AI for Advanced Materials

Connecting artificial intelligence and experimentation to explore advanced materials.

The scientific scope, experimental resources and collaboration arrangements are being defined.

Project Overview

BECOME is structuring an applied research initiative that aims to combine artificial intelligence, experimental data and physics-based modelling to better guide the development of advanced materials.

The initial application under consideration is materials and nanomaterials for energy storage. The objective is to develop an approach linking composition, processing, structure and properties to inform formulation choices and the experiments to undertake.

In this approach, artificial intelligence supports scientific decision-making. Model-generated proposals must be assessed through testing and experimental validation.

The scientific challenge

Which variables best explain a material's behaviour? How can different formulations be compared using the available data? Which experiment would provide the most useful information before the next development decision?

The first research case will be scoped around a specific materials family and an application need agreed with partners. The aim is to make the question testable, identify the information available and determine where further measurements are needed. It is not to claim that a general-purpose system can already discover suitable materials automatically.

The first pilot will focus on one materials family and a question defined with the partners. In energy storage, this could involve predicting an electrochemical property or comparing formulations before a new experimental campaign.

From data to models, experiments and learning

The proposed approach connects data preparation, modelling, recommendations, experimentation and interpretation. Each stage informs the next, with experimental assessment central to the process.

  • Organise the scientific information

    The work will relate composition, processing conditions, structure, characterisation and measured properties. Data sources and test conditions will need to remain traceable so that comparisons can be interpreted in context.

  • Develop models around the research question

    AI and physics-based modelling will be considered according to the agreed problem and the information available. Their purpose is to help compare candidate formulations or identify informative tests, not to replace scientific interpretation.

  • Test the proposals and review the results

    Recommendations will be confronted with experiments. The findings will be used to examine model performance, identify limitations and decide whether the approach should be revised or extended. Evaluation criteria and comparison methods will be defined before drawing conclusions about improvements.

Models will be designed around the agreed research question, rather than applying a technology independently of the need. Predictive approaches may be combined with physics-based simulations and, where relevant, a digital twin of the system under investigation.

Predictions will be compared with a reference method and assessed using data or experiments not used to develop or tune the model. Validation should establish where the approach improves the decision, where uncertainty remains and under which conditions it can be reused.

BECOME's role

BECOME's contribution is to frame the scientific question, coordinate complementary expertise, develop the research methods and help define a suitable validation process. Experimental resources and responsibilities will be established with the participating organisations; the initiative does not assume that all required facilities are already available within the Institute.

The development and exploitation of outputs will follow the rights and responsibilities agreed by the contributors. BECOME will assess transfer opportunities for results it is entitled to use, without presuming ownership of partner data or background knowledge.

The industrial value we aim to establish

The intended value is better-directed testing, more structured interpretation of characterisation results and a clearer basis for deciding which formulations warrant further work.

A pilot should help a team examine whether the approach improves its research decisions and under which conditions. Possible benefits in development time, cost or performance remain objectives to evaluate. No percentage improvement, discovery rate or production-readiness claim is established at this stage.

Depending on the agreed problem, outputs could include an evaluated model, a comparison of formulations, recommendations for further experiments and documentation of performance, uncertainty and the intended scope of use.

Success criteria will be set before development begins. They may concern predictive quality, the relevance of shortlisted formulations or improvements in the experimental strategy.

A staged pathway towards application

The initiative is currently being structured. Feasibility, a demonstrator or pilot, validation and eventual transfer are planned stages whose scope will depend on the research case and the contributions agreed.

Where validation supports further use, potential routes include research contracts, co-development, additional test campaigns and licensing. BECOME's integrated international DeepTech Venture Studio can help examine the appropriate pathway, subject to the findings and the rights of the contributors.

Following validation, suitable methods and components could be adapted to further applications: data-preparation tools, predictive models, optimisation modules or digital twin functions.

  1. Structuring (current)
  2. Feasibility (planned)
  3. Demonstrator (planned)
  4. Validation (planned)
  5. Transfer or Industrial Deployment (planned)

Current stage: Structuring. Later stages are planned and remain subject to feasibility, resources, validation results and the relevant agreements.

Scientific context and a distinct project identity

The initiative has thematic connections with BIO-MICRO-GRADE, a collaborative research proposal involving advanced materials, modelling and energy-storage applications. Its public presentation provides related scientific context.

AI for Advanced Materials is a separate initiative. It is not a funded work package or an automatically authorised extension of BIO-MICRO-GRADE. The relationship does not transfer consortium membership, commitments, data access or intellectual-property rights.

