The dilemma of quantum neural networks

Yang, Qian; Wang, Xinbiao; Du, Yuxuan; Wu, Xingyao; Tao, Dacheng

doi:10.48550/arxiv.2106.04975

Cited by 8 publications

(9 citation statements)

References 66 publications

(80 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Huang et al (2021) proved that for quantum processes, QML provides no advantage in minimizing the average prediction error, only providing an advantage for minimizing the worst-case prediction error. Kübler et al (2021) and Qian et al (2021) showed there is little indication that QML can improve supervise learning.…”

Section: Findings and Discussionmentioning

confidence: 99%

Playing Atari with Hybrid Quantum-Classical Reinforcement Learning

Lockwood,

2021

Preprint

View full text Add to dashboard Cite

Despite the successes of recent works in quantum reinforcement learning, there are still severe limitations on its applications due to the challenge of encoding large observation spaces into quantum systems. To address this challenge, we propose using a neural network as a data encoder, with the Atari games as our testbed. Specifically, the neural network converts the pixel input from the games to quantum data for a Quantum Variational Circuit (QVC); this hybrid model is then used as a function approximator in the Double Deep Q Networks algorithm. We explore a number of variations of this algorithm and find that our proposed hybrid models do not achieve meaningful results on two Atari games -Breakout and Pong. We suspect this is due to the significantly reduced sizes of the hybrid quantumclassical systems.

show abstract

Section: Findings and Discussionmentioning

confidence: 99%

Playing Atari with Hybrid Quantum-Classical Reinforcement Learning

Lockwood,

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…While there is substantial evidence supporting reductions in generalization error [226,227,228,229], evidence of poor generalization performance under certain constructions also exists (e.g. see this recent paper [417]). This may be partly attributable to shallower quantum circuits providing better utility bounds than deeper circuits [418], which contrasts with classical neural network intuition where increased layer depth is associated with an exponential increase in model expressiveness [419].…”

Section: Quantum Machine Learningmentioning

confidence: 98%

Biology and medicine in the landscape of quantum advantages

Cordier¹,

Sawaya²,

Guerreschi³

et al. 2021

Preprint

View full text Add to dashboard Cite

Quantum computing holds significant potential for applications in biology and medicine, spanning from the simulation of biomolecules to machine learning approaches for subtyping cancers on the basis of clinical features. This potential is encapsulated by the concept of a quantum advantage, which is typically contingent on a reduction in the consumption of a computational resource, such as time, space, or data. Here, we distill the concept of a quantum advantage into a simple framework that we hope will aid researchers in biology and medicine pursuing the development of quantum applications. We then apply this framework to a wide variety of computational problems relevant to these domains in an effort to i) assess the potential of quantum advantages in specific application areas and ii) identify gaps that may be addressed with novel quantum approaches. Bearing in mind the rapid pace of change in the fields of quantum computing and classical algorithms, we aim to provide an extensive survey of applications in biology and medicine that may lead to practical quantum advantages.

show abstract

“…However, quantum supervised and unsupervised learning models may encounter trainability issues, where the gradients exponentially vanish for the number of qubits [64,65]. Moreover, a recent study has shown that, in fact, the performance of quantum supervised learning models on realworld datasets could be worse than that of classical learning models [66].…”

Section: Introductionmentioning

confidence: 99%

Unentangled quantum reinforcement learning agents in the OpenAI Gym

Hsiao¹,

Du²,

Chiang³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Classical reinforcement learning (RL) has generated excellent results in different regions ; however, its sample inefficiency remains a critical issue. In this paper, we provide concrete numerical evidence that the sample efficiency (the speed of convergence) of quantum RL could be better than that of classical RL, and for achieving comparable learning performance, quantum RL could use much (at least one order of magnitude) fewer trainable parameters than classical RL. Specifically, we employ the popular benchmarking environments of RL in the OpenAI Gym, and show that our quantum RL agent converges faster than classical fully-connected neural networks (FCNs) in the tasks of CartPole and Acrobot under the same optimization process. We also successfully train the first quantum RL agent that can complete the task of LunarLander in the OpenAI Gym. Our quantum RL agent only requires a single-qubit-based variational quantum circuit without entangling gates, followed by a classical neural network (NN) to post-process the measurement output. Finally, we could accomplish the aforementioned tasks on the real IBM quantum machines. To the best of our knowledge, none of the earlier quantum RL agents could do that.

show abstract

The dilemma of quantum neural networks

Cited by 8 publications

References 66 publications

Playing Atari with Hybrid Quantum-Classical Reinforcement Learning

Playing Atari with Hybrid Quantum-Classical Reinforcement Learning

Biology and medicine in the landscape of quantum advantages

Unentangled quantum reinforcement learning agents in the OpenAI Gym

Contact Info

Product

Resources

About