Internship Online Simulation-Based Inference For Large-Scale Scientific Models H/F INRIA

Internship: Online Simulation-Based Inference for Large-Scale Scientific Models
Le descriptif de l'offre ci-dessous est en Anglais
Type de contrat : Convention de stage

Niveau de diplôme exigé : Bac +4 ou équivalent

Fonction : Stagiaire de la recherche

A propos du centre ou de la direction fonctionnelle

The Centre Inria de l'Université de Grenoble groups together almost 600 people in 23 research teams and 9 research support departments.

Staff is present on three campuses in Grenoble, in close collaboration with other research and higher education institutions (Université Grenoble Alpes, CNRS, CEA, INRAE, ...), but also with key economic players in the area.

The Centre Inria de l'Université Grenoble Alpe is active in the fields of high-performance computing, verification and embedded systems, modeling of the environment at multiple levels, and data science and artificial intelligence. The center is a top-level scientific institute with an extensive network of international collaborations in Europe and the rest of the world.

Contexte et atouts du poste

The candidate will be supervised by Bruno Raffin (Inria Grenoble) and Pedro L. C. Rodrigues (Inria Grenoble). He or she will work mainly at the DataMove team, located at the IMAG building in UGA campus, and in close collaboration with the Statify team, located in Inria Montbonnot. He or she will have access to a team of experts in high-performance computing and machine learning that will help him or her to kickstart the project under the best conditions. The candidate will also have access to supercomputers to run experiments.

The length of the internship is4 months minimumand the start date is flexible, but need a 2 months delay before starting the interhsip due to administrative constraints. The DataMove and STATIFY teams are friendly and stimulating environment that gathers Professors, Researchers, PhD and Master students all leading research on High-Performance Computing and Machine Learning. The city of Grenoble is a student-friendly city surrounded by the Alps mountains, offering a high quality of life and where you can experience all kinds of mountain-related outdoor activities.

Mission confiée

Context

Researchers are turning to machine learning to tackle various problems in science, from biology to astro-physics and fluid dynamics. The project that we propose is part of this growing AI4Science movement, focusing on a key challenge in experiments: figuring out which model parameters best match the data we observe (Figure 1). More specifically, we use simulation-based inference (SBI) [1], a Bayesian approach that leverages deep generative models, such as conditional normalizing flows and score-diffusion models, to approximate the posterior distribution assigning higher probability to parameter values most likely to have produced an observed data.

Despite recent successes of the SBI framework across various applied domains, its applicability is currently constrainedto relatively small-scale models. The primary goal of this project is to extend the capabilities of SBI to accommodate simulators that rely on solving large systems of differential equations to generate observations. As such, the candidate will have the opportunity to work in the exciting intersection between modern machine learning methods (e.g. sampling with diffusion models, embeddings with transformers, training with flow matching) and high performance computing (e.g. handling large-scale parallel simulators, multi-node and GPU training on large supercomputers).

Principales activités

When considering large scale simulations the amount of data produced can be overwhelming and the execution time too long, calling to resort to supercomputers and High Performance Computing (HPC). To optimize performance and reduce costs (power, storage, compute time), training can be performed online. Multiple simulations are executed concurrently and continuously to produce data that are used asap, without being stored in files, by a training process that also runs concurrently with these simulations (Fig. 2) [2, 3]. The traditional SBI workflow consisting of (simulate store, then store train) has to be re-visited to properly support and leverage this online training workflow (simulate buffer train).

To be more specific, consider the usual SBI approach of training a conditional neural density estimator q that approximates the target posterior through the minimization of

The batch of N pairs of parameters (i) and simulations (xi) is provided upfront and the loss function is minimized via some variant of stochastic gradient descent. Note that this is well motivated because when N , one can show that the minimizer of Equation 1 is indeed the target posterior p( x). However, it is not clear how the minimization behaves when the training samples are obtained sequentially due to simulation constraints and/or sampled from a different distribution than p(,x) as one would do when trying to reduce the number of calls to the simulator.

During the M2 internship, the candidate will explore the following questions:

- What is the direct impact of an online paradigm for simulations on the usual batched SBI training? What are the precise bottlenecks and challenges to this transition?

- Are there other loss functions more appropriate to minimize instead of Equation 1 when working under the paradigm of large-scale simulators?

- Can approaches from online reinforcement learning to make smart queries to the simulator and minimize total cost of compute can be reused and adapted?

References

[1] MichaelDeistler, JanBoelts, PeterSteinbach, GuyMoss, ThomasMoreau, ManuelGloeckler, PedroLCRodrigues, Julia Linhart, Janne K Lappalainen, Benjamin Kurt Miller, et al. Simulation-based inference: A practical guide. arXiv preprint arXiv:2508.12939, 2025.

[2] Sofya Dymchenko and Bruno Raffin. Loss-driven sampling within hard-to-learn areas for simulation-based neural network training. In MLPS 2023-Machine Learning and the Physical Sciences Workshop at NeurIPS 2023-37th conference on Neural Information Processing Systems, pages 1-5, 2023.

[3] LucasThibautMeyer,MarcSchouler,RobertAlexanderCaulk,AlejandroRibés,andBrunoRaffin. Highthroughput training of deep surrogates from large ensemble runs. In Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis, pages 1-16, 2023.

Compétences

- Strong mathematical background, specially advanced concepts in machine learning and statistics.
- Good working knowledge on Python and its scientific computing ecosystem (scipy, numpy, pytorch, etc).
- Some practical experience with running experiments on parallel machines will be a plus.
- Excellent writing and oral skills in French and English.

Avantages

- Subsidizedmeals
- Partial reimbursement of public transport costs
- Leave: for annual work contract 7 weeks of annual leave + 10 extra days off due to RTT (statutory reduction in working hours) + possibility of exceptional leave (sick children, moving home, etc.)
- Possibility of teleworking (90 days / year for an annual contract) and flexible organization of working hours at the condition of team leader approval
- Social, cultural and sports events and activities

Rémunération

€4.35 per hour of actual presence at 1 January 2025.

About 590€ gross per month (internship allowance)