Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

arXiv:2605.01319

Barren Plateaus as Destructive Interference: A Diagnostic Framework and Implications for Structured Ansatzes

Pilsung Kang

quant-phcs.LGSubmitted May 2, 2026• Announced May 5, 2026• 0 citations
v1 · May 2, 2026v2 · Aug 26, 2026
arXivHTMLPDF

Abstract

Barren plateaus (BPs) are usually described by the exponential suppression of gradient variance, but the mechanism by which gradient signal disappears remains unclear. We show that this phenomenon can be understood as destructive interference among termwise gradient contributions. To make this perspective operational, we introduce a diagnostic framework based on the cancellation ratio $R_k$, the effective term count $N_{\mathrm{eff},k}$, and the interference-quality measure $B_{\mathrm{eff},k}=R_k\sqrt{N_{\mathrm{eff},k}}$. Under a random-sign model, $B_{\mathrm{eff},k}$ remains near a stable baseline, defining a random-sign cancellation regime. For the transverse-field Ising model (TFIM), we find that the hardware-efficient ansatz (HEA) remains close to this regime across system sizes and depths, whereas the Hamiltonian variational ansatz (HVA) systematically escapes it. In particular, HVA exhibits larger $B_{\mathrm{eff},k}$ not merely because $N_{\mathrm{eff},k}$ is larger, but because $R_k$ also remains systematically larger despite the broader term participation. This pattern indicates improved sign organization rather than simple term suppression. We further establish an exact identity that connects the proposed interference diagnostics directly to the standard variance-based theory of BPs. These results position destructive interference as a mechanistic interpretation of BP-like behavior in the regimes studied here, but they do not imply that BPs and destructive interference are universally interchangeable across all architectures and settings.

Related papers

From Trainability Diagnostics to Optimization Claims: Boundaries and Controls in Variational Quantum Optimization

Pilsung Kang

Barren plateau diagnostics characterize whether gradient signal remains available for training, but surviving signal need not translate into successful optimization. We study this trainability--optimization gap at the level of optimizer steps. Treating coefficient-weighted Hamiltonian-term gradients as task-like components, we introduce step-level diagnostics and derive an exact bridge between signed termwise organization, directional activity, and first-order descent. Resolving this bridge into standard first-order geometry shows that the apparent organization--activity factors are not independent optimization axes and that, at fixed state and update norm, the raw gradient maximizes first-order descent of the summed objective. We compare vanilla gradient descent, a deterministic Hamiltonian-term PCGrad variant, and probe-gated LSO-PCGrad on transverse-field Ising model instances with hardware-efficient and Hamiltonian variational ansatzes, together with matched controls for update norm and probe budget. Blind projection can improve an organization diagnostic while worsening final energy and first-order predictability. After conditioning on standard first-order geometry, residual term-space composition shows no reproducible material incremental association with realized descent, while optimizer-relative update norm shows positive material associations in some settings without cross-regime reproducibility. Matched controls provide no resolved final-energy benefit attributable to the projected direction, and the improvement of LSO-PCGrad is more consistent with probe-based search and step-norm adaptation than with Hamiltonian-term projection itself. These results show that gradient-structure diagnostics can characterize trainability and update geometry without serving as standalone evidence of optimization benefit, which requires controls matched on update norm and search budget.

