Reinforcement Learning in Different Phases of Quantum Control

The ability to prepare a physical system in a desired quantum state is central to many areas of physics such as nuclear magnetic resonance, cold atoms, and quantum computing. Yet, preparing states quickly and with high fidelity remains a formidable challenge. In this work we implement cutting-edge Reinforcement Learning (RL) techniques and show that their performance is comparable to optimal control methods in the task of finding short, high-fidelity driving protocol from an initial to a target state in non-integrable many-body quantum systems of interacting qubits. RL methods learn about the underlying physical system solely through a single scalar reward (the fidelity of the resulting state) calculated from numerical simulations of the physical system. We further show that quantum state manipulation, viewed as an optimization problem, exhibits a spin-glass-like phase transition in the space of protocols as a function of the protocol duration. Our RL-aided approach helps identify variational protocols with nearly optimal fidelity, even in the glassy phase, where optimal state manipulation is exponentially hard. This study highlights the potential usefulness of RL for applications in out-of-equilibrium quantum physics.

[1]  M. Bukov,et al.  QuSpin: a Python package for dynamics and exact diagonalisation of quantum many body systems. Part II: bosons, fermions and higher spins , 2018, SciPost Physics.

[2]  Pankaj Mehta,et al.  Glassy Phase of Optimal Quantum Control. , 2018, Physical review letters.

[3]  David J. Schwab,et al.  A high-bias, low-variance introduction to Machine Learning for physicists , 2018, Physics reports.

[4]  Hartmut Neven,et al.  Universal quantum control through deep reinforcement learning , 2018, npj Quantum Information.

[5]  R. Sarpong,et al.  Bio-inspired synthesis of xishacorenes A, B, and C, and a new congener from fuscol† †Electronic supplementary information (ESI) available. See DOI: 10.1039/c9sc02572c , 2019, Chemical science.

[6]  Marin Bukov,et al.  Reinforcement Learning to Autonomously Prepare Floquet-Engineered States: Inverting the Quantum Kapitza Oscillator , 2018 .

[7]  F. Mintert,et al.  Fast adiabatic evolution by oscillating initial Hamiltonians , 2018, Physical Review A.

[8]  Enrique Solano,et al.  Measurement-based adaptation protocol with quantum reinforcement learning , 2018, Quantum Reports.

[9]  Xin Wang,et al.  Automatic spin-chain learning to explore the quantum speed limit , 2018, Physical Review A.

[10]  Florian Marquardt,et al.  Reinforcement Learning with Neural Networks for Quantum Feedback , 2018, Physical Review X.

[11]  José Miguel Hernández-Lobato,et al.  Taking gradients through experiments: LSTMs and memory proximal policy optimization for black-box quantum control , 2018, ISC Workshops.

[12]  Herschel Rabitz,et al.  Data-driven gradient algorithm for high-precision quantum control , 2017, 1712.01780.

[13]  Marin Bukov,et al.  Broken symmetry in a two-qubit quantum control landscape , 2017, 1711.09109.

[14]  Mario Krenn,et al.  Active learning machine learns to create new quantum experiments , 2017, Proceedings of the National Academy of Sciences.

[15]  Armin Rahmani,et al.  Optimal control of superconducting gmon qubits using Pontryagin's minimum principle: Preparing a maximally entangled state with singular bang-bang protocols , 2017, Physical Review A.

[16]  Stefano Carretta,et al.  Superadiabatic driving of a three-level quantum system , 2017 .

[17]  Armin Rahmani,et al.  Optimal control of Gmon qubits using Pontyagin's minimum principle: preparing a maximally entangled state with singular bang-bang protocols , 2017 .

[18]  Christopher R. Laumann,et al.  Clustering in Hilbert space of a quantum optimization problem , 2017, ArXiv.

[19]  Antonio Celani,et al.  Flow Navigation by Smart Microswimmers via Reinforcement Learning , 2017, Physical review letters.

[20]  D. Schuster,et al.  Speedup for quantum optimal control from automatic differentiation based on graphics processing units , 2016, 1612.04929.

[21]  Grant M. Rotskoff,et al.  Geometric approach to optimal nonequilibrium control: Minimizing dissipation in nanomagnetic spin systems. , 2016, Physical review. E.

[22]  A. Polkovnikov,et al.  Minimizing irreversible losses in quantum systems by local counterdiabatic driving , 2016, Proceedings of the National Academy of Sciences.

