Peer Revenue Equity in science

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

Mathematics

math Watch Mathematics (sign in to use)

Sunday 8 March

Logic

The reals as a subset of an ultraproduct of finite fields

In this paper we present new ways to construct external subsets of nonstandard models of arithmetic using mostly internal sets, and show that if an ultraproduct of prime finite fields includes a copy of the algebraic real numbers then either this copy or its algebraic closure can be constructed in some of these ways. We also show that no copy of the field of real numbers inside such an ultraproduct can ever be constructed in any of these ways, but there is either a hyperreal field or an algebraically closed field of cardinality larger or equal to the continuum that can be.
Read the full abstract

Functional Analysis

The Integration of Stepanov Remotely Almost Periodic Functions

The aim of this paper is to study the problem of the integration of Stepanov remotely almost periodic functions. We prove that every compact primitive of a Stepanov remotely almost periodic function with a minimal $\omega$-limit set is remotely almost periodic. This fact proves the conjecture previously formulated by the author.
Read the full abstract

Combinatorics

Spectral bounds for the independence number of graphs and even uniform hypergraphs

In this paper, we give spectral upper bounds for the independence number of even uniform hypergraphs and graphs, extend the Hoffman bound to even uniform hypergraphs, and give a simple spectral condition for determining the independence number, the Shannon capacity and the Lovász number of a graph. The Hoffman bound on the Lovász number is also extended from regular graphs to general graphs.
Read the full abstract

Functional Analysis

Properties of best approximations with respect to the Ky Fan $p$-$k$ norm, and the strict spectral approximant of a matrix

Some questions raised in [K. Ziętak, From the strict Chebyshev approximant of a vector to the strict spectral approximant of a matrix, Warsaw : Banach Center Publ., 112 Polish Acad. Sci. Inst. Math. (2017)] are discussed. To do so, the subdifferential set of the Ky Fan $p$-$k$ norm is computed. A characterization for the best approximations with respect to the Ky Fan $p$-$k$ norms is given. Further, necessary and sufficient conditions for $\varepsilon$-Birkhoff orthogonality with respect to the Ky Fan $p$-$k$ norm are also derived.
Read the full abstract

Information Theory

Pushing Bistatic Wireless Sensing toward High Accuracy at the Sub-Wavelength Scale

Contactless sensing using wireless communication signals has garnered significant attention due to its non-intrusive nature and ubiquitous infrastructure. Despite the promise, the inherent bistatic deployment of wireless communication introduces clock asynchronism, which leads to unknown phase offsets in channel response and hinders fine-grained sensing. State-of-the-art systems widely adopt the cross-antenna channel ratio to cancel these detrimental phase offsets. However, the channel ratio preserves sensing feature accuracy only at integer-wavelength target displacements, losing sub-wavelength fidelity. To overcome this limitation, we derive the first quantitative mapping between the distorted ratio feature and the ideal channel feature. Building on this foundation, we develop a robust framework that leverages channel response amplitude to recover the ideal channel feature from the distorted ratio. Real-world experiments across Wi-Fi and LoRa demonstrate that our method can effectively reconstruct sub-wavelength displacement details, achieving nearly an order-of-magnitude improvement in accuracy.
Read the full abstract

All authors Wenwei Li, Jiarun Zhou, Qinxiao Quan, Fusang Zhang, Daqing Zhang

Operator Algebras

On permanence of regularity properties II

In this article, we study the permanence of topological and algebraic dimension type properties of simple unital $C\sp*$-algebras. When a pair of unital $C\sp*$-algebras $(A, B)$ is associated by a $*$-homomorphism $\phi: A\to B$ which is tracially sequentially-split by order zero, we show that under suitable assumptions on either $A$ or $B$, or both $A$ and $B$ \begin{itemize} \item tracial $m$-comparison passes from $B$ to $A$; \item tracial $m$-almost divisibility passes from $B$ to $A$; \item tracial nuclear dimension less than $m$ passes from $B$ to $A$. \end{itemize}
Read the full abstract

Commutative Algebra

Generators of the initial ideal of simplicial toric ideals

We describe a generating set for the initial ideal of simplicial toric ideals with respect to the graded reverse lexicographic order, using representations of elements of affine monoids as sums of irreducible elements. Although the resulting generating set is not necessarily minimal, we demonstrate, through an example, how one can obtain the reduced Gröbner basis from it. Moreover, we compare the maximal degree of the Gröbner basis and the Castelnuvo-Mumford regularity.
Read the full abstract

