Explicit counterexamples to Frank-Loss optimizer conjectures for critical Dirac and vector Sobolev inequalities

arXiv:2610.15787 · math.AP · Submitted 2026-10-11 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Mathematics. The day in mathematics: new theorems, new proofs, and what they open.

Ada: We're looking at the paper: "Explicit counterexamples to Frank-Loss optimizer conjectures for critical Dirac and vector Sobolev inequalities".

Paper summary: Ada: So, to get into more detail about this work on "Explicit counterexamples to Frank-Loss optimizer conjectures for critical Dirac and vector Sobolev inequalities", the core idea is that they are proving these conjectures wrong > They're showing concrete examples of competitors whose quotients are strictly smaller than the conjectured optimal values > This matters because it shows that those specific families, like the spinorial family or the higher-dimensional Loss–Yau potential, aren't actually optimizers for these inequalities >

Emmy: And they achieve this by constructing two elementary families of explicit competitors with strictly smaller quotients > They define a spinorial competitor for every dimension greater than or equal to two, which is a constant spinor multiplied by a scalar Aubin–Talenti optimizer > And for dimensions greater than or equal to three, the vector competitor is this planar azimuthal field >

Ada: That planar azimuthal field they compute its quotient in closed form and that's what gives them an explicit proof of nonoptimality for the higher-dimensional Loss–Yau–Dunne–Min potential in dimension three > It’s a very specific result about three dimensions, which is interesting because it connects to other work >

Kolya: From a probabilistic view, what this implies is that the tail behavior governed by these competitors doesn't match the expected optimal bound they were aiming for > If you're looking at applications where these inequalities show up, you can't just assume those specific functions are going to give you the best result >

Ada: Exactly. So, we have this comparison between the spinorial competitor and the Loss–Yau pair, showing that one is strictly smaller across every dimension greater than or equal to two > This suggests a broader structure where simpler choices might outperform more complex ones in these specific Sobolev problems >

Conclusion: Ada: Looking at the whole paper, "Explicit counterexamples to Frank-Loss optimizer conjectures for critical Dirac and vector Sobolev inequalities" by Bieganowski and Mederski, the authors are really pushing back on some long-standing ideas about what functions should be optimal > They're not just saying something isn't optimal; they are providing the actual functions that beat the conjectured ones >

Emmy: And the implication is that for these critical Dirac and vector Sobolev inequalities, we can’t rely on a single family of candidates—like just using spinorial functions or just using the Loss–Yau potential—to find a guaranteed optimum > You have to compare them against these explicit competitors they built >

Ada: The real point seems to be that the sharp threshold in this problem doesn't come from multiplying the optimal constants for Cs and Cv separately, which is what some strategies discussed in other papers might suggest > This means the way we combine separate inequalities matters a lot when you're trying to find the best overall bound >

Kolya: From my side, it just means that when we apply these inequalities in real physical or mathematical models, we have to be careful about which structure—the one they found explicitly or the one Frank and Loss conjectured—actually governs the limiting behavior > It forces us to check more concrete examples instead of relying on abstract structural arguments alone >

University of Warsaw · Institute of Mathematics, Polish Academy of Sciences · Nicolaus Copernicus University

math.AP

Submitted: 2026-10-11

Updated: 2026-11-16

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 92/100

The gist: Frank and Loss proposed critical Sobolev inequalities for the Dirac operator and for vector fields, and these results provide explicit counterexamples to their conjectures regarding optimality…

Key concepts

Frank-Loss Conjectures
These are conjectures proposed by Frank and Loss concerning the optimal constants for critical Sobolev inequalities related to the Dirac operator and vector fields. The paper demonstrates that these assumed optimal solutions are not actually optimizers in certain dimensions.
Spinorial Competitor
This competitor is constructed by taking a constant spinor and multiplying it by a scalar Aubin-Talenti optimizer. This function serves as an explicit example of a trial function that yields a quotient strictly smaller than the conjectured optimal value for every dimension d >= 2.
Planar Azimuthal Field
For dimensions d >= 3, this is the specific vector field used as a competitor. It lies within the nonlinear gauge and its quotient is computed exactly in closed form, providing a concrete proof that higher-dimensional Loss-Yau-Dunne-Min potentials are not optimal in three dimensions.