quant-pharXiv:2609.21243•Sep 21, 2026•0 citations

Hardware-efficient ansatz without barren plateaus in any depth

Chae-Yeun Park, Minhyeok Kang, Joonsuk Huh

Variational quantum circuits have recently gained much interest due to their relevance in real-world applications, such as combinatorial optimizations, quantum simulations, and modeling a probability distribution. Despite their huge potential, the practical usefulness of those circuits beyond tens of qubits is largely questioned. One of the major problems is the so-called barren plateaus phenomenon. Quantum circuits with a random structure often have a flat cost-function landscape and thus cannot be trained efficiently. In this paper, we propose two novel parameter conditions in which the hardware-efficient ansatz (HEA) is free from barren plateaus for arbitrary circuit depths. In the first condition, the HEA approximates to a time-evolution operator generated by a local Hamiltonian. Utilizing a recent result by [Park and Killoran, Quantum 8, 1239 (2024)], we prove a constant lower bound of gradient magnitudes in any depth both for local and global observables. On the other hand, the HEA is within the many-body localized (MBL) phase in the second parameter condition. We argue that the HEA in this phase has a large gradient component for a local observable using a phenomenological model for the MBL system. By initializing the parameters of the HEA using these conditions, we show that our findings offer better overall performance in solving many-body Hamiltonians. Our results indicate that barren plateaus are not an issue when initial parameters are smartly chosen, and other factors, such as local minima or the expressivity of the circuit, are more crucial.

arXiv:2403.04844•Mar 7, 2024•3 citations

H-EFT-VA: An Effective-Field-Theory Variational Ansatz with Provable Barren Plateau Avoidance

Eyad I. B Hamid

Variational Quantum Algorithms (VQAs) are critically threatened by the Barren Plateau (BP) phenomenon. In this work, we introduce the H-EFT Variational Ansatz (H-EFT-VA), an architecture inspired by Effective Field Theory (EFT). By enforcing a hierarchical "UV-cutoff" on initialization, we theoretically restrict the circuit's state exploration, preventing the formation of approximate unitary 2-designs. We provide a rigorous proof that this localization guarantees an inverse-polynomial lower bound on the gradient variance: $Var[\partialθ] \in Ω(1/poly(N))$. Crucially, unlike approaches that avoid BPs by limiting entanglement, we demonstrate that H-EFT-VA maintains volume-law entanglement and near-Haar purity, ensuring sufficient expressibility for complex quantum states. Extensive benchmarking across 16 experiments on the Transverse Field Ising Model confirms a 109x improvement in energy convergence and a 10.7x increase in ground-state fidelity over standard Hardware-Efficient Ansätze (HEA), with statistical significance of $p < 10^{-88}$. The static framework is most effective for Hamiltonians with moderate reference-state overlap; extension to systems with larger reference-state gaps is addressed through dynamic UV-cutoff relaxation strategies explored in concurrent work.

quant-pharXiv:2601.10479•Jan 16, 2026•0 citations

Forward–Inverse Fragmented Hardware-Efficient Ans¨atze Mitigate Barren Plateaus on Molecular Variational Quantum Eigensolver Benchmarks

Oguike Ahaneku

Hardware-efficient ans¨atze (HEA) for the variational quantum eigensolver (VQE) suffer barren plateaus (BP), in which gradient variance decays exponentially for molecular Hamiltonians at chemically relevant scales. The present work tests whether wrapping the HEA into forward–inverse layer pairs (“fragmentation”) removes this observed exponential-decay signature over the tested depth range. Under the canonical McClean et al. random-uniform protocol, plain HEA and an architecture-only fragmented variant are compared on H2, LiH, BeH2, stretched N2 in a complete active space [CAS(10,6): 10 electrons in 6 orbitals], and a reduced-active-space [Fe2S2(SH)4]2− model. On the four molecules in the BP regime, plain HEA exhibits the canonical exponential decay with BP exponent α ∈ [0.32, 0.41], whereas the fragmented variant shows no clear exponential trend (|α| ≲ 0.05) and retains gradient variance roughly two to four orders of magnitude above plain HEA at the deepest tested depths. A separate identity-initialized fragmented variant shows a distinct negative-slope profile, reflecting variance growth from a near-identity starting point rather than the canonical BP protocol. Short-descent tests on LiH and N2 show that unconstrained HEA trajectories can leave the target particle-number sector; thus any below-Hartree–Fock (HF) claim requires explicit sector verification. On H2, where no BP is present, plain HEA reaches the full configuration interaction (FCI) energy while fragmented variants remain trapped near HF. These results indicate that forward–inverse fragmentation targets the BP regime specifically.

