RESEARCH
Juan Pablo focuses on learning generative causal models from sparse temporal observations during cellular reprogramming. His work centers on developing efficient algorithms to generate ensembles of dynamical models and applying supervised deep learning for pattern discovery in large-scale simulation datasets, aiming to address fundamental questions about human nature and inspire transformative, nature-inspired intelligent technologies.
At King Abdullah University of Science and Technology (KAUST), he is affiliated with the Applied Mathematics and Computational Science (AMCS) program and collaborates with the Biological and Environmental Sciences and Engineering (BESE) program. As a Ph.D. candidate under Jesper Tegner, Narsis Kiani, and David Gomez Cabrero, his doctoral research, “Discovery of Hidden Control Variables and Modeling Non-Linear Biological Systems,” combines machine learning and bifurcation theory to uncover dynamical systems in a data-driven approach. He gained international experience through a research internship at the University of Cambridge (2024), working on high-dimensional data compression using Koopman operator theory under Prof. Pietro Liò, and has optimized machine learning methods for time-series prediction, achieving superior accuracy with transformer architectures. He has also contributed to statistical analyses of pharmaceutical nanostructures, enhancing protocols for their development.
Juan Pablo has presented his work at prestigious conferences, most recently at the EMBO | EMBL Symposium: Theory and Concepts in Biology (2025, Heidelberg, Germany), where he won the Best Poster Prize for his poster “Reconstructing Transitions Between Dynamical Regimes Driven by Unknown Variables Using Known Universal Normal Forms.” In 2023, he spoke at the 22nd International Conference on Systems Biology (Connecticut, USA) on “Synergizing Normal Form Dynamical Systems with Foundational Machine Learning: Towards Efficient Decoding of Nonlinear Time-Series Phenomena,” and at the 21st International Conference on Systems Biology (2022, Berlin, Germany) with the talk “Discovery of Hidden Control Variables and Modeling Non-Linear Dynamical Biological Systems.” He also presented at the SIAM Conference on Applications of Dynamical Systems (2021, virtual) with “Discovery of Non-Linear Regime-Switching Dynamical Models in State-Space using Normal Forms.” Earlier in his career, he spoke at the 5th International and Interdisciplinary Workshop on Mathematical Modeling (2018, Villa de Leyva, Colombia), presenting on sustainable development models. His publications include the arXiv preprint “IHCV: Discovery of Hidden Time-Dependent Control Variables in Non-Linear Dynamical Systems” (2023, ready for submission), co-authored with Jesper Tegner, Narsis Kiani, and others.

BIOGRAPHY
Juan Pablo Munoz Diaz holds a B.Sc. in Mathematics (2014) and an M.Sc. in Applied Mathematics (2018) from the National University of Colombia, Manizales. His master’s thesis, under Gerard Olivar Tost, analyzed bifurcations in sustainable development models, and his undergraduate thesis explored solutions to the Schrödinger equation for cubic crystalline networks. He has extensive teaching experience, having lectured on calculus, discrete mathematics, probability and statistics, dynamical systems, and linear algebra at the University of Caldas and National University of Colombia (2016–2018). He also served as a tutor and coordinator for the Youth Mathematics Contest at the National University of Colombia (2013).
Juan Pablo received the Best Academic Average Award for four periods during his undergraduate studies (2010–2014) and the Best Poster Prize at the EMBO | EMBL Symposium (2025). He has completed diplomats in university teaching and rural development contextualization, supporting his educational outreach in Colombia. His computational skills include Python, PyTorch, scikit-learn, MATLAB, LaTeX, and Mathematica. He is fluent in Spanish and English, with proficiency in French (DELF A2) and basic German (A1.2).