pith. sign in

arxiv: 2605.20623 · v1 · pith:BVQDYCYNnew · submitted 2026-05-20 · 🧮 math.AP · cs.AI

Lower Bounds for Advection-Diffusion Equations: An Exploration with AI-Generated Proofs

Chenyang An , Xiaoqian Xu This is my paper

Pith reviewed 2026-05-21 04:24 UTC · model grok-4.3

classification 🧮 math.AP cs.AI
keywords advection-diffusionlower boundsmixing scaleshear flowsperiodic flowsAI proofsPDE estimatesQED system
0
0 comments X

The pith

Explicit lower bounds are established for advection-diffusion equations using AI-generated proofs.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper proves three specific lower bounds on the behavior of solutions to advection-diffusion equations for different types of velocity fields. For inviscid shears with limited transverse variation, the negative Sobolev norm decays at most polynomially. Diffusive shears keep a positive lower bound on the mixing scale. Rapidly oscillating periodic flows lead to exponential L2 decay. These bounds are fully explicit, and the proofs come from an AI multi-agent system without human edits, which matters for understanding the limits of mixing in fluids and the potential of AI in rigorous math.

Core claim

We establish explicit lower bounds for advection-diffusion equations in three settings: a polynomial Ḣ^{-1} bound for inviscid shears with u∈L^∞_t W^{1,1}_y, a uniform positive lower bound on the mixing scale for diffusive shears, and an exponential L^2 bound for rapidly oscillating time-periodic flows. All constants are explicit in the data. The proofs were generated entirely by a multi-agent math proving system, QED, without expert human intervention.

What carries the argument

The multi-agent math proving system QED, which generates the rigorous proofs for the lower bound estimates on the advection-diffusion equations.

Load-bearing premise

The multi-agent system QED produces fully rigorous, error-free proofs for these PDE estimates without any expert human intervention or post-generation correction.

What would settle it

Verification by independent experts finding an error in one of the AI-generated proofs, or construction of a counterexample flow where one of the claimed lower bounds fails to hold.

read the original abstract

We establish explicit lower bounds for advection-diffusion equations in three settings: a polynomial $\dot H^{-1}$ bound for inviscid shears with $u\in L^\infty_t W^{1,1}_y$, a uniform positive lower bound on the mixing scale for diffusive shears, and an exponential $L^2$ bound for rapidly oscillating time-periodic flows. All constants are explicit in the data. The proofs were generated entirely by a multi-agent math proving system, QED, without expert human intervention, serving as a test of AI's capability to produce rigorous mathematics.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript claims to establish three explicit lower bounds for advection-diffusion equations: a polynomial bound in the homogeneous Sobolev space Ḣ^{-1} for inviscid shear flows with velocity in L^∞_t W^{1,1}_y, a uniform positive lower bound on the mixing scale for diffusive shears, and an exponential L² lower bound for rapidly oscillating time-periodic flows. All constants are stated to be explicit in the data, and the proofs for these results are asserted to have been generated autonomously by the multi-agent system QED without any expert human intervention or post-generation correction.

Significance. If the QED-generated proofs are verified to be rigorous and free of gaps, the explicit constants in these lower bounds would strengthen existing results on mixing and dissipation rates in advection-diffusion equations, particularly by moving beyond non-quantitative or implicit estimates. The work also represents an attempt to benchmark AI systems on producing proofs in analysis, which could be of interest if accompanied by reproducible verification.

major comments (2)
  1. The central claim that QED produced fully rigorous, error-free proofs without human intervention is load-bearing for both the PDE results and the AI-capability test, yet the manuscript provides no formal verification (e.g., Lean/Coq export), independent expert audit, or even a high-level outline of the generated proof structures that would allow assessment of potential gaps in any of the three derivations.
  2. §2 (inviscid shear case): the polynomial Ḣ^{-1} lower bound is presented with explicit constants, but without reported error estimates, cross-checks against known limiting cases, or details on how QED handled the L^∞_t W^{1,1}_y regularity, it is impossible to confirm that the claimed bound does not inadvertently reduce to a tautology or rely on unstated assumptions.
minor comments (2)
  1. The description of the QED multi-agent architecture and its interaction protocol with the PDE problem statements could be expanded for reproducibility, including any prompt templates or termination criteria used.
  2. Notation for the mixing scale in the diffusive shear setting should be defined explicitly in the introduction or preliminaries to avoid ambiguity with standard definitions in the literature.

