Missions

In the context of the PRAIRIE cluster the History of astronomy team of the LTE laboratory in Paris Observatory is offering a two year postdoctoral position targeted on interdisciplinary research branching Artificial Intelligence and Digital Humanities. More specifically 80% of the Postdoc time will be devoted to research in the context of the EIDA project, and 20% of the Postdoc time will be decided to foster more generally on the national and international levels initiative which actively shape and expand key research communities at the intersection of AI and the humanities.
EIDA (ANR-22-CE38-0014) aims at developing a new critique of astronomical diagrams of the 8th-18th centuries found, among others, in Chinese, Sanskrit, Arabic/Persian, Greek, Hebrew and Latin sources. It is an interdisciplinary project joining a team of historians of sciences with computer vision researchers who will develop specific algorithms for retrieval and analysis of the historical diagrams. EIDA focuses on diagrams as both visual heritage and tools of astronomical or mathematical reasoning in the astral sciences. Their large-scale study requires the automatic fine-grained analysis of their content, which is often uncorrelated with their visual aspect and for which no satisfactory AI solution exists. The ambition of the project is thus to develop and implement a radically new approach in the study of astronomical diagrams in premodern and early modern periods by creating and leveraging novel and specific computer vision tools. The two main objectives of EIDA are:
to shape a specific criticism of astronomical diagrams and standards for their natively critical edition by addressing the variety of their functions and circulation modes;
to take this innovative analysis to an unprecedented Afro-eurasian scale by developing new computer vision approaches, capable of decomposing an astronomical diagram into components that are meaningful for analysis and edition, without relying on dataset-specific human annotation.
EIDA associates the recent progress of 'analysis-by-synthesis' approaches in computer vision and the recent 'visual' and 'practical' turns in history of astronomy in a deeply interdisciplinary perspective. This collaboration is timely in light of data availability and will allow us to conduct refined analysis on an unprecedented scale disclosing otherwise inaccessible historical results. Moreover, in line with open science principles, EIDA shall develop open source tools for harvesting, studying, editing and visualizing diagrams through a web interface, modeled after the existing DISHAS platform for astronomical tables. The new platform will enable the research community to access the tools employed to produce the results published by EIDA and public code release will enable further building on our developments.
Aligned with EIDA's goals, the successful candidate's research will contribute to a deeper understanding of the documentary, epistemological, and broader intellectual and cultural significance of astronomical diagrams. This research should focus on a well-defined corpus of manuscripts and/or printed sources chosen by the candidate. Sources within this corpus can be drawn from, but are not limited to, one or more of the project's core domains: Chinese, Sanskrit, Byzantine, Arabic/Persian, Hebrew and Latin sources. We also encourage proposals that explore sources from other languages, such as those found in Central Asia, Caucasus or Africa.

Activités

Building upon their personal research project, the successful candidate will actively engage and promote the collective efforts of the project in leveraging new AI based tools for the history of astronomy with the project team that includes historians of astronomy, computer vision specialists, and Digital Humanities engineers both at the Observatoire and other institutions. An EIDA core research area investigates the multifaceted nature of defining image similarity. This exploration extends beyond computer vision-generated representations, delving into broader methods for establishing meaningful and dynamic image groupings. These groupings may be influenced by the documentary context, astronomical content, or the visual styles in which the images are created. A second core EIDA research area concerns the possible features of natively digital critical edition of astronomical diagrams. In practice, the successful candidate will actively contribute to the team's exploration of these questions. This may involve leading dedicated seminars or spearheading specific collaborative tasks within the EIDA framework.
The successful candidate will work as part of the project team in Paris Observatory and will spend 80% of their postdoc time working on their research project in the context of this collective, international project. The postdoctoral fellow is expected to participate in the joint EIDA publications and will be encouraged to take part in the organization of scientific events relevant to their research (workshops and seminars). The postdoctoral fellow will also have dedicated research funds especially for travel to relevant libraries or conferences.
In addition to the research activities outlined above, approximately 20% of the postdoc's time will be dedicated to research management, overseeing two key initiatives:
Co-leading the DHAI (Digital Humanities / Artificial Intelligence) Research Initiative DHAI Seminar which imply:
Organizing a monthly seminar focused on the intersection of AI and the humanities.
Co-organizing two intensive weeks of graduate-level teaching on AI and the humanities.
Encouraging and supporting initiatives related to DHAI in the Paris area.
Coordinating the ELLIS Paris Unit ELLIS Paris
Attending the monthly coordinator meetings of ELLIS to stay aligned with the broader ELLIS network.
Helping to coordinate activities within the Paris Unit, including encouraging and facilitating ELLIS-related events in the region.
Exploring the creation of a special track on DHAI within ELLIS, a development that would enhance the unit's engagement with this interdisciplinary area.
These responsibilities offer the opportunity to actively shape and expand key research communities at the intersection of AI and the humanities.

Compétences

Applicants must hold a PhD in history of science, history, or art history. A strong background in handling primary sources, particularly those related to history of astronomy, mathematics, or art history, is highly valued. Additionally, a demonstrated interest and willingness to work with and contribute to the development of digital humanities, computer vision and AI tools is essential. Excellent written and spoken English communication skills are required for successful collaboration with the project's international team.

Contexte de travail

The work will take place in the history of astronomy team of the Laboratoire Temps Espace in Paris Observatory. more information can be found there: https://syrte.obspm.fr/spip/science/histoire/

Le poste se situe dans un secteur relevant de la protection du potentiel scientifique et technique (PPST), et nécessite donc, conformément à la réglementation, que votre arrivée soit autorisée par l'autorité compétente du MESR.

Contraintes et risques

Nothing specific