• Digital Humanities and Cultural Heritage
• Publications close
• Research data close
• Other research products close
• BI close
• Hal-Diderot close

International audience; We combined biostratigraphical analyses, archaeological surveys, and Glacial Isostatic Adjustment (GIA) models to provide new insights into the relative sea-level evolution in the northeastern Aegean Sea (eastern Mediterranean). In this area, characterized by a very complex tectonic pattern, we produced a new typology of sea-level index point, based on the foraminiferal associations found in transgressive marine facies. Our results agree with the sea-level history previously produced in this region, therefore confirming the validity of this new type of index point. The expanded dataset presented in this paper further demonstrates a continuous Holocene RSL rise in this portion of the Aegean Sea. Comparing the new RSL record with the available geophysical predictions of sea-level evolution indicates that the crustal subsidence of the Samothraki Plateau and the North Aegean Trough played a major role in controlling millennial-scale sea-level evolution in the area. This major subsidence rate needs to be taken into account in the preparation of local future scenarios of sea-level rise in the coming decades.

Abstract Artistic composition (the structural organization of pictorial elements) is often characterized by some basic rules and heuristics, but art history does not offer quantitative tools for segmenting individual elements, measuring their interactions and related operations. To discover whether a metric description of this kind is even possible, we exploit a deep-learning algorithm that attempts to capture the perceptual mechanism underlying composition in humans. We rely on a robust behavioral marker with known relevance to higher-level vision: orientation judgements, that is, telling whether a painting is hung “right-side up.” Humans can perform this task, even for abstract paintings. To account for this finding, existing models rely on “meaningful” content or specific image statistics, often in accordance with explicit rules from art theory. Our approach does not commit to any such assumptions/schemes, yet it outperforms previous models and for a larger database, encompassing a wide range of painting styles. Moreover, our model correctly reproduces human performance across several measurements from a new web-based experiment designed to test whole paintings, as well as painting fragments matched to the receptive-field size of different depths in the model. By exploiting this approach, we show that our deep learning model captures relevant characteristics of human orientation perception across styles and granularities. Interestingly, the more abstract the painting, the more our model relies on extended spatial integration of cues, a property supported by deeper layers.

Machine translation (MT) models used in industries with constantly changing topics, such as translation or news agencies, need to adapt to new data to maintain their performance over time. Our aim is to teach a pre-trained MT model to translate previously unseen words accurately, based on very few examples. We propose (i) an experimental setup allowing us to simulate novel vocabulary appearing in human-submitted translations, and (ii) corresponding evaluation metrics to compare our approaches. We extend a data augmentation approach using a pre-trained language model to create training examples with similar contexts for novel words. We compare different fine-tuning and data augmentation approaches and show that adaptation on the scale of one to five examples is possible. Combining data augmentation with randomly selected training sentences leads to the highest BLEU score and accuracy improvements. Impressively, with only 1 to 5 examples, our model reports better accuracy scores than a reference system trained with on average 313 parallel examples. 14 pages includince 3 of appendices

Marine phytoplankton are believed to account for more than 45% of photosynthetic net primary production on Earth, and hence are at the base of marine food webs and have an enormous impact on the entire Earth system. Their members are found across many of the major clades of the tree of life, including bacteria (cyanobacteria) and multiple eukaryotic lineages that acquired photosynthesis through the process of endosymbiosis. Our understanding of their distribution in marine ecosystems and their contribution to biogeochemical cycles have increased since they were first described in the 18th century. Here, we review historical milestones in marine phytoplankton research and how their roles were gradually understood, with a particular focus on insights derived from large-scale ocean exploration. We start from the first observations made by explorers and naturalists, review the initial identification of the main phytoplankton groups and the appreciation of their function in the influential Kiel and Plymouth schools that established biological oceanography, to finally outline the contribution of modern large-scale initiatives to understand this fundamental biological component of the ocean. Fil: Pierella Karlusich, Juan José. Centre National de la Recherche Scientifique. Ecole Normale Supérieure; Francia. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentina Fil: Ibarbalz, Federico Matias. Centre National de la Recherche Scientifique. Ecole Normale Supérieure; Francia. Consejo Nacional de Investigaciones Científicas y Técnicas. Oficina de Coordinación Administrativa Ciudad Universitaria. Centro de Investigaciones del Mar y la Atmósfera. Universidad de Buenos Aires. Facultad de Ciencias Exactas y Naturales. Centro de Investigaciones del Mar y la Atmósfera; Argentina Fil: Bowler, Chris. Centre National de la Recherche Scientifique. Ecole Normale Supérieure; Francia. Centre National de la Recherche Scientifique; Francia

International audience; Imagine you see a video of someone pulling back their leg to kick a soccer ball, and then a soccer ball soaring toward a goal. You would likely infer that these scenes are two parts of the same event, and this inference would likely cause you to remember having seen the moment the person kicked the soccer ball, even if that information was never actually presented (Strickland & Keil, 2011, Cognition, 121[3], 409–415). What cues trigger people to "fill in" causal events from incomplete information? Is it due to the experience they have had with soccer balls being kicked toward goals? Is it the visual similarity of the object in both halves of the video? Or is it the mere spatiotemporal continuity of the event? In three experiments, we tested these different potential mechanisms underlying the "filling-in" effect. Experiment 1 showed that filling in occurs equally in familiar and unfamiliar contexts, indicating that familiarity with specific event schemas is unnecessary to trigger false memory. Experiment 2 showed that the visible continuation of a launched object’s trajectory is all that is required to trigger filling in, regardless of other occurrences in the second half of the scene. Finally, Experiment 3 found that, using naturalistic videos, this filling-in effect is more heavily affected if the object’s trajectory is discontinuous in space/time compared with if the object undergoes a noticeable transformation. Together, these findings indicate that the spontaneous formation of causal event representations is driven by object representation systems that prioritize spatiotemporal information over other object features.

International audience; Some sentences are globally informative but still deviant because they contain an expression that is redundant relative in its local environment, as in: #Ann is staying in Paris and she in France. In this case, the second conjunct seems to be evaluated after the first. In several frameworks, order of evaluation follows from linear order. Focusing on nominal modification in Japanese and Korean, Ingason (2016) argued that the correct notion of order is hierarchical, not linear, and he proposed that structurally higher elements are evaluated before lower elements, a conclusion that might dovetail with Romoli and Mandelkern's (2017) proposal for postposed if-clauses in conditionals. While agreeing with Ingason's conclusion that for nominal modification evaluation order is hierarchical, we amend his theory by considering sentences with several pre-or post-nominal modifiers (in English, Mandarin and French). We argue that a Noun Phrase is evaluated 'inside out', starting with the head noun and adding modifiers by order of structural proximity to the head-with the result that higher modifiers are evaluated later than lower modifiers (against the 'higher is earlier' view). We explore in an Appendix how this finding can be integrated with existing accounts of evaluation order for other constituent types, such as conjunction.