Star formation is a fundamental process that drives the evolution of galaxies through cosmic time. While the last decade has seen major advances in our understanding of star formation on the scale of galaxies and within individual molecular clouds, a key outstanding challenge is to unify Milky Way and extragalactic results. For instance, the DAOISM project (https://www.iram.fr/~pety/ORION-B/) aimed at establishing a synergy between Galactic and Extragalactic studies of the relationships between molecular cloud structure and star formation. 

The wide-field, multi-line position-position-velocity data cubes provided by current and future astronomical observatories encode a tremendous amount of information (kinematics, physics, chemistry, etc.) about the molecular interstellar medium. In order to decode this information reliably, it is necessary to use cutting-edge knowledge in both astrophysics and data analysis. The astronomical challenges for data science are related to the absence of ground truth, the variable signal-to-noise ratio of data, and the control of biases in data and models. To solve these problems, accurate and rapid methods based, for example, on Bayesian inference, machine learning, etc., must be developed or adapted to the specificities of astronomy. 

We invite the community to participate in a five-day workshop (from noon on April 20 to noon on April 24) in Paris. To promote fruitful discussions, we propose to bring together astronomers and data specialists. In astronomy, the focus will be on challenges related to star formation and astrochemistry, from the Milky Way to galaxies. We will divide the time equally between invited presentations and contributions.

Astrophysics

• Filaments and fibers in the Gould Belt molecular clouds, A. Hacar (Vienna University)
• Stratified sampling in the Gould Belt molecular clouds, M. Tafalla (OAN)
• High-angular, wide-field observations of the Milky Way Central Molecular Zone, J. Henshaw (Liverpool John Moores University)
• Star formation at a bar end - The case of W43, F. Motte (IPAG)
• Magnetic field and star formation, A. Maury (ICREA)
• Ionization fraction and magnetic field in Orion B, I. Beslic (LUX)
• Multi-line observations of nearby galaxies, A. Usero (OAN)
• High-angular mutli-line observations in M51, S. Stuber (Japon)
• Sub-beam estimation of column density PDFs in nearby galaxies: A. Zarkardjian (IRAP)
• Comparative lifetime of atomic and molecular clouds in nearby galaxies, L. Ramambason (Heidelberg universtiy)
• Structure of nearby clouds in Ammonia, J. Pineda (MPE)
• Dense core chemistry, H. Mazurek (LUX)
• Revisiting the role of the HCN(1-0) line as a tracer of the dense gas reservoir for star formation, M. Santa-Maria (CSIC)
• Automated line attribution in line surveys, T. Csengeri (LAB)
• Hyperspectral detection of galaxies at different redshift, D. Cornu (LUX)

Data science

• Bayesian inference & regularization, T. Oberlin (ISAE-Sup Aero, ANIT)
• Bayesian inference & inverse problems, F. Forbes (INRIA)
• BEETROOTS, a new bayesian framework for inverse problems, P. Palud (THALES)
• Emulation of sophisticated model: The case of the Meudon PDR code, E. Bron (LUX)
• Application of scattering transforms to the separation of ISM phases, E. Allys, (LPENS)
• From Statistical Mechanics to Machine Learning and back, M. Gabrié (LPENS)
• Remote sensing & satellite imaging, F. Turpin (Telecom Paris) ou L. Denis (Université Jean Monnet) 
• Neural network denoising and non-linear correlation analysis with mutual information, L. Einig (Netatmo)
• Exoplanet (statistical inference and neural network), O. Flasseur (CRAL)
• Maps of kinetic temperature, volume density and kinematic parameters from non-LTE multi-species fit, L. Segal (THALES)