Daily Summary for 2026-11-09
daily
Transcript
Introduction to the show: ident: Mathematics. The day in mathematics: new theorems, new proofs, and what they open.
Ada: It's the ninth of November, twenty twenty-six, and this is the day's research.
Emmy: 15 new papers came out today.
Ada: I'm Ada.
Emmy: And I'm Emmy, here with Kolya.
Ada: We'll take the day in one pass, then pull out the papers we're staying with.
Emmy: Start us off with Symplectic singularities in Relative Langlands duality for products of general linear groups.
The day's research: Ada: Today's focus is making optimization problems easier in Bures-Wasserstein space.
Emmy: That impacts Gaussian variational inference needing efficient solutions.
Kolya: We lift complex problems into Euclidean space of affine transport maps for that.
Ada: This allows accelerated proximal gradient methods to achieve a better objective gap of O k minus two after k iterations.
Emmy: That rate improvement is better than previous ones we saw in the paper An Accelerated Proximal Gradient Method for Bures-Wasserstein Space.
Kolya: The work on counting multigeodesics establishes asymptotic behavior for these paths on compact hyperbolic surfaces.
Ada: That relates directly to understanding spectral properties through its connection to the Selberg zeta function in Counting multigeodesics and the Selberg-Delange method.
Emmy: This paper shows fixed number of geodesics become surface-independent as length increases.
Kolya: However when number of geodesics grows with length it depends on an infinite Euler product over primitive closed geodesics in Counting multigeodesics and the Selberg-Delange method.
Ada: This moves beyond simple counting to characterize the structure of geodesic paths themselves.
Emmy: The distribution of a specific quantity related to multigeodesics approaches a Poisson random variable as length goes to infinity.
Kolya: That derivation comes from analyzing the generating function related to the Selberg zeta function in Counting multigeodesics and the Selberg-Delange method.
Ada: We study composite optimization over Bures--Wasserstein space of Gaussian measures and show an accelerated method achieves minimax-optimal objective gap.
Emmy: We prove that polar foliations of codimension two on compact simply connected irreducible symmetric spaces are hyperpolar in rank at least two.
Kolya: The relative Langlands dual of the T star N has symplectic singularities when G is a product of general linear groups.
Ada: We prove pointwise gradient bounds for proper weak k-convex functions in terms of a local average and a Wolff potential.
Emmy: We establish characterizations of h-dichotomies for noninvertible evolution families using notions of expansiveness and uniform noncriticality.
Kolya: We prove that every countably infinite group with a countable weight topology admits an exact factorization into two dense infinite subsets.
Ada: Every complete noncompact two-sided properly embedded constant mean curvature hypersurface of finite Morse index in Euclidean space is minimal.
Emmy: We establish almost sure bounds on Taylor coefficients of a function defined by a sum of random variables under certain conditions.
Kolya: We prove a large x asymptotic for the counting function of m-multigeodesics on a compact hyperbolic surface.
Ada: We create counterexamples showing that Pólya-Szegő type inequalities fail for weighted nonlocal seminorms in all three regimes.
Emmy: We characterize when a submodule of the Leavitt path algebra generated by an indicator function is simple.
Kolya: A family of bipartite graphs constructed shows that every sufficiently large real number is the norm of an idempotent Schur multiplier on bounded operators on l2.
Ada: We show that the sum of the Morse index and number of intersections of min-max geodesics achieving p-widths is bounded above by p.
Emmy: We show that any weighted sum of random coordinates under a geometrical constraint is asymptotically Gaussian in directions where no single weight dominates.
Kolya: In this note, we show that the sum of the Morse index and number of intersections of min-max geodesics achieving the p-widths of a closed surface are bounded above by p.
Lucky paper: 2611.08464: Ada: Welcome back to our session on the ninth of November, twenty twenty six, with a deep dive into Symplectic singularities in Relative Langlands duality for products of general linear groups.
Emmy: We're looking at this paper today and it connects some really deep areas of mathematics that are usually quite separate.
Kolya: I’ve been reading the setup, and it seems to start by recalling results from Braverman, Finkelberg, and Nakajima about the regular perverse sheaf on the affine Grassmannian of type A.
Ada: So they establish an isomorphism between that sheaf and a ring object associated with a Gelfand-Tsetlin quiver.
Emmy: And then Theorem one point one says for any finite-dimensional representation N of G, there is a G equivariant isomorphism between the Poisson schemes T star N V and MC of G V plus p star N.
Kolya: That leads directly to Corollary one point three which implies that for any finite-dimensional representation N of GL nj, the relative Langlands dual of T star N has symplectic singularities.
