Information Theory · Cross-listed from cs.LG
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
Monday 6 April
Modern machine learning systems, such as generative models and recommendation systems, often evolve through a cycle of deployment, user interaction, and periodic model updates. This differs from standard supervised learning frameworks, which focus on loss or regret minimization over a fixed sequence of prediction tasks. Motivated by this setting, we revisit the classical model of learning from equivalence queries, introduced by Angluin (1988). In this model, a learner repeatedly proposes hypotheses and, when a deployed hypothesis is inadequate, receives a counterexample. Under fully adversarial counterexample generation, however, the model can be overly pessimistic. In addition, most prior work assumes a full-information setting, where the learner also observes the correct label of the counterexample, an assumption that is not always natural. We address these issues by restricting the environment to a broad class of less adversarial counterexample generators, which we call symmetric. Informally, such generators choose counterexamples based only on the symmetric difference between the hypothesis and the target. This class captures natural mechanisms such as random counterexamples (Angluin and Dohrn, 2017; Bhatia, 2021; Chase, Freitag, and Reyzin, 2024), as well as generators that return the simplest counterexample according to a prescribed complexity measure. Within this framework, we study learning from equivalence queries under both full-information and bandit feedback. We obtain tight bounds on the number of learning rounds in both settings and highlight directions for future work. Our analysis combines a game-theoretic view of symmetric adversaries with adaptive weighting methods and minimax arguments.
Group Theory
Characterizing finite solvable groups through the nilpotency probability
Given a finite group $G$, we denote by $\nu(G)$ the probability that two randomly chosen elements of $G$ generate a nilpotent subgroup. We prove that if $\nu(G)>1/12,$ then $G$ is solvable.
Analysis of PDEs
Low-regularity global well-posedness for the Boltzmann equation near vacuum
We study the Boltzmann equation near vacuum in anisotropic low-regularity Besov spaces. We establish the global existence and uniqueness of strong solutions with the critical regularity index $2/p$ for $p\in[1,\infty)$ in $\mathbb{R}^3$. The proof relies on a new bilinear estimate for the nonlinear collision operator. Combined with a div-curl type lemma we develop, this allows us to close the a priori estimates and thereby obtain global well-posedness.
Read the full abstract
Number Theory
Settled Elements in Arboreal Galois Groups of Quadratic PCF Polynomials
Let $f(x) \in K(x)$ be a quadratic polynomial where $K$ is a field of characteristic not equal to $2$. The associated arboreal Galois representation of the absolute Galois group of $K$ acts on a regular rooted binary tree. Boston and Jones conjectured that, for $f \in \mathbb{Z}[x]$, the image of this representation contains a dense set of settled elements. Roughly speaking, a cycle of an automorphism $\tau$ of the tree is called stable if its length strictly increases at each subsequent level, and $\tau$ is called settled if the proportion of vertices contained in stable cycles goes to $1$ as the level goes to infinity. In this article, we prove that the arithmetic iterated monodromy groups of postcritically finite quadratic polynomials in $K[x]$ with periodic postcritical orbits are densely settled. In the number field case, by a result of Benedetto–Ghioca–Juul–Tucker \cite{BGJT2025s}, it follows that for infinitely many $a \in K$, the associated arboreal Galois representations are densely settled. In particular, our results apply to the arithmetic IMG of the Basilica map $f(x)=x^2-1$.
