Internalisation of external knowledge by foundation models
To perform an unknown task, a subject (human or robot) has to consult external information, which involves a cognitive cost. After several similar experiments, it masters the situation and can act automatically. The 1980s and 1990s saw explorations in AI using conceptual graphs and schemas, but their large-scale implementation was limited by the technology available at the time.
Today's neural models, including transformers and LLM/VLMs, learn universal representations through pre-training on huge amounts of data. They can be used with prompts to provide local context. Fine-tuning allows these models to be specialised for specific tasks.
RAG and GraphRAG methods can be used to exploit external knowledge, but their use for inference is resource-intensive. This thesis proposes a cognitivist approach in which the system undergoes continuous learning. It consults external sources during inference and uses this information to refine itself regularly, as it does during sleep. This method aims to improve performance and reduce resource consumption.
In humans, these processes are linked to the spatial organisation of the brain. The thesis will also study network architectures inspired by this organisation, with dedicated but interconnected “zones”, such as the vision-language and language models.
These concepts can be applied to the Astir and Ridder projects, which aim to exploit foundation models for software engineering in robotics and the development of generative AI methods for the safe control of robots.
Custom synthesis of diamond nanoparticles for photocatalytic hydrogen production
Diamond nanoparticles (nanodiamonds) are used in nanomedicine, quantum technologies, lubricants and advanced composites [1-2]. Our recent results show that nanodiamond can also act as a photocatalyst, enabling the production of hydrogen under solar illumination [3]. Despite its wide band gap, its band structure is adaptable according to its nature and surface chemistry [4]. Moreover, the controlled incorporation of dopants or sp2 carbon leads to the generation of additional bandgap states that enhance the absorption of visible light, as shown in a recent study involving our group [5]. The photocatalytic performance of nanodiamonds is therefore highly dependent on their size, shape and concentration of chemical impurities. It is therefore essential to develop a "tailor-made" nanodiamond synthesis method, in which these different parameters can be finely controlled, in order to provide a supply of "controlled" nanodiamonds, which is currently lacking.
This PhD aims to develop a bottom-up approach to grow nanodiamond using a sacrificial template (silica beads) to which diamond seeds < 10 nm are attached by electrostatic interaction. The growth of diamond nanoparticles from these seeds will be achieved by microwave-enhanced chemical vapor deposition (MPCVD) using a homemade rotating reactor available at CEA NIMBE. After growth, the CVD-NDs will be collected after dissolution of the sacrificial pattern. Preliminary experiments have demonstrated the feasibility of this approach with the synthesis of faceted <100 nm nanodiamonds (so called CVD-ND), as shown in the scanning electron microscopy image.
During the PhD work, the nature of the diamond seeds (ultra-small NDs [size ˜ 5 nm] synthesized by detonation or HPHT, or molecular derivatives of adamantane) as well as CVD growth parameters will be studied to achieve better controlled CVD-NDs in terms of crystallinity and morphology. Nanodiamonds doped with boron or nitrogen will be also considered, playing on the gas phase composition. The crystalline structure, morphology and surface chemistry will be studied at CEA NIMBE using SEM, X-ray diffraction and Raman, infrared and photoelectron spectroscopies. A detailed analysis of the crystallographic structure and structural defects will be carried out by high-resolution transmission electron microscopy (collaboration). CVD FNDs will then be exposed to gas-phase treatments (air, hydrogen) to modulate their surface chemistry and stabilize them in water. The photocatalytic performance for hydrogen production under visible light of these different CVD-NDs will be evaluated and compared using the photocatalytic reactor recently installed at CEA NIMBE.
References
[1] Nunn et al., Current Opinion in Solid State and Materials Science, 21 (2017) 1.
[2] Wu et al., Angew. Chem. Int. Ed. 55 (2016) 6586.
[3] Marchal et al., Adv. Energy Sustainability Res., 2300260 (2023) 1-8.
[4] Miliaieva et al., Nanoscale Adv. 5 (2023) 4402.
[5] Buchner et al., Nanoscale 14 (2022) 17188.
Modeling and prediction of electromagnetic emissions from power converters using deep learning
In recent years, electromagnetic compatibility (EMC) in power converters based on wide bandgap (WBG) semiconductors has attracted growing interest, due to the high switching speeds and increased frequencies they enable. While these devices improve power density and system efficiency, they also generate more complex conducted and radiated emissions that are challenging to control. In this context, this thesis focuses on the prediction, modeling, and characterization of electromagnetic interference (EMI) (> 30 MHz), both conducted and radiated, in high-frequency power electronic systems. The work is based on a multi-subsystem partitioning method and an iterative co-simulation approach, combined with in situ characterization to capture non-ideal and nonlinear phenomena. In addition, deep learning techniques are employed to model EMI behavior using both measured and simulated data. Generative artificial intelligence (Generative AI) is also leveraged to automatically generate representative and diverse configurations commonly encountered in power electronics, thereby enabling efficient exploration of a wide range of EMI scenarios. This hybrid approach aims to enhance analysis accuracy while accelerating simulation and design phases.
