• Digital Humanities and Cultural Heritage
• 2017-2021close
• Open Access close
• Publications close
• Research data close
• Other research products close
• FR close
• Oxford University Research Archive close

Recent work on Model Informed Drug Discovery and Development (MID3) has noted the need for clarity in model description used in quantitative disciplines such as pharmacology and statistics. 1-3 Currently, models are encoded in a variety of computer languages and are shared through publications that rarely include original code and generally lack reproducibility. The DDMoRe Model Description Language (MDL) has been developed primarily as a language standard to facilitate sharing knowledge and understanding of models.

Nearly all existing techniques for automated video annotation (or captioning) describe videos using natural language sentences. However, this has several shortcomings: (i) it is very hard to then further use the generated natural language annotations in automated data processing, (ii) generating natural language annotations requires to solve the hard subtask of generating semantically precise and syntactically correct natural language sentences, which is actually unrelated to the task of video annotation, (iii) it is difficult to quantitatively measure performance, as standard metrics (e.g., accuracy and F1-score) are inapplicable, and (iv) annotations are language-specific. In this paper, we propose the new task of knowledge graph extraction from videos, i.e., producing a description in the form of a knowledge graph of the contents of a given video. Since no datasets exist for this task, we also include a method to automatically generate them, starting from datasets where videos are annotated with natural language. We then describe an initial deep-learning model for knowledge graph extraction from videos, and report results on MSVD* and MSR-VTT*, two datasets obtained from MSVD and MSR-VTT using our method. 10 pages, 4 figures

Based on the cross-referencing of ethnographic materials collected in Fiji and Vanuatu, this article explores the diverse ways faith and climate change are connected together and how these connections are sustained and transformed over time. It does so through the prism of three ‘practices of assemblage’ identified by Tania Murray Li, namely forging alignments, authorising knowledge, and reassembling. It emphasises the partnership and combined efficacy of faith-based organisations and the Bible in these practices, while revealing the role of various other actants including God, NGOs, youth activists, cyclones, a Ni-Vanuatu canoe, and Fiji’s Presidency of COP23. This approach highlights the coexistence, in both Fiji and Vanuatu, of a religiously informed and adaptation-oriented environmental stewardship narrative, stressing human responsibility in the face of climate change, with a counter-narrative considering climate change as God’s business. This coexistence sometimes creates tensions between worldly and religious responses to climate change. These different religious perspectives of climate change can also be deployed as a political resource in nation-building processes, regional power relations, and international climate negotiations. In Oceania indeed, climate change appears as a new arena in which the inextricable entanglement of Christianity and politics is revealed.

Over the last 40 years or so, there has been an explosion of cultural evolution research in anthropology and archaeology. In each discipline, cultural evolutionists investigate how interactions between individuals translate into group level patterns, with the aim of explaining the diachronic dynamics and diversity of cultural traits. However, while much attention has been given to deterministic processes (e.g. cultural transmission biases), we contend that current evolutionary accounts of cultural change are limited because they do not adopt a systematic stochastic approach (i.e. accounting for the role of chance). First, we show that, in contrast with the intense debates in ecology and population genetics, the importance of stochasticity in evolutionary processes has generated little discussion in the sciences of cultural evolution to date. Second, we speculate on the reasons, both ideological and methodological, why that should be so. Third, we highlight the inadequacy of genetically-inspired stochastic models in the context of cultural evolution modelling, and ask which fundamental stochastic processes might be more relevant to take up. We conclude that the field of cultural evolution would benefit from a stochastic revolution. For that to occur, stochastic models ought to be developed specifically for cultural data and not through a copy-pasting of neutral models from population genetics or ecology.

An x-ray radiography technique based upon phase contrast imaging using a lithium fluoride detector has been demonstrated for goals of high energy density physics experiments. Based on the simulation of propagation an x-ray free-electron laser beam through a testobject, the visibility of phase-contrast image depending on an object-detector distance was investigated. Additionally, the metrological capabilities of a lithium fluoride crystal as a detector were demonstrated. International audience

Several questions remain regarding the timing and nature of the Neanderthal-anatomically modern human (AMH) transition in Europe. The situation in Eastern Europe is generally less clear due to the relatively few sites and a dearth of reliable radiocarbon dates. Claims have been made for both notably early AMH and notably late Neanderthal presence, as well as for early AMH (Aurignacian) dispersal into the region from Central/Western Europe. The Kostenki-Borshchevo complex (European Russia) of Early Upper Paleolithic (EUP) sites offers high-quality data to address these questions. Here we revise the chronology and cultural status of the key sites of Kostenki 17 and Kostenki 14. The Kostenki 17/II lithic assemblage shares important features with Proto-Aurignacian material, strengthening an association with AMHs. New radiocarbon dates for Kostenki 17/II of ~41–40 ka cal BP agree with new dates for the recently excavated Kostenki 14/IVw, which shows some similarities to Kostenki 17/II. Dates of ≥41 ka cal BP from other Kostenki sites cannot be linked to diagnostic archaeological material, and therefore cannot be argued to date AMH occupation. Kostenki 14’s Layer in Volcanic Ash assemblage, on the other hand, compares to Early Aurignacian material. New radiocarbon dates targeting diagnostic lithics date to 39–37 ka cal BP. Overall, Kostenki’s early EUP is in good agreement with the archaeological record further west. Our results are therefore consistent with models predicting interregional penecontemporaneity of diagnostic EUP assemblages. Most importantly, our work highlights ongoing challenges for reliably radiocarbon dating the period. Dates for Kostenki 14 agreed with the samples’ chronostratigraphic positions, but standard pre-treatment methods consistently produced incorrect ages for Kostenki 17/II. Extraction of hydroxyproline from bone collagen using prep-HPLC, however, yielded results consistent with the samples’ chronostratigraphic position and with the layer’s archaeological contents. This suggests that for some sites compound-specific techniques are required to build reliable radiocarbon chronologies. International audience

This paper presents new archaeological material and first dates on Upper Pleistocene layers at the site of La Crouzade cave (Gruissan, Aude, France). The site was first excavated by T. and P. Héléna at the beginning of the twentieth century, and the excavations were recently completed during three years (2016–2018) of systematic campaigns. We obtained dates from Middle Palaeolithic layers using two methods: AMS 14C dates were obtained from bone and charcoal, and combined ESR-U series dating was undertaken on horse teeth. Together, these methods allowed us to date this Mousterian sequence to 49,776–44805 cal BP for the deepest level (layer C8) and from 42,000 ± 3000 years BP for the top (layer C6). The Upper Palaeolithic layers are preserved only as patches in the actual excavation area, but a date was obtained from a piece of charcoal collected from a small hearth preserved in the first layer (C5) above Middle Palaeolithic deposits, which indicates an age similar to that of a modern human maxillary previously analysed and re-dated here from 36,014 to 34402 cal BP, confirming its stratigraphic attribution. The Middle Palaeolithic lithics at the site were first described as para-Charentian cultural facies following typological analyses. The revision of the earlier collection supplemented with the new material, using a technological approach, allow to identify two layers dominated by Levallois production followed by discoid production (Layers C8 and C6) surrounding an original assemblage (layer C7), characterised by a dominant Levallois production completed by three secondary production systems of equal importance, including discoid, SSDA and a Quina-like production. The faunal spectrum predominantly comprises an assemblage of Pleistocene large mammals, and biochronological studies corroborate the dates obtained.