Self-Attention for Quantum Entanglement Prediction

Anuj Gore, Roopayan Ghosh, Dylan Lewis, Sougato Bose

Keywords: Quantum entanglement estimation, Randomized measurements, Self-Attention, Machine learning for quantum physics

TL;DR: ML models, including self-attention networks, estimate Renyi-2 entanglement entropy from classical shadows with far fewer measurements, enabling low-shot, scalable entanglement characterization on near-term quantum devices.

Abstract: Quantum entanglement is a powerful resource in quantum mechanics and quantum information processing. However, its reliable quantification remains challenging due to the exponential growth of the underlying Hilbert space with system size, which renders full state reconstruction infeasible. Moreover, experimentally estimating entanglement typically requires a large number of measurement samples leading to a significant overhead. In this paper, we present two models, a feed-forward neural network and an attention-based model, to accurately predict the entanglement of random states. Our results demonstrate that machine-learning method consistently outperform conventional analytical approaches across a range of qubit numbers, highlighting the advantages of machine learning for the efficient quantification of quantum resources.

Submission Number: 75