Read the full abstract
Information Theory
Pinching Antenna Systems (PASS): Enabling Reconfigurable and Controllable Wireless Channels – A Comprehensive Survey
The evolution of wireless networks is driving new paradigms for consideration in upcoming generations. To this end, the 6G anticipates the development of several data-rate-hungry applications, in addition to a forecast growth in sensing-centric applications. Such an evolution, however, is unbalanced on the other side by the accentuated scarcity of spectrum, which opens up urgent needs to develop spectrum-efficient communication and sensing techniques. Due to the inability of the traditional multi-antenna schemes to enhance a wireless channel quality, increasing interest has been paid to wireless channel-altering schemes, such as reconfigurable intelligent surfaces and movable antennas. Recently, a new technique in this category, called pinching antennas (PAs), was introduced and tested. PA systems (PASS) are based on extending the reach of a base station by connecting its radio-frequency chains to long waveguides, on which one or many radiating antennas are pinched at custom positions of interest. Thus, such a technique can provide a means of overcoming several unfavorable channel conditions, such as the absence of a line-of-sight and increased free-space path loss. Importantly, such a channel-tuning feature can provide notable enhancements in terms of sensing, network coverage, data rate, and resilience against eavesdropping. In this work, we provide a comprehensive review of research on PASS, designed to meet various system design objectives, such as network coverage and data rate, information-theoretically secure transmission, sensing, integrated sensing and communication, and energy efficiency. A categorization of the surveyed work is established by comparing the various PASS schemes presented. Several takeaways are illustrated on the proposed schemes' potential and limitations, along with several directions forward discussed, in terms of future deployment and implementation.
Read the full abstract
Information Theory
Linear Exact Repair in MDS Array Codes: A General Lower Bound and Its Attainability
For an $(n,k,\ell)$ MDS array code over $\mathbb{F}_q$, how small can the repair bandwidth and repair I/O be under linear exact repair? We study this question in the regime where the field size $q$, the redundancy $r=n-k$, and the sub-packetization level $\ell$ are fixed, while the code length $n$ varies, and we develop a geometric approach to this setting. Our starting point is an intrinsic reformulation of linear exact repair for MDS array codes in terms of subspace intersections and, for repair I/O, the projective point configurations induced by a parity-check realization. This viewpoint yields a simple projective counting argument establishing the general lower bound $$\beta_{\mathrm{avg}},\beta_{\max},\gamma_{\mathrm{avg}},\gamma_{\max}\;\ge\;\ell(n-1)-\frac{q^{(r-1)\ell}-1}{q-1}$$ for linear exact repair of every $(n,k,\ell)$ MDS array code over $\mathbb{F}_q$ with redundancy $r=n-k\ge 2$. To our knowledge, this is the first lower bound of this form that applies to arbitrary redundancy $r\ge 2$ and sub-packetization level $\ell$. At first glance, the projective counting bound appears rather coarse and therefore unlikely to be attained. We prove that this intuition is correct whenever $r\ge 3$ and $\ell\ge 2$. For $r=2$, the picture changes completely. Using Desarguesian spreads from finite geometry, we construct MDS array codes that attain the bound over a broad interval of code lengths, up to the maximum possible length $q^{\ell}+1$, and do so simultaneously for both repair bandwidth and repair I/O. In the smallest nontrivial case $(r,\ell)=(2,2)$, we also prove a converse within the regular-spread model. Together, these results identify a uniform obstruction governing linear exact repair and show that, in the two-parity case, this obstruction is tight.
Read the full abstract
Number Theory
Joint extreme values of Dirichlet (L)-functions and their logarithmic derivatives
In this paper, we establish joint extreme values of Dirichlet (L)-functions and their logarithmic derivatives using the resonance method. Our results extend previous work of Aistleitner et al. (2019) and Yang (2023).
Representation Theory
Bicompact torsion classes and conjectures on brick infinite algebras
In the category $\operatorname{\mathsf{mod}} A$ of finitely generated modules over a finite dimensional algebra $A$, a torsion class $\mathcal{T} \subset \operatorname{\mathsf{mod}} A$ is said to be compact if there exists a module $M \in \operatorname{\mathsf{mod}} A$ such that $\mathcal{T}$ is the smallest torsion class containing $M$. If a torsion class satisfies this and the dual condition, then we call it a bicompact torsion class. We conjecture that bicompact torsion classes are precisely functorially finite torsion classes, and prove it for hereditary algebras and also for semistable torsion classes. This gives that Demonet's Lattice Point Conjecture implies Enomoto's Infinite Semibrick Conjecture, both of which are important conjectures on brick infiniteness.
