MEIDNet Prism

Learn, build and benchmark multimodal AI for materials discovery

Go from multimodal learning to working models and reproducible materials-discovery workflows. Explore architectures, train on your own data, compare results, and use the MEIDNet methodology to learn shared representations of materials and their properties.

▶ Try MEIDNetUse your datasetLearn multimodality

A. Babu et al., npj Comput. Mater. 12, 287 (2026) · Article · Code

cubic ABX3 perovskite · drag to rotate
Encodehi ← φh(hi, Σj mij)
Alignℒ = −log exp(sii/τ)Σj exp(sij/τ)
Fusez = ½ (zs + zp) ∈ ℝ128
Designz* = argminz ‖g(z) − y*‖²
Decodex* = D(z*)
ValidateEhull(x*) ≤ ε

MEIDNet: from concept to demonstration

From the properties you want to candidate crystals, block by block. Click a block to jump to it.

    0:00 Open the Studio →

    Contents

    Background on multimodal learning, model architectures, the MEIDNet Studio, datasets, benchmarks and contributed results.

    LearnWhat is multimodal learning?

    Modalities, the five challenges of multimodal learning, contrastive learning with a playground, and an interactive map of materials modalities.

    Learn multimodality →

    ArchitecturesWhich design fits my data?

    Early and late fusion, shared latent spaces, cross-attention and contrastive learning, drawn side by side, with an advisor that recommends one.

    Architecture Atlas →   Advisor →

    BuildMEIDNet Studio

    Upload structures and properties, train in the browser, set targets and rules, run the search and inspect every candidate in 3D. Every setting is saved in one YAML file.

    Try MEIDNet →   Recipes by problem →

    DatasetsDatasets and databases

    Perov-5, the published multimodal benchmark, other datasets used with MEIDNet, and 31 computed and experimental databases grouped by application.

    Datasets →   Databases →

    BenchmarksLeaderboards

    Fixed protocols per dataset: inverse design scored by stable, unique and novel candidates, property prediction and representation, with baselines. Any model can be scored with the same code.

    Benchmarks →

    CommunityContribute results

    Results are submitted through GitHub, re-run by the maintainer and marked as verified when they reproduce. Re-runs that disagree are published as well.

    Contribute a result →

    MEIDNet Studio

    The workflow from a dataset of structures and properties to candidate materials.

    Datayour materials
    Modelshared space
    Familycrystal type
    Ruleschemistry
    Targetswhat you want
    Searchinverse design
    Candidatesnew materials

    Citation

    A. Babu, R. A. Gouvêa, P. Vandergheynst and G.-M. Rignanese, MEIDNet: Multimodal generative AI framework for inverse materials design, npj Computational Materials (2026). doi:10.1038/s41524-026-02153-3 · arXiv:2601.22009

    @article{meidnet2026,
  title   = {MEIDNet: Multimodal generative AI framework for inverse materials design},
  author  = {Anand Babu and Rog{\'e}rio Almeida Gouv{\^e}a and Pierre Vandergheynst and Gian-Marco Rignanese},
  journal = {npj Computational Materials},
  year    = {2026},
  doi     = {10.1038/s41524-026-02153-3}
}

    Further reading

    1. Y. Bengio, A. Courville and P. Vincent, Representation learning: a review and new perspectives, IEEE Transactions on Pattern Analysis and Machine Intelligence 35, 1798–1828 (2013). doi:10.1109/TPAMI.2013.50
    2. Y. LeCun, Y. Bengio and G. Hinton, Deep learning, Nature 521, 436–444 (2015). doi:10.1038/nature14539
    3. B. Sanchez-Lengeling and A. Aspuru-Guzik, Inverse molecular design using machine learning: generative models for matter engineering, Science 361, 360–365 (2018). doi:10.1126/science.aat2663
    4. A. Babu, R. Almeida Gouvêa and G.-M. Rignanese, Toward automated discovery with generative models multimodal learning and closed loop workflows in inverse materials design, Cell Reports Physical Science 7, 103561 (2026). doi:10.1016/j.xcrp.2026.103561
    5. A. Babu and N. M. A. Krishnan, Multimodal and cross-modal learning techniques, APL Machine Learning 4, 030901 (2026). doi:10.1063/5.0346744