PhD Student in Computer Science (M/F) - Explainable Causal Inference and Prediction for Geopolymer Data

Job Description

Thesis Subject

The GEOLIANT project (France 2030) aims to develop an artificial intelligence tool to accelerate the design of geopolymer formulations, by predicting their technical properties (porosity,mechanical strength, durability), environmental properties (life-cycle assessment) and economic properties (cost, viability), and by bidirectionally identifying candidate formulations from desired target properties.

The analyses currently available on the project's data — whether from partners, from the literature, or augmented/generated through other work in the project — allow correlations to be identified between formulation parameters (precursor composition, Na₂O/SiO₂ ratios, curing conditions) and material properties, but are not sufficient to guarantee the reliability of design recommendations: a correlation does not indicate whether changing a parameter will actually change a property. This limitation is especially critical given that the project's ultimate goal — inverse design of formulations from target properties — implicitly relies on causal relationships.

The goal of the thesis is to develop explainable causal inference methods over the project's property-graph datalake, in order to identify causal (rather than merely correlational) relationships between formulation parameters and material properties, and to produce prediction models whose outputs are interpretable by domain experts. The work will focus in particular on causal discovery from heterogeneous and partially incomplete observational data, the representation of causal chains as paths within the property graph, and the production of explanations usable both for direct prediction (formulation → properties) and for inverse design (target properties → candidate formulations).

The PhD candidate will build on the project's property-graph datalake and will work closely with the other PhD candidates on the project, with the causal and predictive models developed relying directly on the consolidated and augmented data produced elsewhere. The work will combine causal inference, explainable machine learning, graph learning, and a strong interdisciplinary dimension with the materials science experts of the consortium (LMDC), needed to validate the physico-chemical plausibility of the identified causal relationships.

Your Work Environment

The thesis is part of the GEOLIANT project, supported by BPI France under the France 2030 program, which aims to develop and industrialize low-carbon geopolymer binders from construction waste, as a sustainable alternative to traditional cement, a heavy CO₂ emitter. The project rests on the development of innovative formulations and on the establishment of digital and predictive tools based on artificial intelligence, in order to accelerate research, evaluation, and performance validation.

The PhD candidate will be affiliated with the LIRIS laboratory and will work within an academic and industrial consortium bringing together stakeholders from civil engineering, materials science, and the environment. The thesis work will take place in a multidisciplinary environment, at the interface between computer science (data extraction and integration, graphs, AI) and materials engineering.

The thesis will be carried out in close interaction with the project's partners, in a setting that fosters co-design, expert validation, and the transfer of results toward operational applications, particularly in the context of pilot construction sites. It offers a stimulating research environment, combining methodological contributions with industrial and environmental impact.

The position falls within a sector subject to the protection of scientific and technical potential (PPST — protection du potentiel scientifique et technique), and therefore requires, in accordance with regulations, that your arrival be authorized by the competent authority of the French Ministry of Higher Education and Research (MESR).

Constraints and risks

Thesis carried out within the framework of a multi-partner collaborative project, involving constraints related to coordination, planning, and dependence on partners' contributions.

Dependence on the quality and availability of data consolidated and augmented elsewhere in the project (partner data, extracted literature, generated data), whose heterogeneity and partially incomplete nature directly condition the identifiability and robustness of the causal models.

Risk inherent to causal inference from observational data: no a priori guarantee of causal identifiability in the presence of unobserved confounding factors, requiring explicit assumptions and iterative validation with the consortium's materials science experts.

Difficulty in evaluating the causal validity and quality of the explanations produced in a domain where the underlying physico-chemical mechanisms are not always fully established, which may require evolving methodological and technical choices (causal discovery algorithms, explainability methods).

Need for a balance between AI-based automation and human validation, which may lengthen certain experimental phases.

Constraints linked to the contractual milestones and deliverables of the project funded by BPI France / France 2030.

Interdisciplinarity of the subject, requiring time to become familiar with concepts in materials science and civil engineering.

Occasional travel possible for meetings, workshops, or validation sessions at partner sites.

Compensation and benefits

Compensation

2300 € gross monthly

Annual leave and RTT

44 jours

Remote Working practice and compensation

Pratique et indemnisation du TT

Transport

Prise en charge à 75% du coût et forfait mobilité durable jusqu’à 300€

About the offer

About the CNRS

The CNRS is a major player in fundamental research on a global scale. The CNRS is the only French organization active in all scientific fields. Its unique position as a multi-specialist allows it to bring together different disciplines to address the most important challenges of the contemporary world, in connection with the actors of change.

CNRS

The research professions