Quantum Reinforcement Learning

By Daoyi Dong, Chunlin Chen, Han-Xiong Li, Tzyh Jong Tarn. IEEE Transactions on Systems Man and Cybernetics Part B (Cybernetics), 2008

425 citations recorded by the source corpus.

The key approaches for machine learning, particularly learning in unknown probabilistic environments, are new representations and computation mechanisms. In this paper, a novel quantum reinforcement learning (QRL) method is proposed by combining quantum theory and reinforcement learning (RL). Inspired by the state superposition principle and quantum parallelism, a framework of a value-updating algorithm is introduced. The state (action) in traditional RL is identified as the eigen state (eigen action) in QRL. The state (action) set can be represented with a quantum superposition state, and the eigen state (eigen action) can be obtained by randomly observing the simulated quantum state according to the collapse postulate of quantum measurement. The probability of the eigen action is determined by the probability amplitude, which is updated in parallel according to rewards. Some related characteristics of QRL such as convergence, optimality, and balancing between exploration and exploitation are also analyzed, which shows that this approach makes a good tradeoff between exploration and exploitation using the probability amplitude and can speedup learning through the quantum parallelism. To evaluate the performance and practicability of QRL, several simulated experiments are given, and the results demonstrate the effectiveness and superiority of the QRL algorithm for some complex problems. This paper is also an effective exploration on the application of quantum computation to artificial intelligence.

Subfields: Artificial Intelligence

View Paper: https://www.semanticscholar.org/paper/86b2a1b37ae0d4591dc9902138e9db8ea303fabf

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Quantum Models of Cognition and Decision (2012, background): “1416 and allow L to be the integer value of c · [r (t) + γ · maxl Q (e, al, t)], where c > 0 is a free parameter (Dong et al., 2008).”
A novel DDPG method with prioritized experience replay (2017, background): “In addition, more experience selection and action selection methods for deep reinforcement learning may be inspired by quantum probability to design new exploration strategies [23]-[26].”
Quantum-enhanced machine learning (2016, result): “RL has been less investigated from a quantum information perspective, although some results have been reported [9, 10].”
Quantum Machine Learning for 6G Communication Networks: State-of-the-Art and Vision for the Future (2019, background): “In [174], a quantum value updating algorithm for quantum-powered reinforcement learning has been proposed.”
Quantum control theory and applications: A survey (2009, neutral): “its) and the state |ψi of a qubit can be written as |ψi = cos θ 2 |0i +eiϕsin θ 2 |1i (1) 4 where θ∈ [0,π], ϕ∈ [0,2π]. Then |0i and |1i correspond to the states zero and one for a classical bit [80], [85]. In practical applications, the quantum systems to be controlled are usually not simple closed ”
Variational Quantum Circuits for Deep Reinforcement Learning (2019, methodology): “DISCUSSIONS A. OVERVIEW OF QUANTUM MACHINE LEARNING A general review of the ﬁeld of quantum machine learning can refer to [46], [47]. As for quantum reinforcement learning,the early work can refer to [48], and in this work, the authors use the amplitude ampliﬁcation method to perform the action sele”