Simulated Author's Rebuttal

2 responses · 1 unresolved

We thank the referee for their thorough reading and insightful comments on our manuscript. We address each major comment point by point below, proposing revisions where they strengthen the presentation without altering the core claims. The work aims to explore AI-generated proofs in analysis while providing explicit bounds, and we value the feedback on verification and details.

read point-by-point responses

  1. Referee: The central claim that QED produced fully rigorous, error-free proofs without human intervention is load-bearing for both the PDE results and the AI-capability test, yet the manuscript provides no formal verification (e.g., Lean/Coq export), independent expert audit, or even a high-level outline of the generated proof structures that would allow assessment of potential gaps in any of the three derivations.

    Authors: We acknowledge that the manuscript does not include formal verification in a proof assistant or an independent audit, which would indeed help substantiate the no-intervention claim. The full proofs are included in the paper for direct reader inspection, as is standard for mathematical claims. To address the concern about assessing gaps, we will add a high-level outline of the key logical structures and steps from each QED-generated proof in a new section or appendix of the revised manuscript. This will allow better evaluation of the derivations while preserving the original generation process. We note that QED outputs human-readable proofs intended to be verifiable by experts, though machine-checkable export remains a direction for future development of the system. revision: partial

  2. Referee: §2 (inviscid shear case): the polynomial Ḣ^{-1} lower bound is presented with explicit constants, but without reported error estimates, cross-checks against known limiting cases, or details on how QED handled the L^∞_t W^{1,1}_y regularity, it is impossible to confirm that the claimed bound does not inadvertently reduce to a tautology or rely on unstated assumptions.

    Authors: We agree this additional information would improve clarity and verifiability. In the revised manuscript, we will expand §2 to include: (i) explicit error estimates derived alongside the main bound, (ii) cross-checks against known limiting cases (e.g., the pure diffusion case and stationary shear flows), and (iii) a brief description of how the QED system incorporated the L^∞_t W^{1,1}_y regularity assumption into the estimates without introducing unstated hypotheses. These additions confirm the bound is non-tautological and arises from the quantitative analysis of the advection-diffusion operator under the given regularity. revision: yes

standing simulated objections not resolved
  • Providing a formal Lean or Coq export or conducting a post-generation independent expert audit of the QED proofs, as these were outside the scope of the original autonomous generation process and would require separate development of the AI system.

Circularity Check

0 steps flagged

No significant circularity in mathematical derivations

full rationale

The paper presents explicit lower bounds for advection-diffusion equations across three settings, with proofs claimed to be autonomously generated by the QED multi-agent system. No equations, derivation steps, fitted parameters, or self-referential definitions appear in the provided text or abstract. The central claims do not reduce to inputs by construction, self-citation chains, or renamed empirical patterns; the AI-generation assertion is a methodological claim rather than a load-bearing mathematical loop. This is consistent with a self-contained presentation against external benchmarks, warranting only a minimal score for the unverified AI aspect.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Because only the abstract is available, the ledger cannot be populated with specific free parameters, axioms, or invented entities from the paper. The central claim implicitly rests on the unverified assumption that the QED system produces correct mathematics.

pith-pipeline@v0.9.0 · 5622 in / 1317 out tokens · 28788 ms · 2026-05-21T04:24:27.906602+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

26 extracted references · 26 canonical work pages

  1. [1]

    Qed: An open-source multi-agent system for generating mathematical proofs on open problems, 2026