Ada: The authors then use that result, linking the affine variety WhH over G to the Coulomb branch associated with a larger group containing G as a subgroup.
Emmy: That identification is what gets them to conclude that T star N V has symplectic singularities because of a result from Bellamy about Coulomb branches having symplectic singularities.
Kolya: The construction of those affine varieties WhH over G is important because it plays a key role in functoriality for geometric Satake, which is described in the paper.
Ada: Theorem one point two establishes an isomorphism between MC of H times K and WhH over G using general facts about these structures involving H, K, and G.
Emmy: So the final consequence is that if H equals Qr i=one GL mi, then for any algebraic group homomorphism phi from H to G, the affine variety WhH over G has symplectic singularities.
Kolya: That result seems to be the main payoff here because it connects these deep representation theory concepts directly to a concrete geometric property like having symplectic singularities.
Ada: It’s interesting how they move from general group settings down to specifically products of general linear groups for this conclusion.
Emmy: I wonder how the construction of WhH over G itself behaves when we consider different types of groups H and K in that isomorphism.
Kolya: They define WhH over G as a product involving JH V, rho star JH V times G V, and cH V.
Ada: And they show an isomorphism between MC of H times K and this constructed affine variety WhH over G using the diagram involving H, K, and G.
Emmy: So the whole path is about showing that these geometric spaces have symplectic singularities when you use specific types of groups for H.
Kolya: That ties it back to the larger context of geometric Satake and functoriality they are trying to establish in this research area.
Lucky paper: 2611.08184: Ada: So we are looking at Gradient potential estimates for k-Hessian equations today.
Emmy: It’s from Peking University and Nankai University, and it focuses on pointwise gradient bounds for proper weak k-convex functions when two is less than or equal to k which is less than n.
Kolya: The main result establishes that for those conditions, the gradient of a proper weak k-convex function u at Lebesgue points of Du is bounded by a local average and a Wolff potential of its full Hessian measure Fku.
Ada: That estimate gives local Lipschitz regularity for Fku equal to f if f is in L n over k over one over k locally, and also for Fku equal to Dup when zero.
Emmy: And the second Lorentz index one over k and the upper bound on p are sharp in the full weak class as shown in that paper.
Kolya: The paper also derives local Lipschitz regularity for Fku equal to f if f is in L n over k over one over k locally based on mapping properties of potentials.
Ada: Then there's Theorem one point five which proves that for smooth solutions satisfying Fku equal to f in omega where f is less than or equal to LambdaDup plus h almost everywhere, every weak k-convex solution of Fku equal to Dup in omega is locally Lipschitz whenever p satisfies the inequality one point seven.
Emmy: But the paper also shows that the sharpness of those second Lorentz exponent and upper bound on p is demonstrated through specific constructions.
Kolya: The radial example in proposition three point one shows that for every two less than or equal to k less than n there is a radial function u in Phi k of B one and W one over p star of B one where Fku equals Dup star in B one, which is smooth and strictly k-convex on B one.
Ada: That sounds like a strong result because it shows that for every two less than or equal to k less than n you can find a function that is smooth and strictly k-convex.
Emmy: Proposition four point three provides averaged estimates stating that for BR of x in omega the mean oscillation of u minus l at x plus the mean oscillation of Du minus a is bounded by CP R alpha mu at x plus CR one over alpha A.
Kolya: That leads to corollary four point four which provides bounds on u minus l C zero alpha BR over four z and u minus l BMO BR over four z.
Ada: There's also a section showing the failure of gradient BMO bounds, proposition three point one shows there is a convex v in C one with bounded gradient whose potential WRk of Fkv is unbounded near zero for every fixed zero.
Emmy: That means the Wolff term in one point one cannot be replaced by I R k of Fku over k, where I R k mu at x equals hat R0 mu at Bt(x) t n minus k dt t.
Kolya: And the sequence in section four point two demonstrates why an interior estimate cannot be extended uniformly to the boundary using only the same measure data and the height of the solution.
Ada: So for k greater than or equal to two, gradient BMO can fail even under vanishing critical Morrey density.
Emmy: That’s a key limitation we need to keep in mind when thinking about applying these estimates to real-world problems at scale.
Kolya: We also have the declaration that they used ChatGPT and Codex from OpenAI to assist with mathematical exploration and drafting, specifically for some counterexample constructions and the design of the two point test function for proposition two point eight.
Ada: It’s good to see how those tools helped explore those specific difficult cases in constructing counterexamples.
Emmy: We need to remember that they took full responsibility for the mathematical content and the final manuscript itself.