Terminology

Summary

Frank and Loss proposed critical Sobolev inequalities for the Dirac operator and for vector fields, and these results provide explicit counterexamples to their conjectures regarding optimality

Pruning redundant reasoning tokens

The paper presents explicit counterexamples to Frank–Loss optimizer conjectures for critical Dirac and vector Sobolev inequalities The authors compute closed-form quotients for two elementary families of explicit competitors with strictly smaller quotients For every odd d ≥ 3, the vector competitor is the planar azimuthal field This field gives an explicit proof of nonoptimality of the higher-dimensional Loss–Yau–Dunne–Min potential in dimension three

How the training signal is built

The authors define two elementary families of explicit competitors with strictly smaller quotients. For every d ≥ 2, the spinorial competitor is a constant spinor multiplied by a scalar Aubin–Talenti optimizer. p. 0

Where the paper says it

We compute closed-form quotients for two elementary families of explicit competitors with strictly smaller quotients. For every d ≥ 2, the spinorial competitor is a constant spinor multiplied by a scalar Aubin–Talenti optimizer.

p. 0
Where the paper says it

We compute closed-form quotients for two elementary families of explicit competitors with strictly smaller quotients. For every d ≥ 2, the spinorial competitor is a constant spinor multiplied by a scalar Aubin–Talenti optimizer.

For every d ≥ 3, the vector competitor is the planar azimuthal field This vector field lies in the nonlinear gauge and its quotient is computed in closed form. p. 0
Where the paper says it

For every d ≥ 3, the vector competitor is the planar azimuthal field Ad (x) = d (1 + |x| 2) 2 (−x2, x1 , 0, . It lies in the nonlinear gauge and its quotient is computed in closed form.

p. 0
Where the paper says it

For every d ≥ 3, the vector competitor is the planar azimuthal field Ad (x) = d (1 + |x| 2) 2 (−x2, x1 , 0, . It lies in the nonlinear gauge and its quotient is computed in closed form.

Where the gains come from

The spinorial trial function yields a strictly smaller quotient than the Loss–Yau pair. For every integer d ≥ 2, a direct comparison in [12, Proposition 5. p. 3

Where the paper says it

In [11, Proposition 5. 1] the authors give a direct comparison for the gauge quotient: combining positive and negative first curl eigenforms and applying Hölder’s inequality produces explicit competitors with smaller quotients.

5 and Appendix C] gives a strict competitor whose quotient lies below the common value of the conformal candidate family. p. 6
Where the paper says it

Their direct comparison in [12, Proposition 5. 5 and Appendix C] gives a strict competitor whose quotient lies below the common value of the conformal candidate family.

In dimension three, this field is a specific member of the competitor family of Wang and Zhang. p. 0
Where the paper says it

For every odd d ≥ 3, it gives an explicit proof of nonoptimality of the higher-dimensional Loss–Yau–Dunne–Min potential. In dimension three, this field is a specific member of the competitor family of Wang and Zhang: it is a linear combination of the Loss–Yau potential and its reflected twin.

Consequences and final remarks

The theorem provides quantitative upper bounds for Cs, Cv and S3/2,curl. For every odd dimension, no nonzero scalar multiple of a conformal pullback of the Loss–Yau–Dunne–Min potential is an optimizer of the Frank–Loss vector inequality. p. 0

Where the paper says it

For every odd d ≥ 3, it gives an explicit proof of nonoptimality of the higher-dimensional Loss–Yau–Dunne–Min potential. In dimension three, this field is a specific member of the competitor family of Wang and Zhang: it is a linear combination of the Loss–Yau potential and its reflected twin.

p. 0
Where the paper says it

For every odd d ≥ 3, it gives an explicit proof of nonoptimality of the higher-dimensional Loss–Yau–Dunne–Min potential. In dimension three, this field is a specific member of the competitor family of Wang and Zhang: it is a linear combination of the Loss–Yau potential and its reflected twin.

The estimates also imply CsCv ≤ 2 1440π 3/49 1/3 = 320 147π/3 S 3 2/3 the strategy discussed in [5, Remark 9. p. 1
Where the paper says it

5) ALY(x) = 3 (1 + |x| 2) 2 ((1 − |x| 2 )w + 2(x · w)x + 2w ∧ x), where η ∈ C2 is a constant spinor with |η| = 1 and w = ⟨η, ση⟩; in particular, |w| = 1. In Remark 9.

