Experimentally realizing efficient quantum control with reinforcement learning

Robust optimization for quantum reinforcement learning control using partial observations

“…For example, Ref. [38] proposed a robust quantum control protocol against systematic errors by combining Short-cuts-To-Adiabatic (STA) and deep RL methods, which has been verified in the trappedion system [39]. Moreover, weak measurements can be implemented to reduce the measurement cost more significantly [40,41], which have been successfully applied to the double-well and dissipative qubit systems, respectively.…”

A Neural Network Assisted 171Yb+ Quantum Magnetometer

“…As schematically shown in Fig. 1 (c), our protocol is executed on a single 171 Yb + ion confined in a Paul trap which is shielded by permalloy to reduce the surrounding magnetic noise [28,29]. A magnetic field is applied to the ion leading to a Zeeman shift ≈ 10.0 MHz between | 0 and |1 .…”

Realizing a deep reinforcement learning agent discovering real-time feedback control strategies for a quantum system

“…Reinforcement learning, a subfield of machine learning, has had outstanding success in tasks ranging from board games [10] to robotics [11]. Reinforcement learning, however, has only very recently been started to be applied to complex physical systems, with training performed either on simulations [12][13][14][15][16][17][18] or directly in experiments [19][20][21][22][23][24][25][26], for example in laser [19,22,26], particle [20,21], softmatter [23] and quantum physics [24,25]. Specifically in the quantum domain, during the past few years, a number of theoretical works have pointed out the great promises of reinforcement learning for tasks covering state preparation [27][28][29][30][31], gate design [32], error correction [33][34][35] and circuit optimization/compilation [36,37], making it an important part of the machine learning toolbox for quantum technologies [38][39][40].…”

“…Specifically in the quantum domain, during the past few years, a number of theoretical works have pointed out the great promises of reinforcement learning for tasks covering state preparation [27][28][29][30][31], gate design [32], error correction [33][34][35] and circuit optimization/compilation [36,37], making it an important part of the machine learning toolbox for quantum technologies [38][39][40]. In first applications to quantum systems, reinforcement learning was experimentally deployed, but training was mostly performed based on simulations, specifically to optimize pulse sequences for atoms and spins [14,15,18]. Beyond that, there are two pioneering works demonstrating the training directly on experiments [24,25] which was used to optimize pulses for quantum gates [24] and to accelerate the tune-up of quantum dot devices [25].…”

“…Beyond that, there are two pioneering works demonstrating the training directly on experiments [24,25] which was used to optimize pulses for quantum gates [24] and to accelerate the tune-up of quantum dot devices [25]. However, in none of these experiments [14,15,18,24,25] real-time quantum feedback was required.…”