Reducing the complexity of France's building stock to better anticipate anticipate energy demand flexibility and the integration of solar solar resources
The aim of this work is to respond to the current challenges of energy transition in the building sector, France's leading energy consumer. French public policies are currently proposing far-reaching solutions, such as support for energy-efficient home renovation and incentives for the installation of renewable energy production systems. On a large scale, this is leading to structural changes for both building managers and energy network operators. As a result, players in the sector need to review their energy consumption and carbon impact forecasts, integrating flexibility solutions adapted to the French standard. Some flexibility levers are already in place to meet the challenges of energy and greenhouse gas emission reduction, but others need to be anticipated, taking into account long-term scenarios for energy renovation and the deployment of renewable energy sources, particularly photovoltaic energy, across the whole of France. The issue of massification is therefore an underlying one. That's why this thesis proposes to implement a methodology for reducing the size of the French installed base based on previously defined criteria. In particular, the aim will be to define a limited number of reference buildings that are statistically representative of the behavior resulting from the application of flexibility strategies that meet the challenges of energy efficiency and limiting greenhouse gas emissions. To this end, the CSTB (Centre Scientifique et Technique du Bâtiment) is developing and making available a database of French buildings (BDNB: Base de Données Nationale des Bâtiments), containing information on morphology, uses, construction principles and energy consumption and performance.
Development of microfluidic photoreactors for reproducible, quantitative evaluation of photoactive materials, coupled with on-line analysis by mass spectrometry and gas chromatography
The development of high performance photoactive materials (catalysts, semiconductors, sensitive films) for chemical conversion under light irradiation requires precise, reproducible and quantitative evaluation methods. Conventional batch approaches suffer from major limitations: poor control over residence time, temperature or light gradients, low exposed specific surface area and variable reproducibility. In this context, microfluidic photoreactors offer a promising alternative for structured screening and fine evaluation of photoactive materials, in particular thanks to their high surface/volume ratio, flow control and geometry adaptable to different irradiation configurations.
This work, linked to the PEPR LUMA SUNRISE project, aims to design, fabricate and characterize photonic microreactors specifically adapted to the fine evaluation of photoactive materials. The aim is to create a platform capable of generating quantitative and comparable data on the performance and stability of these materials, under well-defined conditions of throughput, irradiation and reaction environment, and then to couple them to high-level analytical techniques (GC, MS) for on-line identification of the products generated.
We propose to develop 4 axes during this thesis project: 1) development, characterization and optimization of the microfluidic platform for online liquid and gas measurement; 2) implementation of protocols for the deposition of photoactive materials 3) evaluation of photochemical performance and validation of the system with samples provided (SUNRISE partners) and on the degradation of pollutant by photochemistry (collaboration with a thesis in progress at the laboratory) and 4) Coupling of the reactor to online analytical methods (GC, MS).
Fine-grained and spatio-temporally grounded large multimodal models
This PhD project focuses on enhancing Large Multimodal Models (LMMs) through the integration of fine-grained and spatio-temporal information into training datasets. While current LMMs such as CLIP and Flamingo show strong performance, they rely on noisy and coarse-grained image-text pairs and often lack spatial or temporal grounding. The thesis aims to develop automatic pipelines to enrich image datasets with geographic and temporal metadata, refine captions using fine-grained semantic descriptors, and balance dataset diversity and compactness by controlling class-wise sample sizes.
Training strategies will incorporate hierarchical class structures and adapt protocols to improve alignment between caption elements and image regions. The work will also explore joint training regimes that integrate fine-grained, spatial, and temporal dimensions, and propose set-based inference to improve the diversity of generated outputs. The enriched datasets and models will be evaluated using existing or newly developed benchmarks targeting contextual relevance and output diversity. The project also addresses challenges in metadata accuracy, efficient model adaptation, and benchmarking methodologies for multi-dimensional model evaluation.
Applications include improved synthetic data generation for autonomous driving, enhanced annotation of media archives through contextual captioning, and better visual reasoning in industrial simulation scenarios.
In-depth electrical and material characterization of low-K spacer
As part of the European Chip Act, CEA-Leti is pioneering a new generation of transistors using FDSOI architecture. Our goal is to deliver advanced performance with a strong emphasis on materials and energy efficiency. As we push the limits of planar transistors at 10 nm and 7 nm, we face significant physical challenges, particularly in reducing parasitic elements like capacitance and access resistance, which are critical for minimizing energy loss and optimizing performance. We are eager to tackle these challenges together.