1] combines the two separate inequalities. p. 0
Where the paper says it

EXPLICIT COUNTEREXAMPLES TO FRANK–LOSS OPTIMIZER CONJECTURES FOR CRITICAL DIRAC AND VECTOR SOBOLEV INEQUALITIES BARTOSZ BIEGANOWSKI AND JAROSŁAW MEDERSKI ABSTRACT. Frank and Loss proposed critical Sobolev inequalities for the Dirac operator and for vector fields.

p. 0
Where the paper says it

EXPLICIT COUNTEREXAMPLES TO FRANK–LOSS OPTIMIZER CONJECTURES FOR CRITICAL DIRAC AND VECTOR SOBOLEV INEQUALITIES BARTOSZ BIEGANOWSKI AND JAROSŁAW MEDERSKI ABSTRACT. Frank and Loss proposed critical Sobolev inequalities for the Dirac operator and for vector fields.

The sharp threshold in this conformally compactifiable class does not arise by multiplying the separate optimal constants Cs and Cv. p. 0
Where the paper says it

INTRODUCTION AND MAIN RESULTS Let σ = (σ1, σ2, σ3), σ1 = 0 1 1 0 , σ2 = 0 −i i 0 , σ3 = 1 0 0 −1 , be the Pauli matrices. Frank and Loss considered the optimal constants in (1.

p. 0
Where the paper says it

INTRODUCTION AND MAIN RESULTS Let σ = (σ1, σ2, σ3), σ1 = 0 1 1 0 , σ2 = 0 −i i 0 , σ3 = 1 0 0 −1 , be the Pauli matrices. Frank and Loss considered the optimal constants in (1.

The paper presents explicit counterexamples to Frank–Loss optimizer conjectures for critical Dirac and vector Sobolev inequalities The authors compute closed-form quotients for two elementary families of explicit competitors with strictly smaller quotients. p. 0

Where the paper says it

Wang and Zhang recently disproved both optimizer claims using direct comparison and instability arguments. We compute closed-form quotients for two elementary families of explicit competitors with strictly smaller quotients.

For every odd d ≥ 3, the vector competitor is the planar azimuthal field. p. 0
Where the paper says it

For every d ≥ 2, the spinorial competitor is a constant spinor multiplied by a scalar Aubin–Talenti optimizer. For every d ≥ 3, the vector competitor is the planar azimuthal field Ad (x) = d (1 + |x| 2) 2 (−x2, x1 , 0, .

p. 0
Where the paper says it

For every d ≥ 2, the spinorial competitor is a constant spinor multiplied by a scalar Aubin–Talenti optimizer. For every d ≥ 3, the vector competitor is the planar azimuthal field Ad (x) = d (1 + |x| 2) 2 (−x2, x1 , 0, .

This field gives an explicit proof of nonoptimality of the higher-dimensional Loss–Yau–Dunne–Min potential in dimension three. p. 0
Where the paper says it

For every odd d ≥ 3, it gives an explicit proof of nonoptimality of the higher-dimensional Loss–Yau–Dunne–Min potential. In dimension three, this field is a specific member of the competitor family of Wang and Zhang: it is a linear combination of the Loss–Yau potential and its reflected twin.

p. 0
Where the paper says it

For every odd d ≥ 3, it gives an explicit proof of nonoptimality of the higher-dimensional Loss–Yau–Dunne–Min potential. In dimension three, this field is a specific member of the competitor family of Wang and Zhang: it is a linear combination of the Loss–Yau potential and its reflected twin.

The authors define two elementary families of explicit competitors with strictly smaller quotients. For every d ≥ 2, the spinorial competitor is a constant spinor multiplied by a scalar Aubin–Talenti optimizer. For every d ≥ 3, the vector competitor is the planar azimuthal field. p. 0

Where the paper says it

For every d ≥ 2, the spinorial competitor is a constant spinor multiplied by a scalar Aubin–Talenti optimizer. For every d ≥ 3, the vector competitor is the planar azimuthal field Ad (x) = d (1 + |x| 2) 2 (−x2, x1 , 0, .

This vector field lies in the nonlinear gauge and its quotient is computed in closed form.