Lucky paper: 2611.08171: Ada: Okay, moving on to this paper, it's about Dense Exact Factorizations of Countable Topological Groups.
Emmy: So we're looking at groups with a countable weight topology where every open set is infinite. What's the main thing they prove here?
Kolya: The main result is that any finite partial exact factorization extends to an exact factorization where both factors meet every member of a prescribed at most countable family of infinite subsets in infinitely many points.
Ada: That sounds complicated, so how does that relate to the structure of these groups?
Emmy: The proof relies on an auxiliary condition, condition R, which requires that for every choice of pairwise distinct elements h one through r not equal to e and arbitrary t one through r in G, the set G without the union of translates of fibers of the squaring map is infinite.
Kolya: That condition R is first proven for amenable groups and then generalized for all infinite groups, and lemma four point two shows a structural consequence that if every element x in G without F has at least one i such that xaix equals bi, then the FC of G is less than or equal to r squared.
Ada: So they tackle the difficulty of extending a partial factorization by ensuring this condition R holds for any infinite group.
Emmy: The proof for amenable groups uses invariant finitely additive probability measures on amenable groups, and lemma three point two demonstrates that in an amenable group, the measure of the set of square roots of a finite set S is less than one.
Kolya: Proposition three point three verifies condition R for amenable groups by showing that the union of translated fibers has measure less than one which is crucial for the inductive construction.
Ada: That's interesting because amenability seems to be the stepping stone, but they prove it holds for all infinite groups through theorem four point three, showing every group satisfies condition R.
Emmy: Then we have corollary five point two establishing an exact factorization into two dense countably infinite subsets for a group with a countable base topology where all nonempty open sets are infinite, and corollary five point three extends that to every countable nondiscrete Hausdorff left topological group of countable weight.
Kolya: The authors also show that they can choose the factors A and B such that each has an infinite intersection with every nonempty open set, which settles a question posed by Protasov and Slobodianiuk for the countable case.
Lucky paper: 2611.08454: Ada: So we're looking at the paper on On do Carmo's Question The Properly Embedded Case.
Emmy: It establishes that every complete noncompact two-sided properly embedded constant mean curvature hypersurface with a finite Morse index in Euclidean space is minimal.
Kolya: That theorem states that if you have a smooth, connected, complete, noncompact, two-sided hypersurface without boundary and with constant mean curvature and its Morse index is finite then the mean curvature has to be zero.
Ada: But the paper says this theorem doesn't impose requirements on things like the second fundamental form or how many ends it has at all.
Emmy: The authors say the noncompactness assumption is necessary because round spheres have nonzero mean curvature and a finite Morse index, and they also stress that the finite-index assumption is essential even within this properly embedded class.
Kolya: For nonzero mean curvature, the proof strategy involves using the radii of inscribed balls to define a positive weight which gives a polynomial bound on weighted volume.
Ada: And its reciprocal satisfies a differential inequality involving the Jacobi operator.
Emmy: They combine those properties with stability outside a compact set to get finite total area, which directly contradicts Frensel’s infinite-volume theorem, proving the mean curvature must be zero.
Kolya: The geometric tools they use involve defining a function mu of x as the supremum of a specific quantity involving distances to other points on the hypersurface.
Ada: And its reciprocal rho is defined as one over mu, which is essentially the radius of the largest tangent ball on the side into which N points.
Emmy: This geometry allows them to produce a weighted volume estimate, and they prove that mu satisfies the differential inequality delta mu plus A squared mu greater than or equal to n H mu squared in the sense of distributions on sigma.
Kolya: The proof proceeds by assuming H is greater than zero and constructing a locally Lipschitz function mu that has two properties.
Ada: What are those two properties they establish for this function mu?
Kolya: The first property is that mu satisfies the weighted stability inequality n H Z sigma mu cubed phi squared less than or equal to Z sigma mu squared gradient phi squared for every compactly supported Lipschitz function phi on sigma outside a compact set E.
Emmy: And the second property involves its reciprocal rho, which is defined as one over mu, satisfying a geometric volume estimate involving the integral of rho over R.
Kolya: Specifically, they show that Z sigma intersect BR rho is less than or equal to n plus one omega n plus one R plus H minus ten n plus one for every R greater than zero.
Ada: So combining those two things with exterior stability leads to finite total area, which then contradicts Frensel’s infinite-volume theorem, proving H equals zero.
Emmy: It's a strong result because it tackles do Carmo's question in the properly embedded case for hypersurfaces of constant mean curvature.
Kolya: The paper relies on references like Hilário Alencar and Manfredo P do Carmo’s work from one thousand nine hundred ninety-three and one thousand nine hundred ninety-five as well as some more recent papers on non-collapsing flows.