Read the full abstract
Complex Variables
Weighted $L^2$ theory for the Euclidean Dirac operator in higher dimensions
We study weighted $L^{2}$ solvability for the Euclidean Dirac operator in dimensions $n\ge 3$. We prove that, on the exterior domain $\mathbb{R}^{n}\setminus\overline{B(0,1)}$ with logarithmic weight $\varphi=n\log|x|$, no higher-dimensional analogue of the two-dimensional Hörmander estimate can be controlled solely by $\Delta\varphi$; we then establish weighted solvability for the weights $|x|^{m}$ with $m\neq 0$, for the quadratic weight $x_{1}^{2}$, and for sufficiently small anisotropic perturbations of the Gaussian weight, with sharp constant $1/4$ in the Gaussian case. The obstruction arises because, in dimensions $n\ge 3$, the classical weighted identity is coercive only under a structural relation between $\Delta\varphi$ and $|\nabla\varphi|^{2}$, a condition that excludes the Gaussian weight and many polynomial weights. The method is based on a weighted identity for the conjugated unknown $U:=ue^{-\varphi/2}$, together with suitable scalar and Clifford-valued multipliers; this identity yields the required coercive estimates and also gives weighted $L^{2}$ solvability for the Poisson equation through the factorization $\Delta=-D^{2}$.
Read the full abstract
Optimization and Control · Cross-listed from eess.SY
Distributed Covariance Steering via Non-Convex ADMM for Large-Scale Multi-Agent Systems
This paper studies the problem of steering large-scale multi-agent stochastic linear systems between Gaussian distributions under probabilistic collision avoidance constraints. We introduce a family of distributed covariance steering (DCS) methods based on the Alternating Direction Method of Multipliers (ADMM), each offering different trade-offs between conservatism and computational efficiency. The first method, Full-Covariance-Consensus (FCC)-DCS, enforces consensus over both the means and covariances of neighboring agents, yielding the least conservative safe solutions. The second approach, Partial-Covariance-Consensus (PCC)-DCS, leverages the insight that safety can be maintained by exchanging only partial covariance information, reducing computational demands. The third method, Mean-Consensus (MC)-DCS, provides the most scalable alternative by requiring consensus only on mean states. Furthermore, we establish novel convergence guarantees for distributed ADMM with iteratively linearized non-convex constraints, covering a broad class of consensus optimization problems. This analysis proves convergence to stationary points for PCC-DCS and MC-DCS, while the convergence of FCC-DCS follows from standard ADMM theory. Simulations in 2D and 3D multi-agent environments verify safety, illustrate the trade-offs between methods, and demonstrate scalability to thousands of agents.
Read the full abstract
Category Theory
Kleisli semantics and hypergraph composition for Greimasian narrative programs
This article proposes a category-theoretic formalization of Greimasian narrative programs (NPs) that makes their compositional structure mathematically precise. Building on a reconstruction of the actantial model as a categorical schema, we introduce a refined typological schema of actants and derive Set-valued instances corresponding to role-indexed elements of a narrative. NPs are represented within a categorical schema whose morphisms are interpreted using monads on Set. In particular, the List monad provides a Kleisli semantics for modeling non-atomic, list-valued actantial configurations, while the Maybe monad encodes optional dependencies between programs. This yields a minimal representation of narrative programs as structured data with an intrinsic compositional interpretation. To account for the dynamics of narrative formation, we lift these constructions into a diagrammatic setting by freely generating a symmetric monoidal category, and subsequently a hypergraph category, from the set of actants. In this framework, narrative programs act as generators of morphisms, and their composition is realized through wiring diagrams. A narrative trajectory is thereby interpreted as a single composite morphism. This approach provides a unified mathematical framework for structural semiotics, connecting data-level representations of narrative elements with their compositional realization in discourse.
Read the full abstract
Logic
An effective version of the Stone duality
The paper studies computability-theoretic aspects of topological $T_0$-spaces. We introduce effective versions of the notions of a countable $c$-poset and a (second-countable) topological space with base. Based on this, we prove an effective version of the known Stone-type duality between the category $\mathbf{AS}$ (whose objects are almost semispectral spaces with base and whose morphisms are spectral mappings) and the category $\mathbf{DP}$ (whose objects are distributive $c$-posets and whose morphisms are strict mappings). Namely, we show that for an arbitrary set $Z\subseteq \omega$, this duality is preserved when one restricts to objects which have $Z$-computably enumerable presentations only. Following this approach, we establish several results in computable topology. We prove that every degree spectrum of a countable algebraic structure can be realized as the degree spectrum of a topological space with base. We show that for any non-zero natural number $N$, there is a computable topological space with base that has precisely $N$-many computable copies, up to effective spectral homeomorphisms.
