Digital Humanities transverse axis (DH)

Through belonging to organisms mostly related with the Humanities and Social sciences (Université Lyon 2, Maison des Sciences de l’Homme Lyon-Saint-Étienne, Institut du Genre) and multidisciplinary research (LabEx Intelligence des Mondes Urbains, Institut rhônalpin des systèmes complexes), by strongly participating in the  MA in Digital Humanities  that we coordinate and that is coaccreditated by Universities  Lyon 2  and  Lyon 3, the École Normale Supérieure de Lyon  and the École nationale supérieure des sciences de l’information et des bibliothèques, and by participating to many multidisciplinary projects with various laboratories of Humanities and Social sciences, ERIC became a well-known actor in DH on the Lyon-Saint-Étienne site, as well as nationally.

The DH transversal axis helps structure and make visible common projects performed by ERIC's research teams  DMD  and  DIS  in terms of DH. Scientifically, our goal is not only to find application fields to our own research, but mainly to hybridize methodologies from Computer science, Statistics, Humanities and Social sciences to achieve original approaches. We also involve into the long-term collaborations that are required for multidisciplinary research to succeed.

Scientific topics

  • Data lakes.  The current trend involves a whole range of Big Data variety (structured, semi-structured and unstructured data, including textual documents that are premium in Humanities and Social sciences), which must be managed and queried altogether. This theme is particularly addressed in our data lake research and structures our reflection. Moreover, the accessibility of such types of data organization, querying and analysis to non-specialists of computer or data science, such as our Humanities and Social sciences partners, raises fundamental issues.
  • Artificial Intelligence explainability.  As soon as non-computer scientist users can use and interact with machine learning model and algorithm results, providing the result of an algorithm such as a  clustering, for instance, is not enough. A concise description of the  produced clusters  and, if possible, getting back to the original data, is much needed for users to understand why some objects, e.g., texts, have been placed into a given category.

Some recent projects