Creativity is a powerful, renewable resource, vital for addressing complex challenges and preserving cultural heritage. AI plays a key role in this journey, offering tools and insights that expand our potential and unlock previously unimaginable possibilities. From my perspective, my work in AI applied to creativity represents an exciting frontier where science and human imagination converge. Foundational disciplines such as physics, mathematics, and statistics provide the tools to model complex systems, identify patterns, and generate novel solutions, all of which are essential for advancing creativity in meaningful ways. AI, supported by these scientific principles, becomes not just a collaborator but a powerful enabler that amplifies our ability to innovate, solve problems, and express ourselves.
My work is driven by a belief in the transformative power of creativity, supported and amplified by AI, to shape a world where cultural and creative expressions flourish. By bridging tradition with innovation, I aim to contribute to a future where creativity becomes a cornerstone of progress, resilience, and sustainable development.
Selected Publications
A. Londei, M. Benati, D. Lanzieri, V. Loreto, “Dreaming Learning”, NeurIPS 2024, 2nd Workshop on Intrinsically Motivated Open-ended Learning (2024).
Zeghal, J., Lanzieri, D., Aubourg, E., Boucaud, A., Lanusse, F., Louppe, G., Bayer, A., LSST Dark Energy Science Collaboration. Simulation-Based Inference Benchmark for LSST Weak Lensing Cosmology. Astronomy & Astrophysics journal (2024).
Lanzieri, D., Zeghal, J., Makinen, L., Boucaud, A., Lanusse, F., Starck, J.L., Optimal Neural Summarisation for Full-Field Cosmological Implicit Inference. Astronomy & Astrophysics journal (2024).
Lanzieri, D., Lanusse, F., Chirag, M., Horowitz, B., Harnois-D´eraps, J., Starck, J.L., LSST Dark Energy Science Collaboration. Forecasting the power of Higher Order Weak Lensing Statistics with automatically differentiable simulations. Astronomy & Astrophysics journal (2023).
Lanzieri, D., Lanusse, F., Starck, J. L. Hybrid Physical-Neural ODEs for Fast N-body Simulations. ICML 2022 Workshop on Machine Learning for Astrophysics (2022).