Read the full abstract
Combinatorics
Immanantal polynomials of the linear combination matrices of graphs
The immanant is a generalized matrix function associated with irreducible characters of the symmetric group. The immanantal polynomial of matrix $M$ is defined as the immanant of matrix $(xI - M)$, where $I$ is the identity matrix. The characteristic polynomials and permanental polynomials of graph matrices (adjacency matrix, Laplacian matrix, signless Laplacian matrix and $A_{\alpha}(G)$ matrix) have been extensively studied, yielding fascinating results. As a unification of the characteristic polynomial and permanental polynomial, the immanantal polynomial has also attracted attention and led to several important findings. In this paper, we focus on the study of immanantal polynomials for linear combination matrices composed of the degree matrix and adjacency matrix of a graph. First, applying the concept of vertex orientation for general graphs, we provide a combinatorial interpretation of the coefficients of the immanantal polynomials for the linear combination matrices of graphs, and we also characterize the bounds of these coefficients. These bounds implicitly encompass the existing results of Chan and Lam on trees and bipartite graphs. Furthermore, we give a solution to the open problem posed by Merris. Second, we characterize the first six coefficients of the hook immanantal polynomial. And the necessary and sufficient condition under which the linear combination matrices of two regular graphs have the same hook immanantal polynomial is proved. Third, we generalize the Frobenius–König theorem and the Laplace expansion theorem to immanants. Using these two theorems, we show that the star degree of a graph is always a lower bound for the multiplicity of a certain root of the immanantal polynomial of its linear combination matrix. Finally, we derive formulas for the first six coefficients of the hook immanantal polynomial for several important graph matrices.
Read the full abstract
Information Theory
Beyond-Diagonal RIS For Enhanced Secrecy and Sensing Gains in Secure ISAC Networks: An Optimization Framework
Integrated sensing and communication (ISAC) has been receiving a notable interest as an energy- and spectrum-efficient enabler for simultaneous communication and sensing. Notably, reconfigurable intelligent surfaces (RIS) is among the key technologies enabling robust communication and sensing, particularly in environments without a line-of-sight (LoS). Recently, a new type of RIS, called beyond-diagonal RIS (BD-RIS), has drawn attention, offering additional degrees of freedom in controlling the propagation medium. In this paper, a novel secure BD-RIS-aided ISAC scheme is proposed and evaluated. The scheme is applicable to a multi-user multi-target ISAC network, where a dual-functional radar-communication (DFRC) base station (BS) simultaneously serves multiple downlink users and senses various targets that aim to eavesdrop on the legitimate signal transmitted to the users. The presence of a BD-RIS enables circumventing the absence of the LoS link and ensures secure transmission and sensing. To this end, an optimization problem is formulated aiming at maximizing a weighted sum of per-target reflected powers, subject to secrecy and transmit power constraints. Thus, by virtue of an Augmented Lagrangian- and Riemannian conjugate gradient-based approach, in addition to semidefinite programming, an alternating optimization (AO)-based algorithm is developed, which provides a local optimum for the BD-RIS scattering matrix, transmit signal beamforming matrices, and artificial noise covariance matrix. Numerical results highlight (i) the notable sensing gains of the BD-RIS-aided design with respect to its diagonal RIS (D-RIS)-based baseline and (ii) the improved secrecy-sensing trade-off, whereby the BD-RIS can ensure an increasing system secrecy without a significant loss in the per-target reflected power.
