Entanglement transitions induced by quantum-data collection

Shane P. Kelly,Jamir Marino

DOI: https://doi.org/10.1103/PhysRevA.111.L010402

2025-01-18

Abstract:We present an entanglement transition in an array of qubits, induced by the transfer of quantum information from a system to a quantum computer. This quantum-data collection is an essential protocol in quantum machine learning algorithms that promise exponential advantage over their classical counterparts. In this and an accompanying work [Phys. Rev. A 111, 012425 (2025)], we identify sufficient conditions for an entanglement transition to occur in the late time state of the system and quantum computer. In this letter, we present an example entanglement transition occurring in a system comprised of a 1D chain of qubits evolving under a random brickwork circuit. After each layer, a fraction $p$ of sites undergo noisy quantum transduction in which quantum information is transferred to a quantum computer but at the cost of introducing noise from an environment. For an entanglement transition to occur, we argue that the environment must obtain the same amount of information as gained by the computer. Under this condition, the circuit shows a transition from volume law to area law entanglement as the rate $p$ is increased above a critical threshold. Our work delineates the prerequisites for quantum-data collection to induce entanglement transitions, thereby establishing a foundational framework for emergent entanglement phenomena in protocols relevant to quantum machine learning.

Quantum Physics,Disordered Systems and Neural Networks,Statistical Mechanics,High Energy Physics - Theory