| Item type |
SIG Technical Reports(1) |
| 公開日 |
2021-06-24 |
| タイトル |
|
|
タイトル |
A study on the expressibility and learnability of quantum circuit learning |
| タイトル |
|
|
言語 |
en |
|
タイトル |
A study on the expressibility and learnability of quantum circuit learning |
| 言語 |
|
|
言語 |
eng |
| 資源タイプ |
|
|
資源タイプ識別子 |
http://purl.org/coar/resource_type/c_18gh |
|
資源タイプ |
technical report |
| 著者所属 |
|
|
|
Grid Inc. |
| 著者所属 |
|
|
|
Engineering department, The University of Electro-Communications |
| 著者所属 |
|
|
|
Grid Inc.／Engineering department, The University of Electro-Communications |
| 著者所属 |
|
|
|
Grid Inc. |
| 著者所属 |
|
|
|
Engineering department, The University of Electro-Communications／i-PERC, The University of Electro-Communications |
| 著者所属 |
|
|
|
Engineering department, The University of Electro-Communications／i-PERC, The University of Electro-Communications／Grid Inc. |
| 著者所属(英) |
|
|
|
en |
|
|
Grid Inc. |
| 著者所属(英) |
|
|
|
en |
|
|
Engineering department, The University of Electro-Communications |
| 著者所属(英) |
|
|
|
en |
|
|
Grid Inc. / Engineering department, The University of Electro-Communications |
| 著者所属(英) |
|
|
|
en |
|
|
Grid Inc. |
| 著者所属(英) |
|
|
|
en |
|
|
Engineering department, The University of Electro-Communications / i-PERC, The University of Electro-Communications |
| 著者所属(英) |
|
|
|
en |
|
|
Engineering department, The University of Electro-Communications / i-PERC, The University of Electro-Communications / Grid Inc. |
| 著者名 |
Chih-Chieh, Chen
Masaya, Watabe
Kodai, Shiba
Masaru, Sogabe
Katsuyoshi, Sakamoto
Tomah, Sogabe
|
| 著者名(英) |
Chih-Chieh, Chen
Masaya, Watabe
Kodai, Shiba
Masaru, Sogabe
Katsuyoshi, Sakamoto
Tomah, Sogabe
|
| 論文抄録 |
|
|
内容記述タイプ |
Other |
|
内容記述 |
Using quantum circuits for supervised machine learning is one potential way to harness quantum advantage for Noisy Intermediate-Scale Quantum hardware. Many implementations and algorithms have been proposed and studied for quantum circuit learning, but the learnability and generalization ability of quantum circuit ansatz is not well-understood yet. In this work, we study the relation between the circuit ansatz, the expressive power, and the PAC-learnability of quantum circuit learning. The model complexity and generalization ability are studied using a KL-divergence based measure, a VC-dimension upper bound, and various numerical simulation. Our result provides a way to understand the learnability of quantum circuit. |
| 論文抄録(英) |
|
|
内容記述タイプ |
Other |
|
内容記述 |
Using quantum circuits for supervised machine learning is one potential way to harness quantum advantage for Noisy Intermediate-Scale Quantum hardware. Many implementations and algorithms have been proposed and studied for quantum circuit learning, but the learnability and generalization ability of quantum circuit ansatz is not well-understood yet. In this work, we study the relation between the circuit ansatz, the expressive power, and the PAC-learnability of quantum circuit learning. The model complexity and generalization ability are studied using a KL-divergence based measure, a VC-dimension upper bound, and various numerical simulation. Our result provides a way to understand the learnability of quantum circuit. |
| 書誌レコードID |
|
|
収録物識別子タイプ |
NCID |
|
収録物識別子 |
AA12894105 |
| 書誌情報 |
量子ソフトウェア（QS）
巻 2021-QS-3,
号 5,
p. 1-5,
発行日 2021-06-24
|
| ISSN |
|
|
収録物識別子タイプ |
ISSN |
|
収録物識別子 |
2435-6492 |
| Notice |
|
|
|
SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc. |
| 出版者 |
|
|
言語 |
ja |
|
出版者 |
情報処理学会 |
IPSJ-QS21003005.pdf (1.1 MB)