Read the full abstract
Optimization and Control
Viscosity Solutions of Hamilton–Jacobi–Bellman Equations for Control Systems Driven by Teugels Martingales
This paper studies discrete-time two-person nonzero-sum linear quadratic stochastic games with random coefficients. Using convex variational analysis, we derive necessary and sufficient conditions for the existence of open-loop Nash equilibria. When weighting matrices are indefinite, the classical first-order conditions are no longer sufficient for optimality; we introduce a global nonnegativity condition to restore sufficiency, which becomes a cornerstone of the subsequent analysis. To characterize the equilibria explicitly, we develop fully coupled forward-backward stochastic difference equations and a system of non-symmetric stochastic Riccati equations (FBS$\Delta$Es) with constraints. that decouple the stochastic Hamiltonian system. A key technical contribution is the provision of sufficient conditions – positive semidefiniteness of the Riccati matrices operators and structural non-degeneracy – that guarantee the invertibility of a related operator, ensuring the well-posedness of the closed-loop feedback representation of the open-loop Nash equilibrium strategies. A distinctive feature of this work is the presence of fully random coefficients, which leads to fully nonlinear higher-order backward stochastic difference equations in the Riccati framework, in contrast to the algebraic Riccati equations in the deterministic setting.
Read the full abstract
Quantum Algebra
Cancellation-free version of the quantum $K$-theoretic divisor axiom for the flag manifold in the quasi-minuscule case
We prove a cancellation-free version of the quantum $K$-theoretic divisor axiom for the flag manifold in the quasi-minuscule case under some assumptions. Namely, we remove the cancellations from the quantum $K$-theoretic divisor axiom due to Lenart-Naito-Sagaki-Xu in the case where the fundamental weight corresponding to the divisor class is quasi-minuscule.
Read the full abstract
Algebraic Geometry
Resolutions and deformations of cyclic quotient surface singularities
In this paper, we investigate the relations among various results concerning the minimal resolution of cyclic quotient singularities of the form $\mathbb{C}^2/G$. We refer to these as "bamboo-type" singularities, since the dual graphs of the exceptional curves in their resolutions resemble the shape of bamboo. We present classical results on the minimal resolution of singularities, the $G$-Hilbert scheme, the generalized McKay correspondence, deformations of singularities, and quiver varieties. These results have been obtained independently in different contexts, and here we provide a unified exposition enriched with numerous examples, which we hope will serve as a useful guide to the study of two-dimensional cyclic singularities. Moreover, this survey aims to offer insights that may inspire generalizations to non-cyclic singularities and to higher-dimensional quotient singularities.
Read the full abstract
Classical Analysis and ODEs
From hyperbolic to complex Euler integrals
Hyperbolic hypergeometric integrals are defined as Barnes-type integrals of products of hyperbolic gamma functions. Their reduction to ordinary hypergeometric functions is well known. We study in detail their degeneration to complex hypergeometric functions. Namely, using uniform bounds on the integrands, we prove that the univariate hyperbolic beta integral and the conical function degenerate to two-dimensional integrals over the complex plane.
Read the full abstract
Representation Theory
Degenerations of the fourth order $q$-Garnier system arising from confluences in quivers
The $q$-Garnier system was first proposed by Sakai, and alternative directions for its discrete time evolutions were provided by Nagao and Yamada. Recently, it was shown that all of those evolution directions are derived from a birational representation of an extended affine Weyl group which arises from the cluster algebraic construction established by Masuda, Okubo and Tsuda. In this article, we investigate a degeneration structure of the fourth order $q$-Garnier system by using confluences in quivers.
Read the full abstract
Mathematical Physics
Ground states and droplet regimes of the extended Gross-Pitaevskii equation with Lee-Huang-Yang correction
We study the ground states of the extended Gross–Pitaevskii equation with the Lee–Huang–Yang correction from both theoretical and numerical perspectives. Starting from the three-dimensional model, we derive reduced one- and two-dimensional equations through nondimensionalization and dimensional reduction. We establish existence and nonexistence results for ground states in different spatial dimensions, both in free space and under confining external potentials. For the numerical computation of ground states, we propose a normalized gradient flow method with a Lagrange multiplier. The numerical results show how the model parameters affect the ground-state profiles, and reveal different regimes in the free-space parameter plane, including no-ground-state, soliton-like, and droplet-like regions. We also introduce a simple flat-top approximation for the droplet regime and present two- and three-dimensional computations to illustrate more general localized structures.
