• Digital Humanities and Cultural Heritage
• Publications close
• Research data close
• Other research products close
• EU close
• OpenAIRE close

Recently, domain-general recurrent neural networks, without explicit linguistic inductive biases, have been shown to successfully reproduce a range of human language behaviours, such as accurately predicting number agreement between nouns and verbs. We show that such networks will also learn number agreement within unnatural sentence structures, i.e. structures that are not found within any natural languages and which humans struggle to process. These results suggest that the models are learning from their input in a manner that is substantially different from human language acquisition, and we undertake an analysis of how the learned knowledge is stored in the weights of the network. We find that while the model has an effective understanding of singular versus plural for individual sentences, there is a lack of a unified concept of number agreement connecting these processes across the full range of inputs. Moreover, the weights handling natural and unnatural structures overlap substantially, in a way that underlines the non-human-like nature of the knowledge learned by the network.

AbstractDuring its first days of existence, the Vichy regime ordered a review of recent naturalizations. In accordance with a law passed on July 22, 1940, denaturalization decisions were made on a ...

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

­The Eastern Fertile Crescent region of western Iran and eastern Iraq hosted major developments in the transition from hunting and gathering to more sedentary agricultural lifestyles through the Early Neolithic period, 10,000-7000 BC. Within the scope of the Central Zagros Archaeological Project, excavations have been conducted at two Early Neolithic sites in the Kurdistan region of Iraq: Bestansur and Shimshara, as well as survey in the region of the Epipalaeolithic site of Zarzi since 2012. Bestansur represents an early stage in the transition to sedentary, agricultural life, where the inhabitants pursued a biodiverse strategy of hunting, gathering, herding and cultivating, maximising the new opportunities afforded by the warmer climate of the Early Holocene. They also constructed a substantial settlement of mudbrick, including a major building with a minimum of 78 human individuals buried under its floor in association with hundreds of beads. ­ These buildings and human remains provide new insights into social relations, mortuary practices, demography, diet, health and disease during the early stages of sedentarisation. ­ The material culture of Bestansur and Shimshara is rich in imported items such as obsidian, carnelian and sea-shells, indicating the extent to which Early Neolithic communities were networked across the Eastern Fertile Crescent and beyond along routes that later became the Silk Roads. This volume includes final reports by a large-scale interdisciplinary team on a wealth of new data from excavations at Bestansur and Shimshara, through application of state-of-the-art scientific techniques, integrated ecological and social approaches and sustainability studies. ­ The net result is to re-emphasise the enormous significance of the Eastern Fertile Crescent in one of the most important episodes in human history: the Neolithic transition.

This article examines a largely unexplored component of China’s classical garden system – the gardens of salt merchants in Tianjin during the Qing Dynasty (1636-1911). Beyond existing works, which tend to focus on imperial and scholar gardens – gardens of the ruling elites – this examination of merchant gardens contributes to garden history by revealing that merchants created gardens to improve their low social status. It further reveals shifts in the functions, architectural design and flora of the gardens which reflects both individual aesthetics and the changing fortunes of Tianjin’s salt merchants in general. Salt merchant gardens in Tianjin initially presented idyllic scenery to create literary-based, self-immersed spaces. Then beginning in the 1720s, they evolved into a showcase of rising merchant power displaying affluence, thereby enabling merchants to improve their social rank. Finally, from the 1840s, salt merchant gardens gradually became extravagant enclosures as the collapse of the established social structure unfolded.

The automatic extraction of a patient’s natural history from Electronic Health Records (EHRs) is a critical step towards building intelligent systems that can reason about clinical variables and support decision making. Although EHRs contain a large amount of valuable information about the patient’s medical care, this information can only be fully understood when analyzed in a temporal context. Any intelligent system should then be able to extract medical concepts, date expressions, temporal relations and the temporal ordering of medical events from the free texts of EHRs; yet, this task is hard to tackle, due to the domain specific nature of EHRs, writing quality and lack of structure of these texts, and more generally the presence of redundant information. In this paper, we introduce a new Natural Language Processing (NLP) framework, capable of extracting the aforementioned elements from EHRs written in Spanish using rule-based methods. We focus on building medical timelines, which include disease diagnosis and its progression over time. By using a large dataset of EHRs comprising information about patients suffering from lung cancer, we show that our framework has an adequate level of performance by correctly building the timeline for 843 patients from a pool of 989 patients, achieving a correct result in 85% of instances.

The last glacial period is characterized by a number of millennial climate events that have been identified in both Greenland and Antarctic ice cores and that are abrupt in Greenland climate records. The mechanisms governing this climate variability remain a puzzle that requires a precise synchronization of ice cores from the two hemispheres to be resolved. Previously, Greenland and Antarctic ice cores have been synchronized primarily via their common records of gas concentrations or isotopes from the trapped air and via cosmogenic isotopes measured on the ice. In this work, we apply ice core volcanic proxies and annual layer counting to identify large volcanic eruptions that have left a signature in both Greenland and Antarctica. Generally, no tephra is associated with those eruptions in the ice cores, so the source of the eruptions cannot be identified. Instead, we identify and match sequences of volcanic eruptions with bipolar distribution of sulfate, i.e. unique patterns of volcanic events separated by the same number of years at the two poles. Using this approach, we pinpoint 82 large bipolar volcanic eruptions throughout the second half of the last glacial period (12–60 ka). This improved ice core synchronization is applied to determine the bipolar phasing of abrupt climate change events at decadal-scale precision. In response to Greenland abrupt climatic transitions, we find a response in the Antarctic water isotope signals (δ18O and deuterium excess) that is both more immediate and more abrupt than that found with previous gas-based interpolar synchronizations, providing additional support for our volcanic framework. On average, the Antarctic bipolar seesaw climate response lags the midpoint of Greenland abrupt δ18O transitions by 122±24 years. The time difference between Antarctic signals in deuterium excess and δ18O, which likewise informs the time needed to propagate the signal as described by the theory of the bipolar seesaw but is less sensitive to synchronization errors, suggests an Antarctic δ18O lag behind Greenland of 152±37 years. These estimates are shorter than the 200 years suggested by earlier gas-based synchronizations. As before, we find variations in the timing and duration between the response at different sites and for different events suggesting an interaction of oceanic and atmospheric teleconnection patterns as well as internal climate variability. International audience