Learning the structure of open quantum systems
summary
Provides efficient algorithms that recover unknown Lindbladians for k‑local, bounded‑degree and some long‑range models, achieving structure learning with total evolution time scaling about D^3 log N/ε^2 and a large time resolution 1/D, at the cost of quasi‑polynomial classical postprocessing (fix planned). Quantum experiments are minimal: prepare–evolve–measure random Pauli eigenstates and Pauli measurements. Also gives the first efficient structure learning of Hamiltonians from high‑temperature Gibbs states; low‑temperature cases remain open.
AI-written summary · about
chapters (8)
- 0:00Learning open quantum systems
- 4:19Problem: Lindbladian learning
- 11:35Prior work and our results
- 18:53Algorithm overview
- 24:15Convex optimization and Richardson iterations
- 33:30Proof techniques and series expansion
- 41:04Results and open questions
- 53:20Q&A
related talks
44mOptimal Learning of Quantum Hamiltonians From High-Temperature Gibbs States
42mLearning Shallow Quantum Circuits and Quantum States Prepared by Shallow Circuits in Polynomial Time
56mRapid quantum ground state preparation via dissipative dynamics
30mQuantum Thermal Simulation
46mLocal Minima in Quantum Systems
01h 15mDissipative State Preparation and the Dissipative Quantum Eigensolver
44mEmergence of Universal Randomness in Quantum Many-body Dynamics
01h 04mLearning Beyond Stabilizer States