Search papers, labs, and topics across Lattice.
This paper establishes that incorporating shared classical randomness into shallow unitary quantum Born models leads to a significant representational advantage over purely unitary models, allowing for the generation of long-range correlations that are otherwise unattainable. The authors demonstrate that this approach can be implemented using bounded-connectivity circuits augmented with local Pauli operations controlled by a single random bit, effectively expanding the family of distributions that can be generated. Their findings indicate that while a purely unitary model may require exponential depth to achieve similar performance, the proposed channel model can operate efficiently at shallow depths, thus addressing a previously open question in quantum generative modeling.
Shared classical randomness can unlock a vast array of distributions in shallow quantum generative models, outperforming purely unitary approaches that struggle with long-range correlations.
Near-term quantum hardware limits circuit depth and often imposes geometrically local connectivity for quantum generative models, restricting the output distributions accessible to shallow unitary Born models. Introducing stochasticity into a unitary quantum Born model can improve the empirical generative performance of the resulting channel model and, for a restricted small-scale architecture, has been proven to represent a strictly larger family of distributions than its unitary counterpart. However, whether such randomness provides a provable separation at fixed shallow depth for arbitrarily large systems has remained open. Here, we show that shared classical randomness, a comparatively weak resource from entanglement theory, is sufficient to establish such a strict scalable representational separation over the corresponding shallow unitary Born model. More specifically, we augment bounded-connectivity shallow unitary circuits, followed by computational-basis measurements, with spatially separated local Pauli operations, whose joint application is controlled by a single classically sampled random bit. The resulting shallow-depth channel model generates long-range correlations in the classical output distribution that no purely unitary shallow-depth model with bounded connectivity can reproduce. For one-dimensional nearest-neighbour architectures, reproducing such distributions with a purely unitary model can require depth $惟(N)$ in the worst case. We further show that measurement-based quantum computation (MBQC) provides a natural implementation of the required shared classical randomness through suitable adaptation of the random measurement outcomes. Numerical experiments on MBQC-based generative models support the analytical results.
