A quantum machine learning-based predictive analysis of CERN collision events.

Area of Science:

• High-Energy Physics
• Quantum Computing
• Machine Learning

Background:

• Quantum computing advancements drive exploration of quantum algorithms.
• Quantum Machine Learning (QML) is investigated for potential advantages in data analysis.

Purpose of the Study:

• To apply and evaluate Quantum Neural Network (QNN) and Quantum Long Short-Term Memory (QLSTM) models for regression tasks.
• To compare the prediction accuracy and computational efficiency of QML models against classical regression methods using CERN datasets.

Main Methods:

• Utilized two CERN datasets: Dielectron events and Proton collision.
• Implemented and analyzed Quantum Neural Network (QNN) and Quantum Long Short-Term Memory (QLSTM) algorithms.
• Compared QML performance against classical algorithms like CatBoost, analyzing various ansatz designs.

Main Results:

• QML models achieved comparable accuracy to some classical methods, but advanced algorithms like CatBoost yielded superior results.
• Increased circuit complexity in QNN and QLSTM did not significantly enhance prediction accuracy.
• QLSTM with simpler ansatz designs showed particular promise for high-energy physics data.

Conclusions:

• QML models, especially QLSTM, present a promising avenue for modeling high-energy physics data.
• Balancing quantum circuit complexity with performance is crucial for effective QML application.
• Further research on quantum hardware is necessary to ascertain the real-world applicability of these QML models.