We warmly welcome Miriam Klopotek, the latest addition to our group of independent junior research group leaders.
Born in Boston, Massachusetts, USA, Miriam Klopotek obtained her Bachelor’s degree in Physics at the Humboldt University in Berlin. Her Master’s degree was obtained at the University of Tübingen where she also pursued her doctoral studies. Working in the field of theoretical statistical physics (of soft condensed matter), with emphasis on computational approaches, she gained her doctoral degree with distinction under the supervision of Prof. Dr. Martin Oettel. Her thesis is entitled “Hard Rods on Lattices between Two and Three Dimensions: Nonequilibrium, Equilibrium, and Machine Learning”.
Miriam Klopotek is working on emergent phenomena in many-body systems with a clear relation to both simulations and machine learning. In her research plan, she aims to work out even stronger possible cross-references between many-body systems and machine learning and thus also to arrive at statements about ML models.
"We are delighted to be able to attract an outstanding young scientist like Miriam Klopotek. She is an expert in the simulation of physical many-body systems, whose structures and regularities she would like to make usable for intelligent systems. Thus, her research agenda fits perfectly with SimTech's ambitious interdisciplinary methodological goals", underlines Steffen Staab.
Before joining the Cluster of Excellence SimTech, Miriam Klopotek was an early career group leader in the Machine Learning Cluster of Excellence in Tübingen. She is also closely related to Cyber Valley where she is an associated faculty member of the International Max Planck Research School for Intelligent Systems.
“I am incredibly excited to be here at SimTech. It feels like a place particularly well-suited to realize my research goals and grow as a scientist: I saw my own aims reflected broadly in those of the cluster – to combine (physical) simulations with data-driven approaches, enabling new modes of thinking about both. I have been trying to interpret and understand more deeply machine learning algorithms by training them on simulation data from many-body model systems – offering me intuitions about learning that are rooted in fundamentals of statistical mechanics. Now, I am interested in “putting in” many-body physics directly into learning algorithms – to leverage their elsuive, shape-shifting capabilities for collective behavior and self-assembly – and to introduce physics-based learning schemes that alleviate the notorious “black box” interpretability problem. With my “physicist” way of viewing things, I sincereley look forward to interdisciplinary exchange and collaboration with experts in other fields", explains Miriam Klopotek.