Unsupervised beyond-standard-model event discovery at the LHC with a novel quantum autoencoder.

Area of Science:

• High Energy Physics
• Quantum Computing
• Machine Learning

Background:

• The Standard Model of particle physics has been incredibly successful but does not explain phenomena like dark matter or dark energy.
• Proton colliders like the Large Hadron Collider (LHC) are crucial for searching for new physics beyond the Standard Model.
• Unsupervised anomaly detection offers a promising avenue for identifying unexpected signals in complex collider data.

Purpose of the Study:

• To explore the application of unsupervised anomaly detection for identifying new physics at the LHC.
• To introduce and evaluate a novel quantum autoencoder circuit ansatz for this task.
• To investigate the role of entanglement and magic in the performance of quantum autoencoder circuits.

Main Methods:

• Development of a novel quantum autoencoder circuit ansatz tailored for anomaly detection.
• Evaluation of the quantum autoencoder's performance on simulated new physics 'signal' events and varying problem sizes.
• Comparison with classical autoencoders and previously proposed quantum autoencoders.
• Investigation of quantum circuit properties, including entanglement (Meyer-Wallach measure) and magic (stabiliser 2-Rényi entropy).

Main Results:

• The novel quantum autoencoder demonstrated superior performance in anomaly detection compared to previous approaches.
• Classical autoencoders were developed that outperformed prior quantum methods but were still surpassed by the new quantum ansatz.
• The quantum autoencoder achieved high performance with significantly fewer trainable parameters.
• Both entanglement and magic metrics decreased during training, suggesting learned parameters reduce these properties without minimizing them.

Conclusions:

• Quantum autoencoders show significant potential for discovering physics beyond the Standard Model at the LHC.
• The developed quantum autoencoder is a robust tool for unsupervised anomaly detection in high-energy physics.
• Further research into the interplay of entanglement and magic in quantum machine learning is warranted.