Sort by
Refine Your Search
-
language processing that address concrete problems and are both theoretically rigorous and interpretable. The PhD is funded by the ERC CoG PANDORA (Deep Multimodal Learning for Mining and Generation of Arguments
-
. The candidate will be hosted within the Molecular Chemistry and Macromolecular Materials (C3M) department at CNRS-ICMPE, a multidisciplinary research environment dedicated to the design and synthesis of advanced
-
learning, particularly flow-matching generative models and protein language models. The research will focus on designing efficient generative models able to produce realistic conformational ensembles while
-
remaining biologically interpretable? The PhD candidate will design and apply integrative computational workflows using methods such as multi-omics integration, spatial modelling, representation learning
-
to align neuromorphic algorithms with the physical constraints of the target hardware. This hardware–software co design effort will involve: • Deepening and extending NSS-related machine learning and
-
and validated within the PhLAM team. Consequently, only limited effort will be required to acquire the experimental skills needed for this part of the project, allowing the PhD candidate to focus
-
used to generate a training dataset for the cavity circulation emulator. The candidate will test different machine learning architectures and design appropriate, physically relevant metrics to assess
-
. Depending on the candidate's profile and interests, the thesis may develop along one or several of the following directions, at the crossroads of statistical physics, biophysics, and machine learning
-
of heritage mediation. The objective of the proposed PhD project is therefore to design a hybrid scientific architecture combining generative AI, heritage ontologies, reinforcement learning, and probabilistic
-
. The developments will also aim to extend existing approaches, which are often restricted to integrated quantities, to the reconstruction of complete statistical fields. The ultimate goal is to design a goal-oriented