We are excited to offer a unique PhD opportunity for motivated students interested in the field of semiconductor device engineering. Join our team to work on the incorporation and characterization of low-k spacer for advanced 7-10nm FDSOI Technology. This PhD offers the chance to work on a groundbreaking project. If you're curious, innovative, and eager for a challenge, this opportunity is perfect for you!
The impact of the dielectric spacer nature has relevant effects on the overall transistor performances, specifically in non-fully overlapped configuration. The dielectric spacer integration, optimization and engineering remains a challenge and becomes crucial to address technology advancement and scaling down demand. Numerous spacer candidates (SiN, SiCO, SiCON, SiCBN) have been introduced and identified as promising solutions, however, they frequently suffer from inherent defects and adverse electrical characteristics, such as charge trapping and presence of undesired interface states, which hinder their and the overall transistors performance.
Within this framework, the objective of this PhD is to conduct a comprehensive investigation and electrical characterization (CV,IV, BTI, HCI…) of the material spacer (interface, volume), providing an in-depth analysis of transistor performance and its underlying mechanisms. Innovative ultrafast CV stress-measurement characterization on dielectric samples will be also carried out and the correlation between trapping performance and the deposition parameters used in their fabrication will be established. Additionally, the candidate will collaborate closely with experts to contribute to the thin film deposition and characterization of new materials through surface analyses and thin-film characterizations (ellipsometry, FTIR, XRR, XPS…)
Throughout this journey, you will gain a broad spectrum of knowledge, spanning microelectronics materials and processes, analog integrated design, all while addressing the unique challenge of advance 7-10 nm FDSOI technology. You'll collaborate with multidisciplinary teams to develop a deep understanding of FDSOI devices and analyze existing and new measurements. You'll also be part of an integrated multidisciplinary lab, working alongside a team composed of several permanent researchers, exploring a wide range of research applications.
Simplified Model for Rotary Tube Calcination
Since the vitrification lines at La Hague began operation in 1989, ORANO (formerly AREVA) has faced difficulties in controlling the calciner. Actions taken to significantly reduce these problems have considerably eased them, but without completely eliminating them. Most of the recommended actions are based on expert opinions, which themselves are based on inactive test results that don't cover all situations encountered by ORANO. To definitively resolve these control difficulties, it was decided to launch a more theoretical modeling study, while simultaneously investigating new calciner control instrumentation.
Microfluidics for biomimetic detection of airbone pathogens
Air represents a complex contamination pathway that is difficult to control and through which numerous biological, biochemical, or chemical agents can affect populations and healthcare workers. Standard detection approaches, whether qPCR, antigen tests, or ELISA tests, rely on reagents specific to known and targeted agents. These approaches are therefore unsuitable for detecting an unknown pathogen that could result in a new pandemic. To face such unknown agents, new biosensors will be needed to distinguish between pathogenic and non-pathogenic agents. Also, these sensors will have to be miniature for deployment.
With a new microfluidic system the present project aims to explore original approaches for conducting such detection without preconceived notions. Based on the laboratory's experience and developments, the PhD will include :
- developing new materials and designs to optimize and to enable multiple bioaerosol sampling;
- developing a biomimetic biochip and optimize molecular interactions using microflows controlled at the micro/milliscale.
You will design a microfluidic card integrating new detection strategies and study them experimentally using prototypes already developed in the laboratory.
Optimization of transports in the Gas Diffusion Layers of Proton Exchange Membrane Fuel Cells: Artificial Intelligent as a support to define optimal porous structures and usage
The design and manufacturing of innovative materials with required properties is a key objective for developing advanced technologies in the field of energy, such as Hydrogen and Alcaline fuel cells and Electrolysers. These improvements will contribute to propose even more attractive low-carbon electrical energy systems, with reduced pollution and green-house effects.
This thesis focuses on the Gas Diffusion Layer (GDL) which plays a crucial role on the performance and durability of Proton Exchange Membrane Fuel Cell (PEMFC).
Your main aim will be to set-up a numerical approach so as to propose improved porous structures to optimize the different transports inside a GDL, for given targets and constraints. To do so, you will make the bridge between advanced modeling of (electrical, heat, liquid, gas) transports in 3D porous media and artificial intelligence. You will then analyze the influence of the operating conditions on such optimal structures and propose design recommendations.
This work will be conducted in close relationship between world-renowned scientific actors : the fuel cells and the modelling teams of CEA/LITEN (Grenoble), the specialists of transports in porous media at CNRS/IMFT (Toulouse), and the specialists of GDL, modeling and AI at FZJ (Juelich, https://www.fz-juelich.de/en).
Scientific publications are expected and patents could also be proposed.