Contributions that can shape the first research case

Industrial organisations can bring a materials, processing or energy-storage application need. Scientific teams can contribute experimental capabilities, methods and interpretation. Technology partners can contribute computing, modelling or data tools.

Institutions and funding partners can discuss resources for a defined stage of work. Contributions, budgets, deliverables and rights would be agreed for the particular collaboration, rather than promised before the research scope is established.

Evaluate computing tools and scientific-data approaches against experimental validation. Findings may support publications or joint references where the agreements permit.

Define a validation case with BECOME

Are you developing a material, process or energy-storage application? Let's discuss a research question and a validation scope.

Tell us about the challenge and the contribution your organisation could make. The first exchange should help define an appropriate next step and the resources it may require.

Discuss a Validation CaseDiscuss Partnership and Project Financing

Please share only non-confidential information in your initial enquiry.

Related expertise

Related researchers

BECOME Institute
ExpertScientific BoardScientist

Dr Mireille Esther Gettler Summa

Global Scientific Chair & France Scientific Lead

Senior Scientific Advisor — Data Science, Machine Learning & Applied AI

Mathematician and Data Science researcher with more than four decades of experience spanning statistical learning, Machine Learning, European R&I and applied AI.

View Scientific Profile
Prof. Sherif M. A. S. Keshk
ExpertScientific BoardScientist

Prof. Sherif M. A. S. Keshk

Tunisia Scientific Chair

Senior Researcher in Polymer Materials, Nanocomposites and Sustainable Materials Engineering

Sherif Keshk specialises in polymer materials, nanocomposites and sustainable materials engineering.

View details

Publications

32 records

Research article

01/2026 · Journal of Physics and Chemistry of Solids

Enhanced lithium-ion transport in oxidized polyvinyl alcohol/La0·56Li0·33TiO3 hybrid electrolytes for solid-state batteries

Amira Siai; Abdullah Y.A.AL. Zahrani; Radhouane Chtourou; et al.

This study explores the electrical and dielectric properties of hybrid solid electrolytes made from hydrothermally synthesized La0·56Li0·33TiO3 (LLTO) oxide and oxidized polyvinyl alcohol (OPVA). The influence of LLTO on ionic conductivity and dielectric behavior was assessed…

View details
Research article

26/11/2025 · Materials for Renewable and Sustainable Energy

Dual proton transport in Schiff base-modified chitosan functionalized with AMPS for fuel cell applications

Sonia Jebri; Walid Mabrouk; Ridha Elleuch; et al.

Proton exchange membranes (PEMs) require high proton conductivity, stability, and durability for fuel cell applications. This study reports the synthesis of N-(2-acrylamido-2-methylpropane sulfonyl) chitosan (CSB) via Schiff base functionalization with 2-acrylamido-2-methylpropane sulfonic acid (AMPS) and the…

View details
Research article

20/11/2025 · Polymers

Comparative Effects of Various Plasticizers on the Physicochemical Characteristics of Polyhydroxybutyrate (PHB) Film for Food Packaging

Siwar Taamallah; Sabrine Douiri; Sherif M. A. S. Keshk; et al.

This work examined the effects of four plasticizers, glycerol (GLY), potassium phosphate (PHOS), polyethylene glycol (PEG), and soy lecithin (SL), on the structural, surface, thermal, optical, and mechanical properties of polyhydroxybutyrate (PHB) films. FTIR spectra demonstrated…

View details
Research article

11/2025 · Materials Science and Engineering: B

In-situ synthesis of ZnO nanoparticles in oxidized polyvinyl alcohol nanocomposites for tunable optoelectronic and proton conductive performance

Ridha Elleuch; Maryam mallek; Khaled Charradi; et al.

This work presents the in-situ synthesis of zinc oxide (ZnO) nanoparticles in an oxidized polyvinyl alcohol (OPVA) matrix, which greatly enhances the final nanocomposites optical, mechanical, and proton conductive properties. Zinc acetate was used as a…

View details
Conference paper

15/07/2025 · 2025 11th International Conference on Control, Decision and Information Technologies (CoDIT)

Transfer Learning for Predicting Thermal Comfort in Office Environments with Climate Similar to Tunisia: Overcoming Data Scarcity with Deep GRU-BiGRU Models

Mohamed Khayri Rahmani; Hajer Chtioui; Jalel Ben Hadj Slama; et al.

This conference paper explores transfer learning with deep GRU-BiGRU models to address data scarcity in thermal comfort prediction for office environments with climates similar to Tunisia.

View details