Public Abstract

DE-SC0020416: Application of Quantum Machine Learning to High Energy Physics Analysis at LHC using IBM Quantum Computer Simulators and IBM Quantum Computer Hardware

Award Status: Inactive

Institution: Board of Regents of the University of Wisconsin System, operating as University of Wisconsin-Madison, Madison, WI
UEI: LCLSJAGTNZQ7
DUNS: 161202122

Most Recent Award Date: 06/12/2021
Number of Support Periods: 2
PM: Love, Jeremy

Current Budget Period: 09/01/2020 - 08/31/2022
Current Project Period: 09/01/2019 - 08/31/2022
PI: Wu, Sau Lan

Supplement Budget Period: N/A

Public Abstract

ABSTRACT: Application of Quantum Machine Learning to High Energy Physics Analysis at LHC using IBM Quantum Computer Simulators and IBM Quantum Computer Hardware Sau Lan Wu, Physics Department, University of Wisconsin-Madison (Principle Investigator) Miron Livny, Computing Sciences Department, University of Wisconsin-Madison (Co-Investigator) Federico Carminati and Alberto Di Meglio, CERN openlab, IT Department, CERN (Co-Investigators) Ivano Tavernelli and Stefan Woerner, IBM Research Zurich (Co-Investigators) (1) Our goal: The ambitious HL-LHC program will require nearly insurmountable computing resources in the next two decades. New technologies are being sought after to replace the present computing infrastructure. A burning question is whether quantum computer can solve the ever growing demand of computing resources in High Energy Physics in general and physics at LHC in particular. Our goal here is to explore and to demonstrate that Quantum Computing can be the new paradigm (Proof of Principle). The experimental programs of PI Wu at the LHC revolve around one major objective: discovery of new physics. This requires the identification of rare signals in immense backgrounds. Using machine learning algorithms greatly enhance our ability to achieve this objective. Our group in the ATLAS/LHC is one of the groups which have pioneered the use of machine learning in high profile physics analysis. We have used machine learning algorithm on the measurement of Higgs coupling to top quark pairs (ttH). The impact of this ttH channel resulted in the CERN press release on June 4, 2018. However, with a fast increasing volume of data in the future HL-LHC program, applying quantum machine learning method may well be a new direction to go. Specifically, our goals are: 1) To Perform Research and Development of Quantum Machine Learning and Data Analysis Techniques, with Qubit Platform, using IBM Quantum Simulators and IBM Quantum Computer Hardware to Enhance Efficiency and Analysis Methods for HEP at LHC; 2) To Enhance the Software Development of Quantum Machine Learning for HEP at the LHC to provide Scalable Quantum Codes and Tools for Future HEP Analysis. (2) Our Interdisciplinary collaborations: We collaborate with an interdisciplinary un-funded team of physicists and computer scientists from Wisconsin Computing Science Department, CERN openlab of CERN IT Department, and IBM Research Zurich. Among us, we have ample expertise in HEP and QIST. In addition, we expect to get access to Quantum Computer Hardware of 20 to 50 qubits in the not too distant future once IBM and CERN sign the cooperation agreement which will happen very soon. (3) Work in Progress: We have made good progress in the LHC physics channel ttH (Higgs coupling to the top quarks) with quantum machine learning algorithms using IBM Quantum Simulator and IBM Quantum Computer Hardware. In the last six months, we have given five presentations in recent QIS conferences on this topic. We are also presenting talks on this subject in EPS-HEP 2019 and LP 2019. Besides the LHC flagship ttH channel, we are extending this ttH experience to three other LHC flagship physics channels: Higgs to two muons, double Higgs production, and search for Dark Matter (mono-Higgs). (4) Impact on HEP: Our physics program on Higgs Physics and Dark Matter Searches at LHC using machine learning at present and using quantum machine learning in the future corroborates the U.S. particle physics community’s visions documented in the 2014 report from the Particle Physics Project Prioritization Panel (P5). We are aligned with 2 out of the 5 science drivers: (1) Using the Higgs boson as a new tool for discovery, (2) Identify the new physics of Dark Matter. (5) Impact on QIST: Our goal is to pioneer the use of qubit platform to solve the technical challenges in deploying quantum machine learning on HEP analysis using IBM Quantum Computer Simulator and IBM Quantum Computer Hardware. We plan to overcome challenges to encode the classical LHC datasets with many variables per event into limited number of qubits by entangling qubits, to investigate and develop a variational quantum circuit to extend our analysis to larger numbers of events, and to enhance quantum algorithms to advance machine learning for HEP. We will work on quantum error mitigation in the context of quantum machine learning algorithms. We will also explore various quantum machine learning methods and enhance their performances.

