Barren plateaus are typically characterized by vanishing gradients, yet the feasibility of curvature-based optimization fundamentally relies on the statistical resolvability of the Hessian matrix. In this work, we quantify the entrywise resolvability of the Hessian for Variational Quantum Algorithms at random initialization. By leveraging exact second-order parameter-shift rules, we derive a structural representation that reduces the variance of Hessian entries to a finite covariance quadratic form of shifted cost evaluations. This framework reveals two distinct scaling regimes that govern the sample complexity required to resolve Hessian entries against shot noise. For global objectives, we prove that Hessian variances are exponentially suppressed, implying that the number of measurement shots must scale as $O(e^{αn})$ with the number of qubits $n$ to maintain a constant signal-to-noise ratio. In contrast, for termwise local objectives in bounded-depth circuits, the variance decay is polynomial and explicitly controlled by the backward lightcone growth on the interaction graph, ensuring that curvature information remains statistically accessible with $O(\mathrm{poly}(n))$ shots. Extensive numerical experiments across varying system sizes and circuit depths demonstrate these theoretical bounds and the associated sampling costs. Our results provide a rigorous criterion for the computational tractability of second-order methods at initialization.
Barren plateaus are typically characterized by vanishing gradients, yet the feasibility of curvature-based optimization fundamentally relies on the statistical resolvability of the Hessian matrix. In this work, we quantify the entrywise resolvability of the Hessian for Variational Quantum Algorithms at random initialization. By leveraging exact second-order parameter-shift rules, we derive a structural representation that reduces the variance of Hessian entries to a finite covariance quadratic form of shifted cost evaluations. This framework reveals two distinct scaling regimes that govern the sample complexity required to resolve Hessian entries against shot noise. For global objectives, we prove that Hessian variances are exponentially suppressed, implying that the number of measurement shots must scale as O(eαn) with the number of qubits n to maintain a constant signal-to-noise ratio. In contrast, for termwise local objectives in bounded-depth circuits, the variance decay is polynomial and explicitly controlled by the backward lightcone growth on the interaction graph, ensuring that curvature information remains statistically accessible with O(poly(n)) shots. Extensive numerical experiments across varying system sizes and circuit depths demonstrate these theoretical bounds and the associated sampling costs. Our results provide a rigorous criterion for the computational tractability of second-order methods at initialization.
Sampling noisy intermediate-scale quantum devices is a fundamental step that converts coherent quantum-circuit outputs to measurement data for running variational quantum algorithms that utilize gradient and Hessian methods in cost-function optimization tasks. This step, however, introduces estimation errors in the resulting gradient or Hessian computations. To minimize these errors, we discuss tunable numerical estimators, which are the finite difference (including their generalized versions) and scaled parameter-shift estimators [introduced in Phys. Rev. A 103, 012405 (2021)], and propose operational circuit-averaged methods to optimize them. We show that these optimized numerical estimators offer estimation errors that drop exponentially with the number of circuit qubits for a given sampling-copy number, revealing a direct compatibility with the barren-plateau phenomenon. In particular, there exists a critical sampling-copy number below which an optimized difference estimator gives a smaller average estimation error in contrast to the standard (analytical) parameter-shift estimator, which exactly computes gradient and Hessian components. Moreover, this critical number grows exponentially with the circuit-qubit number. Finally, by forsaking analyticity, we demonstrate that the scaled parameter-shift estimators beat the standard unscaled ones in estimation accuracy under any situation, with comparable performances to those of the difference estimators within significant copy-number ranges, and are the best ones if larger copy numbers are affordable.
Barren plateaus-regions of exponentially small gradients-are a major obstacle for training variational quantum algorithms on non-trivial problem sizes. Existing results analyze specific causes in isolation, such as random deep circuits or global cost functions, but lack a unified picture that integrates circuit design, data encoding, cost structure, and noise. In this work, we develop a comprehensive theory of barren plateaus for parametrized quantum circuits. We prove that three mechanisms jointly control gradient scaling: (1) ansatz-induced concentration, where highly expressive or sufficiently deep circuits approximate unitary 2-designs causing gradient variance to scale as O(2 −n); (2) data-encoding-induced concentration, where certain feature maps wash out class-conditional structure; and (3) probleminduced concentration, where cost functions effectively average over exponentially many degrees of freedom. For each mechanism, we derive explicit scaling laws for the gradient variance as a function of qubit count n, depth L, and encoding complexity. We show that appropriately constrained expressibility, locality, and symmetry breaking can prevent gradients from vanishing exponentially and obtain conditions under which gradients remain at worst polynomially small. This yields architectural and initialization design principles for plateau-resistant circuits, including shallow, problem-informed ansätze, layer-wise training, and structured data encodings. Numerical experiments on up to 100 qubits for VQE, QAOA, and quantum classifiers confirm our theoretical predictions and demonstrate 10-100× improvements in effective gradient magnitudes over standard architectures. Our framework subsumes existing barren plateau results and provides a practical toolkit for diagnosing and mitigating trainability issues in variational quantum algorithms.