•Apr 26, 2026•0 citations

Contiguous Depth, Not Nominal Depth, Controls Gradient Visibility in Fragmented Hardware-Efficient Variational Circuits

Oguike Ahaneku

The barren plateau (BP) phenomenon—exponential decay of gradient variance in parameterized quantum circuits (PQCs) with system size—is a primary obstacle to scalingvariational quantum algorithms (VQAs). This paper reports a set of empirical benchmarks on hardware-efficient ansatz (HEA) circuits up to n = 20 qubits, showing that identity fragmentation—partitioning a circuit of nominal depth D into pairs of forward and inverse blocks of size k—substantially increases the variance of cost-function gradients, relative to unfragmented HEA at the same nominal depth. The fragmented design exhibits a suppression exponent αfrag = 0.225 ± 0.008 in Var(∂θ C) ∼ e−αn, versus αunfrag = 0.711 ± 0.025 for the unfragmented baseline (R2 ≥ 0.996 in both cases), giving a measured ratio of approximately 4 × 104 at n = 20, D = 24. In the tested fragmented HEA settings, the advantage is monotone in the maximum contiguous depth Lcontig, persists under finite-shot estimation (8000 shots), and is partially robust to depolarizing noise at gate-error rates relevant to current superconducting hardware. A direct training-drift test over ten gradient-descent steps on an X-basis cost at n = 8 shows the fragmentation gradient-variance advantage persists (50× at t = 0, 40× at t = 10) while the fragmented circuit reduces the cost ∼ 50× more than the unfragmented baseline over the same window. The design rule is derived from the Universal Suppression Law in companion theoretical work; the present paper’s role is to establish benchmark-level empirical support across the circuit families and noise settings studied here.

•Apr 26, 2026•0 citations

Barren Plateaus Beyond Observable Concentration

Zi-Shen Li, Bujiao Wu, Xiao-Wei Li +1 more

Parameterized quantum circuits (PQCs) are central to quantum machine learning and near-term quantum simulation, but their scalability is often hindered by barren plateaus (BPs), where gradients decay exponentially with system size. Prior explanations, including expressivity, entanglement, locality, and noise, are often presented in ways that conflate two distinct issues: concentration of the measured observable and loss of parameter sensitivity caused by circuit dynamics.We develop a unified statistical framework that separates these mechanisms. We show that several standard BP explanations, including locality- and entanglement-related effects, can be understood through a single phenomenon that we term observable concentration (OC). Importantly, we prove that avoiding OC is necessary but not sufficient for trainability. Beyond OC, we identify two distinct mid-circuit sources of gradient suppression. First, in circuits with effectively independent forward and backward blocks, parameter perturbations can propagate outside the measurement light cone and become inaccessible to the final observable, yielding information-loss-induced BPs. Second, BPs can also arise in circuits with highly correlated forward and backward blocks, as we demonstrate through an echo-type circuit model that is reminiscent of information scrambling.