ABSTRACT: Application of Quantum Machine Learning to High Energy Physics Analysis at LHC using IBM Quantum Computer Simulators and IBM Quantum Computer Hardware

Sau Lan Wu, Physics Department, University of Wisconsin-Madison (Principle Investigator)

Miron Livny, Computing Sciences Department, University of Wisconsin-Madison (Co-Investigator)

Federico Carminati and Alberto Di Meglio, CERN openlab, IT Department, CERN (Co-Investigators)

Ivano Tavernelli and Stefan Woerner, IBM Research Zurich (Co-Investigators)

(1) Our goal: The ambitious HL-LHC program will require nearly insurmountable computing resources in the next two decades. New technologies are being sought after to replace the present computing infrastructure. A burning question is whether quantum computer can solve the ever growing demand of computing resources in High Energy Physics in general and physics at LHC in particular. Our goal here is to explore and to demonstrate that Quantum Computing can be the new paradigm (Proof of Principle).

The experimental programs of PI Wu at the LHC revolve around one major objective: discovery of new physics. This requires the identification of rare signals in immense backgrounds. Using machine learning algorithms greatly enhance our ability to achieve this objective. Our group in the ATLAS/LHC is one of the groups which have pioneered the use of machine learning in high profile physics analysis. We have used machine learning algorithm on the measurement of Higgs coupling to top quark pairs (ttH). The impact of this ttH channel resulted in the CERN press release on June 4, 2018. However, with a fast increasing volume of data in the future HL-LHC program, applying quantum machine learning method may well be a new direction to go. Specifically, our goals are:

1) To Perform Research and Development of Quantum Machine Learning and Data Analysis Techniques, with Qubit Platform, using IBM Quantum Simulators and IBM Quantum Computer Hardware to Enhance Efficiency and Analysis Methods for HEP at LHC;

2) To Enhance the Software Development of Quantum Machine Learning for HEP at the LHC to provide Scalable Quantum Codes and Tools for Future HEP Analysis.

(2) Our Interdisciplinary collaborations: We collaborate with an interdisciplinary un-funded team of physicists and computer scientists from Wisconsin Computing Science Department, CERN openlab of CERN IT Department, and IBM Research Zurich. Among us, we have ample expertise in HEP and QIST. In addition, we expect to get access to Quantum Computer Hardware of 20 to 50 qubits in the not too distant future once IBM and CERN sign the cooperation agreement which will happen very soon.

(3) Work in Progress: We have made good progress in the LHC physics channel ttH (Higgs coupling to the top quarks) with quantum machine learning algorithms using IBM Quantum Simulator and IBM Quantum Computer Hardware. In the last six months, we have given five presentations in recent QIS conferences on this topic. We are also presenting talks on this subject in EPS-HEP 2019 and LP 2019. Besides the LHC flagship ttH channel, we are extending this ttH experience to three other LHC flagship physics channels: Higgs to two muons, double Higgs production, and search for Dark Matter (mono-Higgs).

(4) Impact on HEP: Our physics program on Higgs Physics and Dark Matter Searches at LHC using machine learning at present and using quantum machine learning in the future corroborates the U.S. particle physics community’s visions documented in the 2014 report from the Particle Physics Project Prioritization Panel (P5). We are aligned with 2 out of the 5 science drivers: (1) Using the Higgs boson as a new tool for discovery, (2) Identify the new physics of Dark Matter.

(5) Impact on QIST: Our goal is to pioneer the use of qubit platform to solve the technical challenges in deploying quantum machine learning on HEP analysis using IBM Quantum Computer Simulator and IBM Quantum Computer Hardware. We plan to overcome challenges to encode the classical LHC datasets with many variables per event into limited number of qubits by entangling qubits, to investigate and develop a variational quantum circuit to extend our analysis to larger numbers of events, and to enhance quantum algorithms to advance machine learning for HEP. We will work on quantum error mitigation in the context of quantum machine learning algorithms. We will also explore various quantum machine learning methods and enhance their performances.