Variational quantum algorithms (VQAs) optimize the parameters θ of a parametrized quantum circuit V(θ) to minimize a cost function C. While VQAs may enable practical applications of noisy quantum computers, they are nevertheless heuristic methods with unproven scaling. Here, we rigorously prove two results, assuming V(θ) is an alternating layered ansatz composed of blocks forming local 2-designs. Our first result states that defining C in terms of global observables leads to exponentially vanishing gradients (i.e., barren plateaus) even when V(θ) is shallow. Hence, several VQAs in the literature must revise their proposed costs. On the other hand, our second result states that defining C with local observables leads to at worst a polynomially vanishing gradient, so long as the depth of V(θ) is [Formula: see text]. Our results establish a connection between locality and trainability. We illustrate these ideas with large-scale simulations, up to 100 qubits, of a quantum autoencoder implementation.
Variational quantum algorithms (VQAs) optimize the parameters θ of a quantum neural network V(θ) to minimize a cost function C. While VQAs may enable practical applications of noisy quantum computers, they are nevertheless heuristic methods with unproven scaling. Here, we rigorously prove two results, assuming V(θ) is a hardware-efficient ansatz composed of blocks forming local 2-designs. Our first result states that defining C in terms of global observables leads to an exponentially vanishing gradient (i.e., a barren plateau) even when V(θ) is shallow. This implies that several VQAs in the literature must revise their proposed cost functions. On the other hand, our second result states that defining C with local observables leads to at worst a polynomially vanishing gradient, so long as the depth of V(θ) is O(logn). Taken together, our results establish a connection between locality and trainability. Finally, we illustrate these ideas with large-scale simulations, up to 100 qubits, of a particular VQA known as quantum autoencoders.
The training of a parameterized model largely depends on the landscape of the underlying loss function. In particular, vanishing gradients are a central bottleneck in the scalability of variational quantum algorithms (VQAs), and are known to arise in various ways. However, a caveat of most existing gradient bound results is the requirement of t-design circuit assumptions that are typically not satisfied in practice. In this work, we loosen these assumptions altogether and derive tight upper and lower bounds on loss and gradient concentration for a large class of parameterized quantum circuits and arbitrary observables, which are significantly stronger than prior work. Moreover, we show that these bounds, as well as the variance of the loss itself, can be estimated efficiently and classically-providing practical tools to study the loss landscapes of VQA models, including verifying whether or not a circuit/observable induces barren plateaus. In particular, our results can readily be leveraged to rule out barren plateaus for a realistic class of ansätze and mixed observables, namely, observables containing a non-vanishing local term. This insight has direct implications for hybrid Quantum Generative Adversarial Networks (qGANs). We prove that designing the discriminator appropriately leads to 1-local weights that stay constant in the number of qubits, regardless of discriminator depth. This implies that qGANs with appropriately chosen generators do not suffer from barren plateaus even at scale-making them a promising candidate for applications in generative quantum machine learning. We demonstrate this result by training a qGAN to learn a 2D mixture of Gaussian distributions with up to 16 qubits, and provide numerical evidence that global contributions to the gradient, while initially exponentially small, may kick in substantially over the course of training.
Recent advances in the variational quantum eigensolver (VQE) have strengthened its promise for molecular electronic-structure calculations on near-term noisy intermediate-scale quantum (NISQ) devices. However, barren plateaus (BPs) remain a major obstacle to scaling VQE, as vanishing gradients can severely obstruct optimization at large scale. While most BP studies focus on randomly initialized circuits and global gradient behavior, practical quantum chemistry VQE typically uses chemically motivated initializations and local optimization around HF- or MP2-derived states. Although chemically informed initializations are widely expected to improve VQE trainability, how they mitigate local barren plateaus remains unclear. We therefore study the occurrence of local barren plateaus in molecular UCCSD-VQE, considering both chemically motivated initialization points and iterates along the corresponding optimization trajectories. In the benchmarks considered here, chemically initialized starting points and variational trajectories show polynomial decay of local gradient variance with system size instead of the exponential BP-like decay observed for random controls, indicating that chemical initialization mitigates barren plateaus in the local optimization regions and highlights the trainability advantage of chemically motivated VQE strategies.