[23]  Matthias Troyer,et al.  Solving the quantum many-body problem with artificial neural networks , 2016, Science.

[24]  L. Hollenberg,et al.  Superadiabatic quantum state transfer in spin chains , 2016, 1604.04885.

[25]  M. Rigol,et al.  Emergent Eigenstate Solution to Quantum Dynamics Far from Equilibrium , 2015, 1512.05373.

[26]  M. Bukov,et al.  QuSpin: a Python Package for Dynamics and Exact Diagonalisation of Quantum Many Body Systems part I: spin chains , 2016, 1610.03042.

[27]  Hans-J. Briegel,et al.  Quantum-enhanced machine learning , 2016, Physical review letters.

[28]  Gautam Reddy,et al.  Learning to soar in turbulent environments , 2016, Proceedings of the National Academy of Sciences.

[29]  Hartmut Neven,et al.  Optimizing Variational Quantum Algorithms using Pontryagin's Minimum Principle , 2016, ArXiv.

[30]  A. Baksic,et al.  Accelerated quantum control using superadiabatic dynamics in a solid-state lambda system , 2016, Nature Physics.

[31]  L. D'alessio,et al.  Adiabatic Perturbation Theory and Geometry of Periodically-Driven Systems , 2016, 1606.02229.

[32]  Matteo G. A. Paris,et al.  Assessing the significance of fidelity as a figure of merit in quantum state reconstruction of discrete and continuous-variable systems , 2016, 1604.00313.

[33]  Roger G. Melko,et al.  Machine learning phases of matter , 2016, Nature Physics.

[34]  Grant M. Rotskoff,et al.  Near-optimal protocols in complex nonequilibrium transformations , 2016, Proceedings of the National Academy of Sciences.

[35]  Dries Sels,et al.  Geometry and non-adiabatic response in quantum and classical systems , 2016, 1602.01062.

[36]  Demis Hassabis,et al.  Mastering the game of Go with deep neural networks and tree search , 2016, Nature.

[37]  A. Baksic,et al.  Supplementary information for “ Speeding up adiabatic quantum state transfer by using dressed states ” , 2016 .

[38]  H. Neven,et al.  Digitized adiabatic quantum computing with a superconducting circuit , 2015, Nature.

[39]  I. Bloch,et al.  Optimal control of complex atomic quantum systems , 2015, Scientific Reports.

[40]  A. Zeilinger,et al.  Automated Search for new Quantum Experiments. , 2015, Physical review letters.

[41]  P. Manju,et al.  Fast machine-learning online optimization of ultra-cold-atom experiments , 2015, Scientific Reports.

[42]  Mads Kock Pedersen,et al.  Exploring the quantum speed limit with computer games , 2015, Nature.

[43]  D. Schuster,et al.  Speedup for quantum optimal control from GPU-based automatic differentiation , 2016 .

[44]  Shane Legg,et al.  Human-level control through deep reinforcement learning , 2015, Nature.

[45]  Barry C Sanders,et al.  High-Fidelity Single-Shot Toffoli Gate via Quantum Control. , 2015, Physical review letters.

[46]  Rahul Nandkishore,et al.  Nonlocal adiabatic response of a localized system to local manipulations , 2014, Nature Physics.

[47]  A. Retzker,et al.  Realization of a Quantum Integer-Spin Chain with Controllable Interactions , 2014, 1410.0937.

[48]  Masoud Mohseni,et al.  Quantum brachistochrone curves as geodesics: obtaining accurate minimum-time protocols for the control of quantum systems. , 2014, Physical review letters.

[49]  Tzyh Jong Tarn,et al.  Fidelity-Based Probabilistic Q-Learning for Control of Quantum Systems , 2014, IEEE Transactions on Neural Networks and Learning Systems.

[50]  Barry C. Sanders,et al.  Evolutionary Algorithms for Hard Quantum Control , 2014, 1403.0943.

[51]  D J Egger,et al.  Adaptive hybrid optimal quantum control for imprecisely characterized systems. , 2014, Physical review letters.

[52]  B. Lanyon,et al.  Quasiparticle engineering and entanglement propagation in a quantum many-body system , 2014, Nature.

[53]  S Montangero,et al.  Information theoretical analysis of quantum optimal control. , 2014, Physical review letters.

[54]  C. Jarzynski,et al.  Classical and Quantum Shortcuts to Adiabaticity for Scale-Invariant Driving , 2014, 1401.1184.

