image at hand while promoting regularisation elsewhere. In comparison with state-of-the-art approaches (based on variational/PDEs and patch-based methods), DIP-based inpainting reduces artefacts and better adapts to contextual/non-local information, thus providing a valuable and effective tool for art historians. As a case study, we apply such approach to reconstruct missing image contents in a dataset of highly damaged digital images of medieval paintings located into several chapels in the Mediterranean Alpine Arc and provide a detailed description on how visible and invisible (e.g., infrared) information can be integrated for identifying and reconstructing damaged image regions. Comment: 26 pages

This work is based upon the collaboration in the COST Action CA18209 – NexusLinguarum “European network for Web-centred linguistic data science”, supported by COST (European Cooperation in Sci- ence and Technology). Marko Robnik-Šikonja received financial support from the Slovenian Research Agency through core research programme P6-0411 and projects J6-2581, J7-3159, and V5-2297. Encar- nación Hidalgo Tenorio was financially supported by the European Social Fund, the Andalusian Gov- ernment, and the University of Granada (Project References: A-HUM-250-UGR18 & P18-FR-5020). Petya Osenova was partially supported by CLaDA-BG, the Bulgarian National Interdisciplinary Re- search e-Infrastructure for Resources and Technologies in favor of the Bulgarian Language and Cultural Heritage, and partially through the EU infrastructures CLARIN and DARIAH, Grant number DO01- 377/18.12.2020. Parliamentary and legislative debate transcripts provide an informative insight into elected politicians’ opinions, positions, and policy preferences. They are interesting for political and social sciences as well as linguistics and natural language processing (NLP) research. While exiting research studied individual parliaments, we apply advanced NLP methods to a joint and comparative analysis of six national parliaments (Bulgarian, Czech, French, Slovene, Spanish, and United Kingdom) between 2017 and 2020. We analyze emotions and sentiment in the transcripts from the ParlaMint dataset collection, and assess if the age, gender, and political orientation of speakers can be detected from their speeches. The results show some commonalities and many surprising differences among the analyzed countries. Slovenian Research Agency P6-0411, J6-2581, J7-3159, V5-2297 COST (European Cooperation in Science and Technology) University of Granada (A-HUM-250-UGR18 & P18-FR-5020) CLaDA-BG DO01-377/18.12.2020 Andalusian Government European Social Fund

The last deglaciation (21 to 8 ka) of the Eurasian ice sheet (EIS) is thought to have been responsible for a sea level rise of about 20 m. While many studies have examined the timing and rate of the EIS retreat during this period, many questions remain about the key processes that triggered the EIS deglaciation 21 kyr ago. Due to its large marine-based parts in the Barents-Kara (BKIS) and British Isles sectors, the BKIS is often considered to be a potential analogue of the current West Antarctic ice sheet (WAIS). Identifying the mechanisms that drove the EIS evolution might provide a better understanding of the processes at play in the West Antarctic destabilization. To investigate the relative impact of key drivers on the EIS destabilization, we used the threedimensional ice sheet model GRISLI (GRenoble Ice Shelf and Land Ice) (version 2.0) forced by climatic fields from five Paleoclimate Modelling Intercomparison Project phases 3 and 4 (PMIP3, PMIP4) Last Glacial Maximum (LGM) simulations. In this study, we performed sensitivity experiments to test the response of the simulated Eurasian ice sheets to surface climate, oceanic temperatures (and thus basal melting under floating ice tongues), and sea level perturbations. Our results highlight that the EIS retreat simulated with the GRISLI model is primarily triggered by atmospheric warming. Increased atmospheric temperatures further amplify the sensitivity of the ice sheets to sub-shelf melting. These results contradict those of previous modelling studies mentioning the central role of basal melting on the deglaciation of the marine-based Barents-Kara ice sheet. However, we argue that the differences with previous works are mainly related to differences in the methodology followed to generate the initial LGM ice sheet. Due to the strong sensitivity of EIS to the atmospheric forcing highlighted with the GRISLI model and the limited extent of the confined ice shelves during the LGM, we conclude by questioning the analogy between EIS and the current WAIS. However, because of the expected rise in atmospheric temperatures, the risk of hydrofracturing is increasing and could ultimately put the WAIS in a configuration similar to the past Eurasian ice sheet. International audience