The theorem provides quantitative upper bounds for Cs, Cv and S3/2,curl. For every odd dimension, no nonzero scalar multiple of a conformal pullback of the Loss–Yau–Dunne–Min potential is an optimizer of the Frank–Loss vector inequality. The estimates also imply CsCv ≤ 2 1440π 3/49 1/3 = 320 147π/3 S 3 > the strategy discussed in [5, Remark 9. p. 1

Where the paper says it

5) ALY(x) = 3 (1 + |x| 2) 2 ((1 − |x| 2 )w + 2(x · w)x + 2w ∧ x), where η ∈ C2 is a constant spinor with |η| = 1 and w = ⟨η, ση⟩; in particular, |w| = 1. In Remark 9.

1] combines the two separate inequalities. The sharp threshold in this conformally compactifiable class does not arise by multiplying the separate optimal constants Cs and Cv.

The paper presents explicit counterexamples to Frank–Loss optimizer conjectures for critical Dirac and vector Sobolev inequalities. The authors compute closed-form quotients for two elementary families of explicit competitors with strictly smaller quotients. For every odd d ≥ 3, the vector competitor is the planar azimuthal field. This field gives an explicit proof of nonoptimality of the higher-dimensional Loss–Yau–Dunne–Min potential in dimension three.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Explicit counterexample generation for critical Sobolev inequalities in Dirac and vector fields allows AI systems to rigorously test conjectures by providing explicit competitors with strictly smaller quotients, moving beyond mere qualitative disproofs like disproved both optimizer claims using direct comparison and instability arguments.

  2. Automated computation of closed-form quotients for elementary competitor families enables AI to determine exact numerical bounds for Sobolev constants, such as calculating the specific quotient (6.7) for the field AT d in dimension three, providing quantitative upper bounds like 240/491/3S32/3.

  3. The ability to analyze and compare different trial families, including spinorial competitors and vector-potential competitors, allows AI systems to systematically disprove conjectures across varying dimensions (e.g., disproves the higher-dimensional vector conjecture in every odd dimension) by constructing explicit counterexamples for each case.

  4. Implementation of gauge seminorm analysis allows AI to verify the equivalence between different formulations of inequalities, such as showing that the gauge formulation in (1.3) is equivalent to (1.2) and using this to establish constraints on minimization problems within specific classes like the axisymmetric azimuthal class T.

  5. Systematic verification of conformal invariance properties, such as the pullback below is the change-of-coordinates rule for a potential viewed as a one-form, enables AI systems to extend nonoptimality results from specific potentials (like Loss–Yau–Dunne–Min) to their entire conformal family.

  6. Automated computation of geometric quantities like the Jacobian determinant and its relationship to the norm of the exterior derivative allows AI to derive explicit pointwise identities, such as the pointwise identity nabla ALYDM,d 2 = 8d/2 (d - 1)s-4, which are critical for proving strict inequalities.

  7. The derivation of explicit bounds on Sobolev constants using trial functions like the constant-direction spinor ψd(x) provides a direct method for AI systems to find sharp upper bounds, such as S v d ≤ 2 d/2S d 1/2 bd √Gd, which are then compared against known thresholds.

Abstract

Frank and Loss proposed critical Sobolev inequalities for the Dirac operator and for vector fields. They conjectured optimality of their spinorial family and wrote that it is conceivable that the higher-dimensional Loss-Yau potential is an optimizer. Wang and Zhang recently disproved both optimizer claims using direct comparison and instability arguments. We compute closed-form quotients for two elementary families of explicit competitors with strictly smaller quotients. For every d at least2, the spinorial competitor is a constant spinor multiplied by a scalar Aubin-Talenti optimizer. For every d at least3, the vector competitor is the planar azimuthal field [ A d(x)= (1+ x 2) squared (-x 2,x 1,0,,0).] It lies in the nonlinear gauge and its quotient is computed in closed form. For every odd d at least3, it gives an explicit proof of nonoptimality of the higher-dimensional Loss-Yau-Dunne-Min potential. In dimension three, this field is a specific member of the competitor family of Wang and Zhang: it is a linear combination of the Loss-Yau potential and its reflected twin. Its axisymmetry gives a negative answer to their question about minimization within the SO(2) times 1-invariant class.

Sources

Related papers