[55]  S. Lloyd,et al.  Complexity of controlling quantum many-body dynamics , 2013, 1301.6015.

[56]  Matteo G. A. Paris,et al.  Drawbacks of the use of fidelity to assess quantum resources , 2013, 1309.5325.

[57]  A. Campo,et al.  Shortcuts to adiabaticity by counterdiabatic driving. , 2013, Physical review letters.

[58]  G. C. Hegerfeldt,et al.  Driving at the quantum speed limit: optimal control of a two-level system. , 2013, Physical review letters.

[59]  C. Jarzynski Generating shortcuts to adiabaticity in quantum and classical dynamics , 2013, 1305.4967.

[60]  Marco Genovese,et al.  Experimental estimation of quantum discord for a polarization qubit and the use of fidelity to assess quantum correlations , 2013, 1305.4475.

[61]  H. Schättler,et al.  Geometric Optimal Control , 2012 .

[62]  Mazyar Mirrahimi,et al.  Real-time quantum feedback prepares and stabilizes photon number states , 2011, Nature.

[63]  C. Monroe,et al.  Onset of a quantum phase transition with a trapped ion quantum simulator. , 2011, Nature communications.

[64]  M. Greiner,et al.  Quantum simulation of antiferromagnetic spin chains in an optical lattice , 2011, Nature.

[65]  Tommaso Calarco,et al.  Chopped random-basis quantum optimization , 2011, 1103.0855.

[66]  S. Glaser,et al.  Second order gradient ascent pulse engineering. , 2011, Journal of magnetic resonance.

[67]  A. Gruslys,et al.  Comparing, optimizing, and benchmarking quantum-control algorithms in a unifying programming framework , 2010, 1011.4874.

[68]  Tommaso Calarco,et al.  Optimal control technique for many-body quantum dynamics. , 2010, Physical review letters.

[69]  M. Berry,et al.  Transitionless quantum driving , 2009 .

[70]  J. P. Garrahan,et al.  Dynamic Order-Disorder in Atomistic Models of Structural Glass Formers , 2009, Science.

[71]  Stuart A Rice,et al.  On the consistency, extremal, and global properties of counterdiabatic fields. , 2008, The Journal of chemical physics.

[72]  J. García-Ripoll Time evolution algorithms for Matrix Product States and DMRG , 2006, cond-mat/0610210.

[73]  Christopher M. Bishop,et al.  Pattern Recognition and Machine Learning (Information Science and Statistics) , 2006 .

[74]  J. García-Ripoll Time evolution of Matrix Product States , 2006, cond-mat/0602305.

[75]  A. Cavagna,et al.  Spin-glass theory for pedestrians , 2005, cond-mat/0505032.

[76]  Stuart A Rice,et al.  Assisted adiabatic passage revisited. , 2005, The journal of physical chemistry. B.

[77]  Timo O. Reiss,et al.  Optimal control of coupled spin dynamics: design of NMR pulse sequences by gradient ascent algorithms. , 2005, Journal of magnetic resonance.

[78]  L. Vandersypen,et al.  NMR techniques for quantum control and computation , 2004, quant-ph/0404064.

[79]  Richard S. Sutton,et al.  Reinforcement Learning: An Introduction , 1998, IEEE Trans. Neural Networks.

[80]  Stuart A. Rice,et al.  Adiabatic Population Transfer with Control Fields , 2003 .

[81]  Shlomo E. Sklarz,et al.  Loading a Bose-Einstein condensate onto an optical lattice: An application of optimal control theory to the nonlinear Schrödinger equation , 2002, cond-mat/0209195.

[82]  M. Mézard,et al.  Analytic and Algorithmic Solution of Random Satisfiability Problems , 2002, Science.

[83]  R. Brockett,et al.  Time optimal control in spin systems , 2000, quant-ph/0006114.

[84]  Kompa,et al.  Whither the future of controlling quantum phenomena? , 2000, Science.

[85]  S. Glaser,et al.  Unitary control in quantum ensembles: maximizing signal intensity in coherent spectroscopy , 1998, Science.

[86]  H. Rabitz,et al.  Teaching lasers to control molecules. , 1992, Physical review letters.

[87]  Y. Saad Analysis of some Krylov subspace approximations to the matrix exponential operator , 1992 .

[88]  W. H. Kirk And and Or , 1921 .