Context

Federated Learning (FL) enables multiple clients to collaboratively train machine learning models without sharing their data. Instead, clients exchange local model updates with a central server, which uses them to improve a global model. While this paradigm enhances data privacy, recent studies have shown that FL remains vulnerable to privacy breaches. In particular, gradient inversion attacks can reconstruct sensitive client data from transmitted updates, posing serious privacy concerns [1]. Traditional approaches such as Deep Leakage from Gradients (DLG) [1], demonstrated that even simple models can leak identifiable information.

At the same time, diffusion models [3,4] have emerged as powerful generative frameworks capable of synthesizing realistic data from noisy or partial information. Recent works demonstrate that diffusion models can enhance gradient-based privacy attacks [5] and inspire novel privacy-preserving strategies [6,7].

Objectives

The goal of this internship is to explore the dual role of diffusion models in attacking and defending Federated Learning systems:

1. Attack Objectives
   • Implement diffusion-based gradient inversion attacks inspired by GGDM [5], conditioning the diffusion process on client gradients or embeddings to reconstruct original inputs.
   • Compare against existing attack baselines, such as DLG [1], iDLG [2], and GAN-based approaches [8], evaluating reconstruction quality and scalability.
2. Defense Objectives
   • Design a diffusion-inspired privacy mechanism based on forward stochastic noising of client updates, making gradient inversion harder while preserving model convergence.
   • Explore hybrid mechanisms integrating differential privacy with diffusion-based noise injection, inspired by DP-Fed-FinDiff [6] and Personalized Federated Diffusion [7].
3. Comparative Study
   • Benchmark the proposed approaches against established FL defenses, such as Secure Aggregation [9], DP-SGD [10].
   • Analyze trade-offs between privacy and robustness in different scenarios, including non-i.i.d. data distributions.

[1] Zhu et al. Deep Leakage from Gradients. NeurIPS 2019. 

[2] Zhao et al. iDLG: Improved Deep Leakage from Gradients. ICLR 2020. 

[3] Ho et al. Denoising Diffusion Probabilistic Models. NeurIPS 2020.

[4] Song et al. Score-Based Generative Modeling through SDEs. ICLR 2021.

[5] Gu et al. Gradient-Guided Diffusion Models for Privacy Attacks. 2024. 

[6] Liu et al. DP-Fed-FinDiff: Differentially Private Federated Diffusion for Tabular Data. 2024.

[7] Chen et al. Personalized Federated Diffusion with Privacy Guarantees. 2025.

[8] Fang et al. GIFD: A Generative Gradient Inversion Method with Feature Domain Optimization ICCV 2023.

[9] Bonawitz et al. Practical Secure Aggregation for Privacy-Preserving ML. CCS 2017.

[10] Abadi et al. Deep Learning with Differential Privacy. CCS 2016.

Qu’attendons-nous de vous ?

The candidate should be in the last year of an engineering school or a master student (Bac+5) in a field related to machine learning/AI, who wishes to conduct research and development in an emerging, yet impactful field, in a collaborative environment. The intern will work in a team of researchers, post-docs, and PhD students who are actively investigating various challenges and aspects of federated learning. The candidate should have knowledge in machine learning and optimisation, and be skilled in Python programming and in using various machine learning libraries and frameworks.

Rejoignez-nous, venez développer vos compétences et en acquérir de nouvelles !

Vous avez encore un doute ? Nous vous proposons :

• L'opportunité de travailler au sein d'une organisation de renommée mondiale dans le domaine de la recherche scientifique,
• Un environnement unique dédié à des projets ambitieux au profit des grands enjeux sociétaux actuels,
• Une expérience à la pointe de l’innovation, comportant un fort potentiel de développement industriel,
• Des moyens expérimentaux exceptionnels et un encadrement de qualité,
• De réelles opportunités de carrière à l’issue de votre stage
• Un poste au cœur de la métropole grenobloise, facilement accessible via la mobilité douce favorisée par le CEA,
• Une participation aux transports en commun à hauteur de 85%,
• Un équilibre vie privée – vie professionnelle reconnu,
• Un restaurant d'entreprise,
• Une politique diversité et inclusion,

Conformément aux engagements pris par le CEA en faveur de l'intégration des personnes handicapées, cet emploi est ouvert à toutes et à tous. Le CEA propose des aménagements et/ou des possibilités d'organisation pour l'inclusion des travailleurs handicapés.