quant-pharXiv:2603.18479•Mar 20, 2026•0 citations
Barren plateaus (BPs) are usually described by the exponential suppression of gradient variance, but the mechanism by which gradient signal disappears remains unclear. We show that this phenomenon can be understood as destructive interference among termwise gradient contributions. To make this perspective operational, we introduce a diagnostic framework based on the cancellation ratio RkR_kRk​, the effective term count Neff,kN_{\mathrm{eff},k}Neff,k​, and the interference-quality measure Beff,k=RkNeff,kB_{\mathrm{eff},k}=R_k\sqrt{N_{\mathrm{eff},k}}Beff,k​=Rk​Neff,k​​. Under a random-sign model, Beff,kB_{\mathrm{eff},k}Beff,k​ remains near a stable baseline, defining a random-sign cancellation regime. For the transverse-field Ising model (TFIM), we find that the hardware-efficient ansatz (HEA) remains close to this regime across system sizes and depths, whereas the Hamiltonian variational ansatz (HVA) systematically escapes it. In particular, HVA exhibits larger Beff,kB_{\mathrm{eff},k}Beff,k​ not merely because Neff,kN_{\mathrm{eff},k}Neff,k​ is larger, but because RkR_kRk​ also remains systematically larger despite the broader term participation. This pattern indicates improved sign organization rather than simple term suppression. We further establish an exact identity that connects the proposed interference diagnostics directly to the standard variance-based theory of BPs. These results position destructive interference as a mechanistic interpretation of BP-like behavior in the regimes studied here, but they do not imply that BPs and destructive interference are universally interchangeable across all architectures and settings.
Barren plateau diagnostics characterize whether gradient signal remains available for training, but surviving signal need not translate into successful optimization. We study this trainability--optimization gap at the level of optimizer steps. Treating coefficient-weighted Hamiltonian-term gradients as task-like components, we introduce step-level diagnostics and derive an exact bridge between signed termwise organization, directional activity, and first-order descent. Resolving this bridge into standard first-order geometry shows that the apparent organization--activity factors are not independent optimization axes and that, at fixed state and update norm, the raw gradient maximizes first-order descent of the summed objective. We compare vanilla gradient descent, a deterministic Hamiltonian-term PCGrad variant, and probe-gated LSO-PCGrad on transverse-field Ising model instances with hardware-efficient and Hamiltonian variational ansatzes, together with matched controls for update norm and probe budget. Blind projection can improve an organization diagnostic while worsening final energy and first-order predictability. After conditioning on standard first-order geometry, residual term-space composition shows no reproducible material incremental association with realized descent, while optimizer-relative update norm shows positive material associations in some settings without cross-regime reproducibility. Matched controls provide no resolved final-energy benefit attributable to the projected direction, and the improvement of LSO-PCGrad is more consistent with probe-based search and step-norm adaptation than with Hamiltonian-term projection itself. These results show that gradient-structure diagnostics can characterize trainability and update geometry without serving as standalone evidence of optimization benefit, which requires controls matched on update norm and search budget.
Variational quantum circuits have recently gained much interest due to their relevance in real-world applications, such as combinatorial optimizations, quantum simulations, and modeling a probability distribution. Despite their huge potential, the practical usefulness of those circuits beyond tens of qubits is largely questioned. One of the major problems is the so-called barren plateaus phenomenon. Quantum circuits with a random structure often have a flat cost-function landscape and thus cannot be trained efficiently. In this paper, we propose two novel parameter conditions in which the hardware-efficient ansatz (HEA) is free from barren plateaus for arbitrary circuit depths. In the first condition, the HEA approximates to a time-evolution operator generated by a local Hamiltonian. Utilizing a recent result by [Park and Killoran, Quantum 8, 1239 (2024)], we prove a constant lower bound of gradient magnitudes in any depth both for local and global observables. On the other hand, the HEA is within the many-body localized (MBL) phase in the second parameter condition. We argue that the HEA in this phase has a large gradient component for a local observable using a phenomenological model for the MBL system. By initializing the parameters of the HEA using these conditions, we show that our findings offer better overall performance in solving many-body Hamiltonians. Our results indicate that barren plateaus are not an issue when initial parameters are smartly chosen, and other factors, such as local minima or the expressivity of the circuit, are more crucial.
