Quantum Machine Learning for Event Classification

An important part of the first-phase Quantum Computing (QC) project to HEP application is qunatum machine learning (QML) for physics data analysis. The ML is a powerful tool to find relation between input and output guided by the ground truth (supervised learning) or find an underlying structure in input data (unsupervised learning). Quantum-enhanced ML technique has potential for improving the conventional ML, e.g, by finding more effective discriminator(s) for new physics signal by mapping data into a higher-dimentional 'quantum' feature space.

Project

Quantum Machine Learning for Event Classfication

Quantum machine learning is considered to be a promising early QC application with the current NISQ device. The variational quantum algorithm (VQC), one of the supervised quantum learning methods, could bring a significant performance improvement in classification task by mapping data into quantum feature space. The variationoal algorithm uses tunable parameters to learn the properties of input data, making it possible to construct a shallow quantum circuit suitable for NISQ device.

We are exploring the application of VQC to the classification of new particles predicted in Supersymmetry (SUSY) theory from background processes using Qulacs and IBM Qiskit quantum computing framework.

SUSY signal classification with VQC

Shown here is a representative quantum circuit used in VQC. The input data x is encided into quantum feature space using Uin(x), then processed via U(𝜽) to produce output states, for which the measurement is performed. The measured output <Z> is used to evalute cost function for supervised learning.

The plot shows the results of VQC algorithm (labelled QCL) for the SUSY classification obtained using the Qulacs simulator, presented in terms of AUC values as a function of the number of training events. The results are compared with those obtained using the standard (classical) ML methods based on a boosted-decision tree (BDT) and a deep neural network (DNN).

Public Results

Publication

Quantum Gate Pattern Recognition and Circuit Optimization for Scientific Applications, W. Jang, K. Terashi, M. Saito, C. W. Bauer, B. Nachman, Y. Iiyama, T. Kishimoto, R. Okubo, R. Sawada, J. Tanaka, arXiv:2002.09935, 2021.
Event Classification with Quantum Machine Learning in High-Energy Physics, K. Terashi, M. Kaneda, T. Kishimoto, M. Saito, R. Sawada, J. Tanaka, Comput. Softw. Big Sci. 5, 2 (2021), arXiv:2002.09935.
Quantum Machine Learning in High Energy Physics, Wen Guan, Gabriel Perdue, Arthur Pesah, Maria Schuld, Koji Terashi, Sofia Vallecorsa, Jean-Roch Vlimant, Mach. Learn.: Sci. Technol. 2, 011003 (2021), arXiv:2005.08582.

Talks/Posters

量子コンピュータ (Japanese), Y. Iiyama, Flavor Physics workshop 2021, September 30, 2021.
Quantum Gate Pattern Recognition and Circuit Optimization for Scientific Applications, K. Terashi et al, 25th International Conference on Computing in High-Energy and Nuclear Physics (online), May 20, 2021.
Quantum Gate Pattern Recognition and Circuit Optimization for Scientific Applications, W. Jang et al (poster), US-Japan Hawaii Symposium of the US-Japan Science and Technology Cooperation Program (online), Apr 23, 2021.
Quantum Gate Pattern Recognition and Circuit Optimization for Scientific Applications (Japanese), 寺師弘二, 他, 量子ソフトウェア研究会, オンライン, 2021年3月29日
擬似量子メモリを用いた長い量子回路の近似手法 (Japanese), 大久保龍之介, 他, 日本物理学会第76回年次大会, オンライン, 2021年3月14日
高エネルギー物理への量子アルゴリズム応用のための量子回路最適化手法の開発 (Japanese), 張元豪, 他, 日本物理学会第76回年次大会, オンライン, 2021年3月14日
Quantum Gate Pattern Recognition and Circuit Optimization for Scientific Applications, K. Terashi et al, CERN openlab Technical Workshop (online), March 11, 2021.
Event classification with quantum machine learning and its application to HEP data analysis, R. Sawada, KEK Workshop on Femtotechnologies by quantum computers in 2021 (online), Mar 23, 2021.
Quantum Computing in High-Energy Physics and Related Fields, K. Terashi, University of Washington EPE seminar (online), Dec 3, 2020.
Quantum Computing ~ Application to Fundamental Science ~, K. Terashi, Colloquium at Baylor University (online), Nov 4, 2020.
高エネルギー実験でのデータ解析に向けた量子機械学習の応用 (Japanese), 寺師弘二, 他, 日本物理学会第76回年次大会, オンライン, 2020年9月14日.
量子機械学習のHEP応用 (Japanese), 寺師弘二, 新テラスケール研究会, オンライン, 2020年8月12日.
Quantum Machine Learning ~ Application to Event Classification in HEP ~, K. Terashi, KIAS seminar (online), May 25, 2020.
量子機械学習を用いた事象選別アルゴリズムとデータ解析への応用 (Japanese), 寺師弘二, 他, 日本物理学会第75回年次大会, 名古屋大学, 2020年3月16日.
Quantum Machine Learning ~ Application to Event Classification in HEP ~, K. Terashi, Machine Learning at LHC, Nagoya, February 6, 2020.
量子ゲート方式コンピュータによる実験素粒子物理データ解析 (Japanese), 齊藤真彦, 他, 「量子コンピュータ研究開発の現在とこれから」, 理化学研究所, 2020年1月29日.
Event Classification with Quantum Machine Learning, K. Terashi et al, CERN Openlab Technical Workshop, Geneva, January 22, 2020.

Members

Yutaro Iiyama, Tomoe Kishimoto, Ryu Sawada, Koji Terashi (in A-Z order)

Google Sites

Report abuse