    Chenyang An, Qihao Ye, Minghao Pan, and Jiayaun Zhang. Qed: An open-source multi-agent system for generating mathematical proofs on open problems, 2026

    work page 2026

  2. [2]

    G. K. Batchelor. Small-scale variation of convected quantities like temperature in turbulent fluid. I. General discussion and the case of small conductivity.J. Fluid Mech., 5:113–133, 1959

    work page 1959

  3. [3]

    Enhanced dissipation, hypoellipticity, and anomalous small noise inviscid limits in shear flows.Archive for Rational Mechanics and Analysis, 224(3):1161–1204, 2017

    Jacob Bedrossian and Michele Coti Zelati. Enhanced dissipation, hypoellipticity, and anomalous small noise inviscid limits in shear flows.Archive for Rational Mechanics and Analysis, 224(3):1161–1204, 2017

    work page 2017

  4. [4]

    Inviscid damping and enhanced dissipation of the boundary layer for 2d navier–stokes linearized around couette flow in a channel

    Jacob Bedrossian and Siming He. Inviscid damping and enhanced dissipation of the boundary layer for 2d navier–stokes linearized around couette flow in a channel. Communications in Mathematical Physics, 379(1):177–226, 2020

    work page 2020

  5. [5]

    Weak and parabolic solu- tions of advection–diffusion equations with rough velocity field.Journal of Evolution Equations, 24(1):1, 2024

    Paolo Bonicatto, Gennaro Ciampa, and Gianluca Crippa. Weak and parabolic solu- tions of advection–diffusion equations with rough velocity field.Journal of Evolution Equations, 24(1):1, 2024

    work page 2024

  6. [6]

    Mixing and diffusion for rough shear flows.arXiv preprint arXiv:2009.12268, 2020

    Maria Colombo, Michele Coti Zelati, and Klaus Widmayer. Mixing and diffusion for rough shear flows.arXiv preprint arXiv:2009.12268, 2020

    work page arXiv 2009

  7. [7]

    Diffusion and mixing in fluid flow.Annals of Mathematics, pages 643–674, 2008

    Peter Constantin, Alexander Kiselev, Lenya Ryzhik, and Andrej Zlatoˇ s. Diffusion and mixing in fluid flow.Annals of Mathematics, pages 643–674, 2008. 62 C. AN AND X. XU

    work page 2008

  8. [8]

    Cellular mixing with bounded palenstrophy

    Gianluca Crippa and Christian Schulze. Cellular mixing with bounded palenstrophy. Mathematical Models and Methods in Applied Sciences, 27(12):2297–2320, 2017

    work page 2017

  9. [9]

    The definition and measurement of some characteristics of mixtures

    PV Danckwerts. The definition and measurement of some characteristics of mixtures. Applied Scientific Research, Section A, 3(4):279–296, 1952

    work page 1952

  10. [10]

    Personal communications

    Hongjie Dong. Personal communications. 2017

    work page 2017

  11. [11]

    Dissipation enhancement by mixing.Nonlinearity, 32(5):1810, 2019

    Yuanyuan Feng and Gautam Iyer. Dissipation enhancement by mixing.Nonlinearity, 32(5):1810, 2019

    work page 2019

  12. [12]

    Mixing for generic rough shear flows.SIAM Journal on Mathematical Analysis, 55(6):7240–7272, 2023

    Lucio Galeati and Massimiliano Gubinelli. Mixing for generic rough shear flows.SIAM Journal on Mathematical Analysis, 55(6):7240–7272, 2023

    work page 2023

  13. [13]

    Personal com- munications

    Yupei Huang, Massimo Sorella, David L Villringer, and Xiaoqian Xu. Personal com- munications. 2025

    work page 2025

  14. [14]

    Exponential lower bounds for the advection-diffusion equation with shear flows.arXiv preprint arXiv:2511.14512, 2025

    Yupei Huang and Xiaoqian Xu. Exponential lower bounds for the advection-diffusion equation with shear flows.arXiv preprint arXiv:2511.14512, 2025

    work page arXiv 2025

  15. [15]