Read the full abstract
Algebraic Geometry
Stokes structure of wild difference modules
We formulate and prove a Riemann–Hilbert correspondence between two categories: wild difference modules and wild Stokes-filtered $\mathscr{A}_{\rm{per}}$-modules. This correspondence is motivated by the Riemann–Hilbert correspondence for germs of meromorphic connections in one variable due to Deligne–Malgrange. It also generalizes the Riemann–Hilbert correspondence for mild difference modules.
Read the full abstract
Optimization and Control
An Alternating Primal Heuristic for Nonconvex MIQCQP with Dynamic Convexification and Parallel Local Branching
We develop a novel primal heuristic for nonconvex Mixed-Integer Quadratically Constrained Quadratic Programs (MIQCQPs). The method is built around a convex approximation that is dynamically adjusted within a feasibility-pump-style alternating heuristic. Approximations are adjusted based on the structure of the MIQCQP instance. Additionally, parallelized local branching is incorporated to further refine detected solutions. This paper builds upon the second-place finalist submission in the 2025 Land-Doig MIP Computational Competition. Our results are validated with computational experiments on instances from QPLIB, finding feasible solutions for three previously unsolved cases and improving the best-known solutions for fifteen instances within five minutes of runtime.
Read the full abstract
Analysis of PDEs
Rigidity for a semilinear Neumann problem with exponential nonlinearity in the large diffusion limit
We consider a semilinear Neumann problem with exponential nonlinearity in a smooth bounded domain $\Omega \subset \mathbb{R}^2$. We prove that there exists a threshold $\bar{\varepsilon}>0$ such that for all $\varepsilon>\bar{\varepsilon}$, any classical solution must be constant. This result provides a positive answer to a conjecture recently posed by Calanchi, Ciraolo, and Messina (2026). Our proof relies on a combination of $L^1$-estimates, a Jensen-type argument via the Neumann Green's function to obtain uniform exponential integrability, and elliptic regularity.
Read the full abstract
Operator Algebras
A universal property for groupoid C*-algebras. II. Fell bundles
We define possibly unsaturated, upper semicontinuous Fell bundles over Hausdorff, locally compact groupoids and establish a universal property for representations of their full section C*-algebras on Hilbert modules over arbitrary C*-algebras. Based on this, we prove that the full section C*-algebra is functorial and exact, and we define a quasi-orbit space and a quasi-orbit map. We deduce and extend Renault's Integration and Disintegration Theorems to general Fell bundles using our universal property.
Read the full abstract
Quantum Algebra
Quantum Borcherds-Bozec Superalgebras
We introduce quantum Borcherds-Bozec superalgebras. We present and prove various results of the quantum superalgebras including a bilinear form, higher Serre relation, quasi-R-matrix, character formula for the irreducible highest weight modules. We also prove the category of integrable representations is semi-simple.
Read the full abstract
Analysis of PDEs
On the Viscosity Solutions of Parabolic p-Laplacian Equations with Capillary-Type Boundary Conditions
In this paper, we establish the well-posedness and large-time asymptotic behavior of viscosity solutions to singular/degenerate parabolic $p$-Laplacian equations with general capillary-type boundary conditions, including Neumann and prescribed contact angle cases, on strictly convex domains. By establishing a gradient estimate independent of the $C^0$ norm of the solution via the maximum principle, and by analyzing the problem through an approximation procedure together with associated elliptic eigenvalue problems, we prove the existence, uniqueness, and asymptotic behavior of solutions. For the elliptic problem with Neumann boundary conditions, we first focus on flat domains with the zero Neumann condition. By reflecting $u$ across the flat boundary $T_1$ and then using inf- and sup-convolution arguments in the reflected domain, we obtain the $C^{1,\alpha}$ result. For the general elliptic case, we obtain sharp global $C^{1,\alpha}$ regularity by flattening the boundary and employing compactness arguments together with an “improvement of flatness” iteration. With an extra condition in the iteration, we can also deal with the singular case $1<p<2$. In the parabolic setting, the spatial Hölder regularity of $Du$ follows from elliptic estimates combined with the Lipschitz continuity of $u$ in time, which in turn yields joint Hölder continuity in $(x,t)$. Extensions to non-convex domains are also discussed by incorporating a suitable forcing term.
