
Marie Didier
Speaker at AMLD Africa 2026
Marie Didier is a scientist and entrepreneur with a Ph.D. in Nonlinear Optics from EPFL and advanced degrees in physics and conservation science. Her career bridges research, innovation, and cultural diplomacy, from developing cutting-edge imaging technologies to building initiatives that connect local expertise with global impact.
As the founder and CEO of MATIS, a Swiss deep-tech startup, Marie leads the development of portable multispectral imaging and AI-based solutions for art assessment, security, and material identification. Her work merges machine learning, optics, and cultural heritage science, transforming traditional methods of preservation and authentication into scalable and secure digital tools.
Before founding MATIS, Marie worked as a Kress-Mellon Fellow at the National Gallery of Art in Washington, D.C., and as a PostDoc for industry at CSEM, where she specialized in advanced microscopy and algorithmic imaging.
Deeply committed to bridging culture, society, and knowledge, Marie co-founded and directs RestArt Beirut, a philanthropic and cultural diplomacy initiative dedicated to connecting education and academic expertise with traditional craftsmanship, fostering local talent, and empowering communities through cultural heritage preservation.
In parallel, she is dedicated to connecting tradition and innovation by bringing technology where it can generate tangible societal impact. A field she is deeply engaged in is the preservation of cultural heritage, where she contributes to the development of advanced tools for assessment, conservation, and the fight against art trafficking helping nations protect the invaluable cultural assets that define their history and identity. Her work bridges science, technology, and culture, fostering collaboration between researchers, institutions, and local communities worldwide.
Rooted in both science diplomacy and cultural diplomacy, Marie’s mission is to create meaningful connections between deep-tech innovation and societal impact, ensuring that technology serves not only progress but also the preservation and transmission of cultural and scientific heritage.
At AMLD Africa
WITS University, Johannesburg, South Africa
Decoding Cultural Heritage: hardware-enabled AI and material DNA
This talk presents a hardware-enabled, physics-informed AI approach to multispectral imaging for cultural heritage material identification. By combining real and synthetic spectral data, we build robust machine learning models to extract material DNA and enable scalable material identification.







