• 2018-2022close
• Part of book or chapter of book close
• CH close
• Hyper Article en Ligne close
• Digital Humanities and Cultural Heritage close

International audience; This paper describes the workflow of the Grammateus project, from gathering data on Greek documentary papyri to the creation of a web application. The first stage is the selection of a corpus and the choice of metadata to record: papyrology specialists gather data from printed editions, existing online resources and digital facsimiles. In the next step, this data is transformed into the EpiDoc standard of XML TEI encoding, to facilitate its reuse by others, and processed for HTML display. We also reuse existing text transcriptions available on . Since these transcriptions may be regularly updated by the scholarly community, we aim to access them dynamically. Although the transcriptions follow the EpiDoc guidelines, the wide diversity of the papyri as well as small inconsistencies in encoding make data reuse challenging. Currently, our data is available on an institutional GitLab repository, and we will archive our final dataset according to the FAIR principles.

International audience; Convolutional neural networks (CNN) are known to learn an image representation that captures concepts relevant to the task, but do so in an implicit way that hampers model interpretability. However, one could argue that such a representation is hidden in the neurons and can be made explicit by teaching the model to recognize semantically interpretable attributes that are present in the scene. We call such an intermediate layer a \emph{semantic bottleneck}. Once the attributes are learned, they can be re-combined to reach the final decision and provide both an accurate prediction and an explicit reasoning behind the CNN decision. In this paper, we look into semantic bottlenecks that capture context: we want attributes to be in groups of a few meaningful elements and participate jointly to the final decision. We use a two-layer semantic bottleneck that gathers attributes into interpretable, sparse groups, allowing them contribute differently to the final output depending on the context. We test our contextual semantic interpretable bottleneck (CSIB) on the task of landscape scenicness estimation and train the semantic interpretable bottleneck using an auxiliary database (SUN Attributes). Our model yields in predictions as accurate as a non-interpretable baseline when applied to a real-world test set of Flickr images, all while providing clear and interpretable explanations for each prediction.

The deposits of volcanic debris avalanches (VDAs) contain diagnostic features that distinguish them from those of other landslides. In this chapter, we summarize the sedimentary characteristics and the different (litho-)facies described over the past four decades, and how findings from individual case studies can be adapted as globally applicable sedimentological tools. A plethora of descriptive terms and partially conflicting definitions emerged in the ever-growing literature on VDA deposits (VDADs). These we summarize and make recommendations for future use. Different facies models that were developed at different volcanoes might point to unique emplacement conditions (e.g. dry versus wet; confined versus unconfined) and, if confirmed, the apparent ‘conflict’ of terminology might help identify the paleo-settings of ancient VDAs. General observations of large unsaturated landslides of different origin show that preservation of source stratigraphy, (mega-)clasts, jigsaw-fractured clasts, and incorporation of runout path material are common features. Their unique composition, grain sizes, and abundance of matrix sets VDADs apart from deposits of large rockslides and debris flows. The latter can be associated with VDAs, and whether they formed syn- or post-VDAD emplacement is reflected in forensic evidence within the depositional sequences. Recent case studies illustrate the advances in analytical techniques and in understanding the processes of debris avalanche transport and deposition forty years after the eruption and lateral collapse of Mount St. Helens volcano.

International audience; This chapter provides methodological information about implementing documentary approaches by which to analyze, discuss, and support river restoration. First, we present the kind of documentary material that can be used to analyze discourse on river restoration. We explain how to build a corpus in order to study stakeholders’ discourses and to collect specific information about river restoration. We based our analysis on scientific publications, newspaper articles, and administrative documents. Then we detail different qualitative and quantitative methods used to analyze such documentary material and the type of results they yield. We present four types of method: bibliometric analysis, content analysis, textual data analysis, and qualitative analysis. Finally, we discuss the pros and cons of documentary methods versus survey methods.