    Suppression of chemotactic explosion by mixing

    Alexander Kiselev and Xiaoqian Xu. Suppression of chemotactic explosion by mixing. Archive for Rational Mechanics and Analysis, 222(2):1077–1112, 2016

    work page 2016

  16. [16]

    A multiscale measure for mixing

    George Mathew, Igor Mezi´ c, and Linda Petzold. A multiscale measure for mixing. Physica D: Nonlinear Phenomena, 211(1-2):23–46, 2005

    work page 2005

  17. [17]

    Time-averaging under fast periodic forcing of parabolic par- tial differential equations: exponential estimates.Journal of Differential Equations, 174(1):133–180, 2001

    Karsten Matthies. Time-averaging under fast periodic forcing of parabolic par- tial differential equations: exponential estimates.Journal of Differential Equations, 174(1):133–180, 2001

    work page 2001

  18. [18]

    Diffusion-limited mixing by incompress- ible flows.Nonlinearity, 31(5):2346, 2018

    Christopher J Miles and Charles R Doering. Diffusion-limited mixing by incompress- ible flows.Nonlinearity, 31(5):2346, 2018

    work page 2018

  19. [19]

    Tracer microstructure in the large-eddy dominated regime.Chaos, Solitons & Fractals, 4(6):1091–1110, 1994

    Raymond T Pierrehumbert. Tracer microstructure in the large-eddy dominated regime.Chaos, Solitons & Fractals, 4(6):1091–1110, 1994

    work page 1994

  20. [20]

    Unique continuation for parabolic equations.Communications in Partial Differential Equations, 21(3-4):521–539, 1996

    Chi—Cheung Poon. Unique continuation for parabolic equations.Communications in Partial Differential Equations, 21(3-4):521–539, 1996

    work page 1996

  21. [21]

    Superexponential dissipation enhancement onT d.arXiv preprint arXiv:2509.02081, 2025

    Keefer Rowan. Superexponential dissipation enhancement onT d.arXiv preprint arXiv:2509.02081, 2025

    work page arXiv 2025

  22. [22]

    On completeness of root functions of sturm–liouville problems with discontinuous boundary operators.Journal of Differ- ential Equations, 255(10):3305–3337, 2013

    Alexander Shlapunov and Nikolai Tarkhanov. On completeness of root functions of sturm–liouville problems with discontinuous boundary operators.Journal of Differ- ential Equations, 255(10):3305–3337, 2013

    work page 2013

  23. [23]

    Using multiscale norms to quantify mixing and transport.Non- linearity, 25(2):R1, 2012

    Jean-Luc Thiffeault. Using multiscale norms to quantify mixing and transport.Non- linearity, 25(2):R1, 2012

    work page 2012

  24. [24]

    Diffusion and mixing in fluid flow via the resolvent estimate.Science China Mathematics, 64(3):507–518, 2021

    Dongyi Wei. Diffusion and mixing in fluid flow via the resolvent estimate.Science China Mathematics, 64(3):507–518, 2021

    work page 2021

  25. [25]

    Linear inviscid damping and enhanced dissipation for the Kolmogorov flow.Advances in Mathematics, 362:106963, 2020

    Dongyi Wei, Zhifei Zhang, and Weiren Zhao. Linear inviscid damping and enhanced dissipation for the Kolmogorov flow.Advances in Mathematics, 362:106963, 2020

    work page 2020

  26. [26]

    On the relation between enhanced dissipation timescales and mixing rates.Communications on Pure and Applied Mathematics, 73(6):1205–1244, 2020

    Michele Coti Zelati, Matias G Delgadino, and Tarek M Elgindi. On the relation between enhanced dissipation timescales and mixing rates.Communications on Pure and Applied Mathematics, 73(6):1205–1244, 2020. LOWER BOUNDS BY AI 63 Chenyang An: University of California, San Diego, La Jolla, United States Email address:cya.portfolio@gmail.com Xiaoqian Xu: Duke...

    work page 2020