Breaking the One-Dimensional Expressibility-Trainability Tradeoff

• URL: http://arxiv.org/abs/2607.04598v1
• Date: Mon, 06 Jul 2026 02:01:59 GMT
• Title: Breaking the One-Dimensional Expressibility-Trainability Tradeoff
• Abstract summary: We show that the growth of expressibility and entangling power (EP) that improves Hilbert-space coverage is not a one-dimensional tradeoff.<n>EP measures how far the circuit has moved along the coverage dial, while EPD monitors whether input-dependent variability remains.<n>We find that ansatz routes can reach high, Haar-like coverage before EPD and gradient variance collapse.
• Score: 0.0
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: Expressive parameterized quantum circuits (PQCs) are often designed under a dilemma: the growth of expressibility and entangling power (EP) that improves Hilbert-space coverage is also expected to randomize an ansatz and activate barren-plateau (BP) conditions. We show that this dilemma is not a one-dimensional tradeoff. The usual picture collapses three inequivalent objects -- parameter-ensemble coverage, fixed-circuit entangling response, and local gradient moments -- into one scalar narrative. For a fixed circuit probed by Haar-product inputs, EP is a global two-copy mean of the output-entanglement distribution, whereas entangling-power deviation (EPD) is a global four-copy fluctuation descriptor. Gradient variance, however, is a local two-copy contraction selected by a parameter light cone and a cost observable. This moment hierarchy yields an analytic separation: equal EP need not imply equal trainability, as witnessed by equal-EP circuits with different EPDs and different gradient variances. These separations turn EP and EPD into a two-dial design rule for PQC ansatzes: EP measures how far the circuit has moved along the coverage dial, while EPD monitors whether input-dependent variability remains. We find that ansatz routes can reach high, Haar-like coverage before EPD and gradient variance collapse, showing that coverage and BP activation are distinct crossover events. The EP/EPD framework thus breaks the apparent one-dimensional expressibility-trainability tradeoff into a practical design rule: search for highly expressive PQCs in the window where coverage is high but BP-like homogenization has not yet erased trainable structure.

Related papers

• AdaFlash: Adaptive Speculative Decoding via On-Policy Distilled Diffusion Drafters [67.51886052453452]
  Speculative decoding has become a prevalent paradigm for accelerating large language model inference.<n>Recent work such as DFlash further boosts drafting efficiency by leveraging diffusion drafters.<n>We propose AdaFlash framework, comprising two components: (i) an on-policy distillation (OPD) algorithm with reverse-KL divergence tailored for diffusion drafters, and (ii) an adaptive length head that dynamically adjusts the candidate sequence length on the fly.
  arXiv  Detail & Related papers  (2026-07-21T15:52:50Z)
• Disentangling Continuous-Time Latent Dynamics: Identifiability of Latent SDEs via Diffusion Shifts [60.95873409673775]
  identifiability in continuous-time latent differential equation models remains largely open.<n>We study additive-noise latent SDEs observed through an unknown nonlinear diffeomorphism.<n>We propose a two-stage estimator for latent disentanglement and optional graph recovery.
  arXiv  Detail & Related papers  (2026-06-26T16:18:28Z)
• Geometry-Controlled Exceptional Points in Helicoidal Orbital-Angular-Momentum Photonics [0.0]
  Exceptional points (EPs) are non-Hermitian spectral singularities where eigenvalues and eigenvectors coalesce.<n>We propose a geometry-controlled route to EPs in orbital-angular-momentum (OAM) photonics.
  arXiv  Detail & Related papers  (2026-06-20T21:53:12Z)
• D-PACE: Dynamic Position-Aware Cross-Entropy for Parallel Speculative Drafting [59.204113363599994]
  Speculative decoding accelerates inference by having a small drafter propose tokens that a larger target model verifies in parallel.<n>Recent diffusion-based parallel drafters such as DFlash predict the full B-token block in one forward pass, enabling deeper drafters and longer accepted blocks.<n>We derive per-position training weights from a differentiable surrogate of expected accepted draft length, matching the weight of each position to its log-probability contribution.<n>Across six benchmarks, two Qwen3-4B draft depths, two decoding temperatures, and two additional target models, D-PACE consistently improves both wall-clock speedup and average
  arXiv  Detail & Related papers  (2026-05-12T06:27:57Z)
• On the Structural Non-Preservation of Epistemic Behaviour under Policy Transformation [51.56484100374058]
  We formalise such information-conditioned interaction patterns as behavioural dependency.<n>This induces a probe-relative notion of $$-behavioural equivalence and a within-policy behavioural distance.<n>Results identify structural conditions under which probe-conditioned behavioural separation is not preserved under common policy transformations.
  arXiv  Detail & Related papers  (2026-02-24T22:55:21Z)
• Entangling Power and Its Deviation: A Quantitative Analysis on Input-State Dependence and Variability in Entanglement Generation [0.0]
  Quantifying the entangling capability of quantum operations is a fundamental task in quantum information science.<n>EP alone cannot capture the intricate input-state-dependent nature of entanglement generation.<n>EPD provides a deeper and more complete characterization of the entangling structure.
  arXiv  Detail & Related papers  (2025-08-01T04:06:55Z)
• Topological transitions in quantum jump dynamics: Hidden exceptional points [45.58759752275849]
  Phenomena associated with exceptional points (EPs) and their applications have been extensively studied.<n>We consider a monitored three level system and find multiple EPs in the Lindbladian eigenvalues considered as functions of a counting field.<n>We demonstrate that these EPs signify transitions between different topological classes.
  arXiv  Detail & Related papers  (2024-08-09T18:00:02Z)
• Creating and controlling exceptional points of non-Hermitian Hamiltonians via homodyne Lindbladian invariance [0.0]
  The Exceptional Points (EPs) of non-Hermitian Hamiltonians (NHHs) are spectral degeneracies associated with coalescing eigenvalues and eigenvectors. These EPs can be generated experimentally in open quantum systems, evolving under a Lindblad equation. We exploit this mechanism to create and control EPs solely by changing the measurement we postselect on.
  arXiv  Detail & Related papers  (2022-06-03T15:35:08Z)
• KERPLE: Kernelized Relative Positional Embedding for Length Extrapolation [72.71398034617607]
  KERPLE is a framework that generalizes relative position embedding for extrapolation by kernelizing positional differences. The diversity of CPD kernels allows us to derive various RPEs that enable length extrapolation in a principled way.
  arXiv  Detail & Related papers  (2022-05-20T01:25:57Z)
• Transition from degeneracy to coalescence: theorem and applications [0.0]
  We present a theorem on the existence of transition between DP and EP for a general system. The obtained EP is robust to the strength of non-Hermitian terms. This finding reveals the ubiquitous connection between DP and EP.
  arXiv  Detail & Related papers  (2021-07-28T06:26:21Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.