Periodic Reporting for period 2 - AI4theSciences (Artificial Intelligence for the Sciences)
Okres sprawozdawczy: 2022-10-01 do 2025-09-30
The Artificial Intelligence for the Sciences (AI4theSciences) cofund Project is an innovative, interdisciplinary and intersectoral PhD program conducted by Université Paris Sciences et Lettres (PSL). AI4theSciences Cofund project participates in the development of an interdisciplinary research community composed of all PSL laboratories at the cutting edge of the use of artificial intelligence techniques in their own disciplines. As a founding member of 3IA PR[A]IRIE and the Paris School of AI within the framework of the IA-Cluster programme, PSL is recognised at the highest international level for its work in mathematics and computer science, at the core of AI and Machine Learning.
AI4theSciences is a flagship project for PSL University. It is an emblematic example to all the PSL scientific community that EU projects are accelerators of career development, advance and transfer knowledge.
The AI4theSciences programme train a new generation of scientists highly sought-after both by private sector and academia, at the interface of AI and other academic fields, and aware of ethical and social issues related to AI. A global strategy for developing AI and its applications has been devised by the French President, planning in particular the creation of 4 leading AI Institutes in France. The PSL Graduate Transverse Program on AI, to which the AI4theSciences programme contribute, is part of a major component of one of the four selected Institutes: PR[AI]RIE 3IA Research Center. The AI4theSciences Cofund reinforces this Graduate Transverse Program and therefore contributes to this national plan on AI and its applications. More specifically, the PSL Graduate Transverse and the AI4theSciences programmes aim exactly at fostering pluridisciplinary researches and training of specialists of various academic fields equipped with the latest AI technologies adapted to their own field, as required both by the private sector and our university.
AI4theSciences is a flagship project for PSL University. It is an emblematic example to all the PSL scientific community that EU projects are accelerators of career development, advance and transfer knowledge.
The AI4theSciences programme train a new generation of scientists highly sought-after both by private sector and academia, at the interface of AI and other academic fields, and aware of ethical and social issues related to AI. A global strategy for developing AI and its applications has been devised by the French President, planning in particular the creation of 4 leading AI Institutes in France. The PSL Graduate Transverse Program on AI, to which the AI4theSciences programme contribute, is part of a major component of one of the four selected Institutes: PR[AI]RIE 3IA Research Center. The AI4theSciences Cofund reinforces this Graduate Transverse Program and therefore contributes to this national plan on AI and its applications. More specifically, the PSL Graduate Transverse and the AI4theSciences programmes aim exactly at fostering pluridisciplinary researches and training of specialists of various academic fields equipped with the latest AI technologies adapted to their own field, as required both by the private sector and our university.
All PhD projects funded under the AI4TheSciences COFUND programme share a common methodological foundation in the development and application of advanced artificial intelligence and machine learning techniques, while intentionally addressing a very broad range of scientific questions. The programme spans domains including physics, cosmology and astrophysics, quantum technologies, advanced materials and energy systems, neuroscience and cognition, medicine and biomedical engineering, biology and ecology, language and cultural studies, as well as economics and social sciences. In the physical sciences, AI is used to model complex systems, integrate observational data with physical constraints, quantify uncertainties, and reveal hidden structures in high-dimensional datasets, contributing to progress in areas such as dark energy studies, space astrometry and materials failure analysis. In engineering-oriented projects, AI supports the design and optimisation of quantum-inspired computing architectures, advanced materials, and energy-related technologies, highlighting the convergence between computation and physical matter.
In the life sciences and medicine, AI-driven approaches are developed for neural signal decoding, brain–computer interfaces, medical image reconstruction, biomechanical modelling, and personalised treatment planning, combining deep learning, physics-informed methods and transfer learning. In parallel, biological and ecological projects employ AI to investigate evolutionary processes, enzyme design, genomic data, to decipher the communication of dolphins, and to monitor biodiversity through environmental DNA, demonstrating the role of AI as a key instrument for modern biological discovery. Language and cognition-oriented projects explore the emergence of linguistic competence in humans and machines, the dynamics of large language models, and the cultural and social dimensions of artificial intelligence. Finally, projects in economics and social sciences apply AI to large-scale behavioural and organisational data to analyse decision-making processes, career dynamics and financial market formation.