Variational Quantum Algorithms (VQAs) are critically threatened by the Barren Plateau (BP) phenomenon. In this work, we introduce the H-EFT Variational Ansatz (H-EFT-VA), an architecture inspired by Effective Field Theory (EFT). By enforcing a hierarchical "UV-cutoff" on initialization, we theoretically restrict the circuit's state exploration, preventing the formation of approximate unitary 2-designs. We provide a rigorous proof that this localization guarantees an inverse-polynomial lower bound on the gradient variance: Var[∂θ]∈Ω(1/poly(N))Var[\partialθ] \in Ω(1/poly(N))Var[∂θ]∈Ω(1/poly(N)). Crucially, unlike approaches that avoid BPs by limiting entanglement, we demonstrate that H-EFT-VA maintains volume-law entanglement and near-Haar purity, ensuring sufficient expressibility for complex quantum states. Extensive benchmarking across 16 experiments on the Transverse Field Ising Model confirms a 109x improvement in energy convergence and a 10.7x increase in ground-state fidelity over standard Hardware-Efficient Ansätze (HEA), with statistical significance of p<10−88p < 10^{-88}p<10−88. The static framework is most effective for Hamiltonians with moderate reference-state overlap; extension to systems with larger reference-state gaps is addressed through dynamic UV-cutoff relaxation strategies explored in concurrent work.
Hardware-efficient ans¨atze (HEA) for the variational quantum eigensolver (VQE) suffer barren plateaus (BP), in which gradient variance decays exponentially for molecular Hamiltonians at chemically relevant scales. The present work tests whether wrapping the HEA into forward–inverse layer pairs (“fragmentation”) removes this observed exponential-decay signature over the tested depth range. Under the canonical McClean et al. random-uniform protocol, plain HEA and an architecture-only fragmented variant are compared on H2, LiH, BeH2, stretched N2 in a complete active space [CAS(10,6): 10 electrons in 6 orbitals], and a reduced-active-space [Fe2S2(SH)4]2− model. On the four molecules in the BP regime, plain HEA exhibits the canonical exponential decay with BP exponent α ∈ [0.32, 0.41], whereas the fragmented variant shows no clear exponential trend (|α| ≲ 0.05) and retains gradient variance roughly two to four orders of magnitude above plain HEA at the deepest tested depths. A separate identity-initialized fragmented variant shows a distinct negative-slope profile, reflecting variance growth from a near-identity starting point rather than the canonical BP protocol. Short-descent tests on LiH and N2 show that unconstrained HEA trajectories can leave the target particle-number sector; thus any below-Hartree–Fock (HF) claim requires explicit sector verification. On H2, where no BP is present, plain HEA reaches the full configuration interaction (FCI) energy while fragmented variants remain trapped near HF. These results indicate that forward–inverse fragmentation targets the BP regime specifically.
The barren plateau (BP) phenomenon—exponential decay of gradient variance in parameterized quantum circuits (PQCs) with system size—is a primary obstacle to scalingvariational quantum algorithms (VQAs). This paper reports a set of empirical benchmarks on hardware-efficient ansatz (HEA) circuits up to n = 20 qubits, showing that identity fragmentation—partitioning a circuit of nominal depth D into pairs of forward and inverse blocks of size k—substantially increases the variance of cost-function gradients, relative to unfragmented HEA at the same nominal depth. The fragmented design exhibits a suppression exponent αfrag = 0.225 ± 0.008 in Var(∂θ C) ∼ e−αn, versus αunfrag = 0.711 ± 0.025 for the unfragmented baseline (R2 ≥ 0.996 in both cases), giving a measured ratio of approximately 4 × 104 at n = 20, D = 24. In the tested fragmented HEA settings, the advantage is monotone in the maximum contiguous depth Lcontig, persists under finite-shot estimation (8000 shots), and is partially robust to depolarizing noise at gate-error rates relevant to current superconducting hardware. A direct training-drift test over ten gradient-descent steps on an X-basis cost at n = 8 shows the fragmentation gradient-variance advantage persists (50× at t = 0, 40× at t = 10) while the fragmented circuit reduces the cost ∼ 50× more than the unfragmented baseline over the same window. The design rule is derived from the Universal Suppression Law in companion theoretical work; the present paper’s role is to establish benchmark-level empirical support across the circuit families and noise settings studied here.
Parameterized quantum circuits (PQCs) are central to quantum machine learning and near-term quantum simulation, but their scalability is often hindered by barren plateaus (BPs), where gradients decay exponentially with system size. Prior explanations, including expressivity, entanglement, locality, and noise, are often presented in ways that conflate two distinct issues: concentration of the measured observable and loss of parameter sensitivity caused by circuit dynamics.We develop a unified statistical framework that separates these mechanisms. We show that several standard BP explanations, including locality- and entanglement-related effects, can be understood through a single phenomenon that we term observable concentration (OC). Importantly, we prove that avoiding OC is necessary but not sufficient for trainability. Beyond OC, we identify two distinct mid-circuit sources of gradient suppression. First, in circuits with effectively independent forward and backward blocks, parameter perturbations can propagate outside the measurement light cone and become inaccessible to the final observable, yielding information-loss-induced BPs. Second, BPs can also arise in circuits with highly correlated forward and backward blocks, as we demonstrate through an echo-type circuit model that is reminiscent of information scrambling.