Towards Faster Reinforcement Learning of Quantum Circuit Optimisation: Exponential Reward Functions

Ioana Moflic; Alexandru Paler

doi:10.1145/3611315.3633259

Towards Faster Reinforcement Learning of Quantum Circuit Optimisation: Exponential Reward Functions

Research output: Chapter in Book/Report/Conference proceeding › Conference article in proceedings › Scientific › peer-review

Abstract

Reinforcement learning for the optimization of quantum circuits uses an agent whose goal is to maximize the value of a reward function that decides what is correct and what is wrong during the exploration of the search space. It is an open problem how to formulate reward functions that lead to fast and efficient learning. We propose an exponential reward function which is sensitive to structural properties of the circuit. We benchmark our function on circuits with known optimal depths, and conclude that our function is reducing the learning time and improves the optimization. Our results are a next step towards fast, large scale optimization of quantum circuits.

Original language	English
Title of host publication	Proceedings of the 18th ACM International Symposium on Nanoscale Architectures, NANOARCH 2023
Publisher	ACM
ISBN (Electronic)	9798400703256
DOIs	https://doi.org/10.1145/3611315.3633259
Publication status	Published - 25 Jan 2024
MoE publication type	A4 Conference publication
Event	ACM International Symposium on Nanoscale Architectures - Dresden, Germany Duration: 18 Dec 2023 → 20 Dec 2023 Conference number: 18

Conference

Conference	ACM International Symposium on Nanoscale Architectures
Abbreviated title	NANOARCH
Country/Territory	Germany
City	Dresden
Period	18/12/2023 → 20/12/2023

Keywords

optimization
quantum circuit
reinforcement learning

Access to Document

10.1145/3611315.3633259

https://arxiv.org/abs/2311.12509

Cite this

@inproceedings{a57c0a05650d4b66afa013e97169b07a,

title = "Towards Faster Reinforcement Learning of Quantum Circuit Optimisation: Exponential Reward Functions",

abstract = "Reinforcement learning for the optimization of quantum circuits uses an agent whose goal is to maximize the value of a reward function that decides what is correct and what is wrong during the exploration of the search space. It is an open problem how to formulate reward functions that lead to fast and efficient learning. We propose an exponential reward function which is sensitive to structural properties of the circuit. We benchmark our function on circuits with known optimal depths, and conclude that our function is reducing the learning time and improves the optimization. Our results are a next step towards fast, large scale optimization of quantum circuits.",

keywords = "optimization, quantum circuit, reinforcement learning",

author = "Ioana Moflic and Alexandru Paler",

note = "Publisher Copyright: {\textcopyright} 2023 ACM.; ACM International Symposium on Nanoscale Architectures, NANOARCH ; Conference date: 18-12-2023 Through 20-12-2023",

year = "2024",

month = jan,

day = "25",

doi = "10.1145/3611315.3633259",

language = "English",

booktitle = "Proceedings of the 18th ACM International Symposium on Nanoscale Architectures, NANOARCH 2023",

publisher = "ACM",

address = "United States",

}

Towards Faster Reinforcement Learning of Quantum Circuit Optimisation: Exponential Reward Functions. / Moflic, Ioana; Paler, Alexandru.
Proceedings of the 18th ACM International Symposium on Nanoscale Architectures, NANOARCH 2023. ACM, 2024. 21.

Research output: Chapter in Book/Report/Conference proceeding › Conference article in proceedings › Scientific › peer-review

TY - GEN

T1 - Towards Faster Reinforcement Learning of Quantum Circuit Optimisation

T2 - ACM International Symposium on Nanoscale Architectures

AU - Moflic, Ioana

AU - Paler, Alexandru

N1 - Conference code: 18

PY - 2024/1/25

Y1 - 2024/1/25

N2 - Reinforcement learning for the optimization of quantum circuits uses an agent whose goal is to maximize the value of a reward function that decides what is correct and what is wrong during the exploration of the search space. It is an open problem how to formulate reward functions that lead to fast and efficient learning. We propose an exponential reward function which is sensitive to structural properties of the circuit. We benchmark our function on circuits with known optimal depths, and conclude that our function is reducing the learning time and improves the optimization. Our results are a next step towards fast, large scale optimization of quantum circuits.

AB - Reinforcement learning for the optimization of quantum circuits uses an agent whose goal is to maximize the value of a reward function that decides what is correct and what is wrong during the exploration of the search space. It is an open problem how to formulate reward functions that lead to fast and efficient learning. We propose an exponential reward function which is sensitive to structural properties of the circuit. We benchmark our function on circuits with known optimal depths, and conclude that our function is reducing the learning time and improves the optimization. Our results are a next step towards fast, large scale optimization of quantum circuits.

KW - optimization

KW - quantum circuit

KW - reinforcement learning

UR - http://www.scopus.com/inward/record.url?scp=85184299313&partnerID=8YFLogxK

U2 - 10.1145/3611315.3633259

DO - 10.1145/3611315.3633259

M3 - Conference article in proceedings

AN - SCOPUS:85184299313

BT - Proceedings of the 18th ACM International Symposium on Nanoscale Architectures, NANOARCH 2023

PB - ACM

Y2 - 18 December 2023 through 20 December 2023

ER -

Towards Faster Reinforcement Learning of Quantum Circuit Optimisation: Exponential Reward Functions

Abstract

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this