Taken together, this portfolio reflects a deliberate strategic choice to position artificial intelligence as a unifying scientific methodology rather than as a standalone discipline. By fostering close interaction between methodological innovation and domain-specific applications, the AI4TheSciences PhD programme contributes to shaping a new paradigm of interdisciplinary, data-driven scientific research in Europe.
In the life sciences and medicine, AI-driven approaches are developed for neural signal decoding, brain–computer interfaces, medical image reconstruction, biomechanical modelling, and personalised treatment planning, combining deep learning, physics-informed methods and transfer learning. In parallel, biological and ecological projects employ AI to investigate evolutionary processes, enzyme design, genomic data, to decipher the communication of dolphins, and to monitor biodiversity through environmental DNA, demonstrating the role of AI as a key instrument for modern biological discovery. Language and cognition-oriented projects explore the emergence of linguistic competence in humans and machines, the dynamics of large language models, and the cultural and social dimensions of artificial intelligence. Finally, projects in economics and social sciences apply AI to large-scale behavioural and organisational data to analyse decision-making processes, career dynamics and financial market formation.
Taken together, this portfolio reflects a deliberate strategic choice to position artificial intelligence as a unifying scientific methodology rather than as a standalone discipline. By fostering close interaction between methodological innovation and domain-specific applications, the AI4TheSciences PhD programme contributes to shaping a new paradigm of interdisciplinary, data-driven scientific research in Europe.
The AI4theSciences COFUND programme has significantly advanced interdisciplinary doctoral training by integrating Artificial Intelligence within diverse scientific fields. Its dual supervision model, which pairs an AI specialist with a domain expert, has enabled PhD fellows to achieve substantial scientific progress. The successful defence of numerous theses from the first cohort demonstrates this impact.
By the end of the project, several results are expected. First, all doctoral projects from cohorts 2 and 3 will be completed, generating scientific advances in areas such as deep generative models, brain and language modelling, AI for seismic hazard monitoring and biomechanical simulation for surgery. Second, the programme will contribute to strengthening the European AI talent pool, with 24 highly trained researchers prepared to pursue careers across academia, industry and the public sector. Finally, further integration within the expanding AI ecosystem of PSL University is anticipated, particularly through the Paris School of AI, known as PRAIRIE PSAI, supported by the France 2030 initiative. This development will increase visibility, training capacity and research excellence.
AI4theSciences also contributes to strengthening Europe’s leadership in AI driven scientific innovation. The research conducted by the fellows advances machine learning methodologies and applies them to strategic domains. Open access dissemination ensures broad and lasting scientific impact. In addition, the programme directly addresses Europe’s shortage of AI skilled professionals. Its graduates will contribute to key sectors including healthcare, risk management, digital technologies and environmental monitoring. In doing so, the programme supports the emergence of a highly qualified and ethically aware workforce capable of driving responsible and sustainable AI innovation across the European Research Area.
By the end of the project, several results are expected. First, all doctoral projects from cohorts 2 and 3 will be completed, generating scientific advances in areas such as deep generative models, brain and language modelling, AI for seismic hazard monitoring and biomechanical simulation for surgery. Second, the programme will contribute to strengthening the European AI talent pool, with 24 highly trained researchers prepared to pursue careers across academia, industry and the public sector. Finally, further integration within the expanding AI ecosystem of PSL University is anticipated, particularly through the Paris School of AI, known as PRAIRIE PSAI, supported by the France 2030 initiative. This development will increase visibility, training capacity and research excellence.
AI4theSciences also contributes to strengthening Europe’s leadership in AI driven scientific innovation. The research conducted by the fellows advances machine learning methodologies and applies them to strategic domains. Open access dissemination ensures broad and lasting scientific impact. In addition, the programme directly addresses Europe’s shortage of AI skilled professionals. Its graduates will contribute to key sectors including healthcare, risk management, digital technologies and environmental monitoring. In doing so, the programme supports the emergence of a highly qualified and ethically aware workforce capable of driving responsible and sustainable AI innovation across the European Research Area.