Read the full abstract
Optimization and Control · Cross-listed from cs.AI
Optimizing Service Operations via LLM-Powered Multi-Agent Simulation
Service system performance depends on how participants respond to design choices, but modeling these responses is hard due to the complexity of human behavior. We introduce an LLM-powered multi-agent simulation (LLM-MAS) framework for optimizing service operations. We pose the problem as stochastic optimization with decision-dependent uncertainty: design choices are embedded in prompts and shape the distribution of outcomes from interacting LLM-powered agents. By embedding key numerical information in prompts and extracting it from LLM-generated text, we model this uncertainty as a controlled Markov chain. We develop an on-trajectory learning algorithm that, on a single simulation run, simultaneously constructs zeroth-order gradient estimates and updates design parameters to optimize steady-state performance. We also incorporate variance reduction techniques. In a sustainable supply chain application, our method outperforms benchmarks, including blackbox optimization and using LLMs as numerical solvers or as role-playing system designers. A case study on optimal contest design with real behavioral data shows that LLM-MAS is both as a cost-effective evaluator of known designs and an exploratory tool that can uncover strong designs overlooked by traditional approaches.
Read the full abstract
Representation Theory
Relativistic Toda lattice of type B and quantum $K$-theory of type C flag variety
We introduce a classical integrable system associated with the torus-equivariant quantum $K$-theory of type C flag variety. We prove that its conserved quantities coincide with the generators of the defining ideal of the Borel presentation of the quantum $K$-ring obtained by Kouno and Naito. In particular, the Hamiltonian of the system is naturally regarded as a type B analogue of the relativistic Toda lattice introduced by Ruijsenaars. We also construct Bäcklund transformations describing the discrete time evolution of the system. This construction makes explicit the integrable structure underlying the quantum $K$-theory and provides a framework for further studies of the $K$-theoretic Peterson isomorphism.
Read the full abstract
Optimization and Control
Polynomial iteration complexity of a path-following smoothing Newton method for symmetric cone programming
It has long remained open whether smoothing Newton methods (SNMs) for symmetric cone programming (SCP) admit polynomial iteration complexity. A key difficulty lies in the lack of an analogue of the self-concordant convex framework underlying interior-point methods (IPMs). In this paper, inspired by Nemirovski's self-concordant convex-concave theory, we address this open problem by introducing a reduced barrier augmented Lagrangian (BAL) function. We prove that the reduced BAL function is self-concordant convex-concave and establish that the parameterized smooth system arising in SNMs coincides with the first-order optimality conditions of an associated minimax problem. Motivated by this equivalence, we propose a path-following smoothing Newton method (PFSNM). The reduced BAL function induces a central path and an associated neighborhood, which provide estimates for the Newton decrement needed for the path-following analysis. As a result, the method achieves an iteration complexity of $\mathcal{O}(\sqrt{\nu}\ln(1/\varepsilon))$, matching the best-known short-step complexity for IPMs. Numerical results on standard benchmarks show that PFSNM is competitive with several well-known interior-point solvers, and the observed performance is consistent with the theoretical development.
Read the full abstract
Rings and Algebras
Profinite tensor powers
We discuss the problem of defining a tensor product of profinitely many copies of a vector space $V$, and propose a definition $\bigotimes_X^{\mathrm{mcc}} V$ in the special situation that (1) $V$ is finite-dimensional over $\mathbf{F}_2$, and (2) the profinite $X$ indexing the tensor factors is acted on with finitely many orbits by a pro-$2$-group. The "mcc" on the tensor sign stands for "magnetized and conditionally convergent." A variant construction makes sense when $V$ is a bimodule over a ring of the form $\mathbf{F}_2 \times \cdots \times \mathbf{F}_2$, and the index set $X$ has the profinite version of a cyclic order. The definition organizes some computations in Heegaard Floer homology: it can be pitched as a computation of the Heegaard Floer theory of some pro-$3$-manifolds, though we do not know how to define such a thing.