Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
TT: Fachverband Tiefe Temperaturen
TT 2: Focus Session: Artificial Intelligence in Condensed Matter Physics I (joint session TT/DY)
TT 2.7: Vortrag
Montag, 18. März 2024, 12:45–13:00, H 0104
Adversarial Hamiltonian learning of quantum dots in a minimal Kitaev chain — •Rouven Koch1, David van Driel2,3, Alberto Bordin2,3, Jose L. Lado1, and Eliska Greplova3 — 1Department of Applied Physics, Aalto University, Espoo, Finland — 2QuTech, Delft University of Technology, Delft, The Netherlands — 3Kavli Institute of Nanoscience, Delft University of Technology, Delft, The Netherlands
Knowledge of the underlying Hamiltonian in quantum devices is key for tuning and controlling experimental quantum systems. Here we demonstrate an adversarial machine learning framework capable of Hamiltonian learning of a quantum dot chain from noisy experimental measurements. We train a convolutional conditional generative adversarial network with simulated data of the differential conductances based on a Kitaev chain model. The trained model is able to predict the parameters determining the sweet spot conditions of the two-quantum-dot system at which the predicted mid-gap bound state emerges. This gives us a fast and numerically efficient way to explore the phase diagram describing the transition between elastic co-tunning and Andreev reflection regimes and thus is suitable to assist the sweet-spot tuning of the Kitaev chains. The application of our methodology to experimental measurements in an InSb nanowire shows promising results in extracting Hamiltonians from measurements, potentially supporting the hard task of tuning quantum-dot systems into distinct Hamiltonian regimes.
Keywords: Deep Learning; Quantum Dots; Kitaev Chain; Machine Learning; Majorana Fermions