Model-independent estimation of the properties of quantum states is a central challenge in quantum technologies, as experimental imperfections, drifts, and imprecise models of the actual quantum dynamics inevitably hinder accurate reconstructions. Here, we introduce a training strategy for photonic quantum extreme learning machines in which both the learning stage and the optimization of the measurement settings are performed entirely with classical light, while inference is carried out on genuinely quantum states. The protocol is based on the identity between the normalized output intensities following the evolution of coherent states through a linear optical reservoir, and the output statistics obtained with separable input quantum states. Building on this correspondence, we implemented a model-free, gradient-based optimization of the reservoir measurement projection directly on experimental data, without relying on a prior model of the device transformation. We experimentally show that the resulting classical-to-quantum transfer enables accurate reconstruction of single-qubit Pauli observables for previously unseen single-photon states, and extends to the estimation of a two-qubit entanglement witness for arbitrary bipartite states. Beyond demonstrating a qualitatively distinct form of out-of-distribution generalization across the classical-to-quantum boundary, our results identify a practical route to fast, adaptive, and resource-efficient training of photonic quantum learning devices.

Entanglement percolation aims at generating maximal entanglement between any two nodes of a quantum network by utilizing strategies based solely on local operations and classical communication between the nodes. As it happens in classical percolation theory, the topology of the network is crucial, but also the entanglement shared between the nodes of the network. In a network of identically partially entangled states, the network topology determines the minimum entanglement needed for percolation. In this work, we generalize the protocol to scenarios where the initial entanglement shared between each two nodes of the network is not the same but has some randomness. In such cases, we find that for classical entanglement percolation, only the average initial entanglement is relevant. In contrast, the quantum entanglement percolation protocol generally performs worse under these more realistic conditions.

Understanding equilibration times in closed quantum systems is essential for characterising their approach to equilibrium. Chaotic many-body systems are paradigmatic in this context: they are expected to thermalise according to the eigenstate thermalisation hypothesis and exhibit spectral properties well described by random matrix theory (RMT). While RMT successfully captures spectral correlations, its ability to provide quantitative predictions for equilibration timescales has remained largely unexplored. Here, we study equilibration within RMT using the framework of equilibration as dephasing, focusing on closed systems whose Hamiltonians are drawn from the Gaussian unitary ensemble (GUE). We derive an analytical expression that approximates the average equilibration time of the GUE and show that it is independent of both the initial state and the choice of observable, a consequence of the rotational invariance of the GUE. Numerical simulations confirm our analytical expression and demonstrate that our approximation is in close agreement with the true average equilibration time of the GUE. We find that the equilibration time decreases with system size and vanishes in the thermodynamic limit. This unphysical result indicates that the true equilibration timescale of realistic chaotic many-body systems must be dominated by physical features not captured by random matrix ensembles – the GUE in particular.

Cascaded emission from the biexciton state of a quantum dot results in polarization entangled photon pairs. However, modelling this system becomes challenging when photon emission is cavity-mediated due to the large Hilbert space and non-Markovian nature of its phonon-induced decoherence. Here, we introduce an algorithm that reduces the computational cost of the numerically exact process tensor method for non-Markovian dynamics simulations when the environmental coupling operator has degenerate eigenvalues, making calculations of the non-Markovian dynamics of large systems feasible. We compute the degree of entanglement of photon pairs generated by pulsed two-photon resonant excitation and find surprisingly good agreement between the numerically exact results and those calculated using the approximate polaron master equation method, permitting an efficient exploration of trends across system parameters.