Ada: It really shows how the geometric constraints on the hypersurface force it to be minimal if its index is finite.
Emmy: We also see connections to other areas, like the work by Ben Andrews and Christine Breiner on mean curvature flow and embedded surfaces in Euclidean three-space.
Kolya: The key part for me is how they use the radii of inscribed balls to define that initial positive weight, which sets up the entire contradiction against Frensel's infinite volume theorem.
Ada: It’s a deep dive into geometric measure theory here, showing that finite index severely restricts the possible shape of these surfaces.
Emmy: This is significant because it gives us a concrete geometric condition for when constant mean curvature surfaces can exist in Euclidean space without being minimal.
Lucky paper: 2611.08165: Ada: So we are diving into Counting multigeodesics and the Selberg-Delange method today, Kolya.
Emmy: We're looking at how the paper Aapo Karvonen does things to count m-multigeodesics of length at most x on a compact hyperbolic surface X.
Kolya: The main result is that they prove a large x asymptotic for the counting function Pm(X, x).
Ada: That leads into the specific formula they give: Pm(X, x) equals e to x divided by x times log m minus one times nu zero X of m minus one over log x plus an error term.
Emmy: And that error term depends only on X because you can make the dependency explicit through asymptotic expansions in powers of one over x.
Kolya: That structure is really interesting because it shows how the counting function behaves as the length increases, which is what we expected from related number theory results.
Ada: But they also look at a different thing: the distribution of omega one in multigeodesics, and they show that as x goes to infinity, this distribution gets close to a Poisson random variable with parameter log x.
Emmy: That means for any epsilon greater than zero, the probability that omega one divided by log x minus one is greater than epsilon goes to zero as x goes to infinity.
Kolya: That convergence relies on them studying the generating function of omega one and relating it back to the Selberg zeta function.
Ada: I wonder how that relationship between the generating function and the Selberg zeta function actually drives that Poisson approximation result.
Emmy: It seems like they use Proposition four point nine and Proposition four point eight to prove the central limit theorem, which is what leads to that Poisson approximation for omega one divided by log x minus one.
Kolya: The leading order term for Pm(X, x) comes from considering the m-th coefficient of the Taylor series of FX x z, and they use Cauchy's integral formula there.
Ada: So they choose r equals m minus one times log x in that formula and use estimates from Proposition three point one and Theorem four point six to get that leading order contribution for Pm(X, x).
Emmy: That leading order term simplifies to nu zero of m minus one over log x times log m minus one times x times m minus one factorial divided by log x plus an error term involving exp C MX plus SX epsilon plus g sys X plus epsilon CMX.
Kolya: That dependence on the surface X in the error term is important because it shows how much the geometry of the surface affects that precision.
Ada: Thinking about what this means for someone just listening, it suggests that even though we're dealing with complex geodesic counting, there's a predictable statistical pattern emerging as you look at longer and longer paths.
Emmy: It moves beyond just saying how many paths exist to describing the underlying randomness of those paths themselves.
Kolya: This whole approach is inspired by the classical Selberg-Delange method in number theory, but they had to deal with the fact that the Selberg zeta function grows much faster than standard zeta functions on that critical strip.
Ada: So even without assuming the Riemann hypothesis, they managed to get a bound where zeta s equals o of t for s equal to sigma plus it with one over two less than or equal to sigma less than or equal to one.
Emmy: That is a strong result because it means the growth rate of that function is controlled even in regions where we don't have full information about its zeros.
Kolya: The final conclusion on the leading order term simplifies to nu zero of m minus one over log x times log log x times m minus one factorial as the main contribution.
Ada: So, looking at this paper Counting multigeodesics and the Selberg-Delange method, it’s a deep connection between geometry, number theory, and probability that helps us understand how things count on curved spaces.
More episodes
- 2509.05860-Concentration Inequalities for Branching Random Walks
- 2509.19309-Small Prime Primitive Roots in Arithmetic Progressions
- 2509.21830-Noncollapsing for Curvature Flows with Inhomogeneous Speeds
- 2510.08346-Separability and Submetrizability in Locally Convex Spaces
- 2510.27290-Borel Combinatorics of Schreier Graphs of Z-actions
- 2511.13500-Sampling Density for Gabor Phase Retrieval
- 2511.19161-Linear dynamics of random products of weighted shifts
- 2511.21548-Metastability of diffusion processes in narrow tubes
- 2512.01536-Abelian approximations for bicrossed products
- 2601.02524-Lagrangian slice disks with symplectomorphic exteriors