After the discovery of the Higgs boson at the LHC, particle physics community is exploring and proposing next accelerators, to address the remaining open questions on the underlying mechanisms and constituents of the present universe. One of the studied possibilities is FCC (Future Circular Collider), a 100 km long collider at CERN. The feasibility study of this future proposed accelerator implies the definition of tolerances on magnets imperfections and of the strategies of correction in order to guarantee the target performances of the High Energy Booster ring. The efficiency of the correction scheme, used to control the orbit, directly bounds the corrector needs and magnet tolerances. Analytic formulae give a first estimation of the average rms value of the required linear correctors' strengths and of the allowed magnets misalignments and field quality along the entire ring. The distribution of the correctors along the ring is simulated in order to verify the quality of the residual orbit after the proposed correction strategy and compared with the analytical predictions. First specifications of the orbit correctors strength and tolerances for the alignment of the main elements of the ring are presented. The limits of the studied correction scheme and method are also discussed. International audience

We present a neuro-symbolic (NeSy) workflow combining a symbolic-based learning technique with a large language model (LLM) agent to generate synthetic data for code comment classification in the C programming language. We also show how generating controlled synthetic data using this workflow fixes some of the notable weaknesses of LLM-based generation and increases the performance of classical machine learning models on the code comment classification task. Our best model, a Neural Network, achieves a Macro-F1 score of 91.412% with an increase of 1.033% after data augmentation. 18 pages, 4 figures, forthcoming chapter in the book Generative Artificial Intelligence for Code - Impact of Large Language Models on Code Generation and Summarization in the book series Transactions on Computer Systems and Networks, Springer International audience

We revisit how super-Hubble cosmological fluctuations induce, at any time in the cosmic history, a non-vanishing spatial curvature of the local background metric. The random nature of these fluctuations promotes the curvature density parameter to a stochastic quantity for which we derive novel non-perturbative expressions for its mean, variance, higher moments and full probability distribution. For scale-invariant Gaussian perturbations, such as those favored by cosmological observations, we find that the most probable value for the curvature density parameter $\Omega_\mathrm{K}$ today is $-10^{-9}$, that its mean is $+10^{-9}$, both being overwhelmed by a standard deviation of order $10^{-5}$. We then discuss how these numbers would be affected by the presence of large super-Hubble non-Gaussianities, or, if inflation lasted for a very long time. In particular, we find that substantial values of $\Omega_\mathrm{K}$ are obtained if inflation lasts for more than a billion e-folds. Comment: 7 pages, 2 figures, uses RevTex. Misprints corrected, references added, matches published version

Reasoning by analogies permeates theoretical developments in physics and astrophysics, motivated by the unreachable nature of many phenomena at play. For example, analogies have been used to understand black hole physics, leading to the development of a thermodynamic theory for these objects and the discovery of the Hawking effect. The latter, which results from quantum field theory on black hole space-times, changed the way physicists approached this subject: what had started as a mere aid to understanding becomes a possible source of evidence via the research programme of `analogue gravity' that builds on analogue models for field effects. Some of these analogue models may and can be realised in the laboratory, allowing experimental tests of field effects. Here, we present a historical perspective on the connection between the Hawking effect and analogue models. We also present a literature review of current research, bringing history and contemporary physics together. We argue that the history of analogue gravity and the Hawking effect is divided into three distinct phases based on how and why analogue models have been used to investigate fields in the vicinity of black holes. Furthermore, we find that modern research signals a transition to a new phase, where the impetus for the use of analogue models has surpassed the problem they were originally designed to solve. Comment: Version accepted for publication in special issue of EPJH on modern developments in quantum theory