Probability

On the slow points of fractional Brownian motion

Esser and Loosveldt have recently resolved a long-standing open problem in the folklore by proving that fractional Brownian motion (fBm) has slow points in the sense of Kahane, following a rich theory of slow points developed for Brownian motion and other, related, self-similar Markov processes. We presently introduce another method for the study of slow points in order to compute the Hausdorff dimension of fBm slow points. Our method follows recent ideas on the points of slow growth for SPDEs but also requires a number of new localization ideas that are likely to have other applications.
Read the full abstract

Optimization and Control · Cross-listed from stat.ME

Evaluating consumption effects of intelligent control algorithms for district heated buildings

As buildings become increasingly connected and sensor-rich, intelligent remote heating control is rapidly superseding conventional local heating control. Such control algorithms often aim at reducing energy consumption by minimizing over-heating and utilizing free solar energy, for instance. Numerous companies offering heating optimization solutions have recently emerged. After installing such a system, end-users naturally want to quantify and verify the effect of such an investment, i.e., monetary return. Methods for tracking buildings' heating efficiency are diverse, ranging from simple weather normalization to more complex modeling approaches, but lack transparency and commonly agreed best practices. The problem is further complicated by the fact that buildings constantly undergo non-control-related changes that affect their energy efficiency, making it difficult to isolate and track only control-related effects using the existing methods. In this paper, we first review and derive methods for monitoring the overall efficiency of buildings, and show their inability to isolate the control effects from other changes happening in the buildings. We then propose a model-based approach for estimating and tracking only the control-related effects. Moreover, we show how the models can decompose the total control effect into sub-components to reveal where the energy effects come from. We demonstrate the methods using real data collected over approximately 10 years from the Danfoss Leanheat Building platform. Our scope focuses on district heated buildings with substation-level (supply temperature) control, but the methodology extends to other cases as well.
Read the full abstract

All authors Antti Solonen, Arttu Häkkinen, Sallamaari Rapo, Antti Mäkinen, Sampo Kaukonen, Felipe Uribe

Dynamical Systems

Three heteroclinic orbits induce a countable family of equivalence classes of regular flows

We solve the problem of topological classification for smooth structurally stable flows on closed four-dimensional manifolds, the non-wandering set of which contains exactly two saddle equilibria, and the wandering set contains isolated trajectories connecting these saddle equilibria (heteroclinic curves). In particular, we show that for a flow of the class under consideration on $\mathbb{CP}^2$, the number of heteroclinic curves is a complete topological invariant, while on the sphere $\mathbb S^4$, there exists a countably many equivalence classes with an arbitrary odd number $\gamma\geq 3$ of heteroclinic curves. These results contrast with a three-dimensional case, where under similar conditions there exists only finite set of equivalence classes for each number of heteroclinic curves.
Read the full abstract

Probability · Cross-listed from stat.ML

Probabilistic Inference and Learning with Stein's Method

This monograph provides a rigorous overview of theoretical and methodological aspects of probabilistic inference and learning with Stein's method. Recipes are provided for constructing Stein discrepancies from Stein operators and Stein sets, and properties of these discrepancies such as computability, separation, convergence detection, and convergence control are discussed. Further, the connection between Stein operators and Stein variational gradient descent is set out in detail. The main definitions and results are precisely stated, and references to all proofs are provided.
Read the full abstract

Optimization and Control · Cross-listed from eess.SY

IQC-Based Output-Feedback Control of LPV Systems with Time-Varying Input Delays

Input delays are a common source of performance degradation and instability in control systems. This paper addresses the $\mathcal{H}_\infty$ output-feedback control problem for LPV systems with time-varying input delays under the integral quadratic constraint (IQC) framework. By integrating parameter-dependent Lyapunov functions with dynamic IQC multipliers, we derive convex, delay-dependent synthesis conditions formulated as parameter-dependent LMIs, enabled by the proposed exact-memory controller structure. An explicit controller reconstruction formula is provided to recover the LPV controller from the LMI solution, avoiding the need to specify the functional form of the parameter-dependent controller gains. While the synthesis problem for memoryless control is inherently non-convex, the proposed approach demonstrates significant performance improvement, reduced conservatism, and computational efficiency for standard output-feedback design. Numerical examples illustrate the effectiveness and broad applicability of the method to LPV systems with time-varying input delays.
Read the full abstract

Statistics Theory · Cross-listed from stat.ME

Dirichlet kernel density estimation on the simplex with missing data

Nonparametric density estimation for compositional data supported on the simplex is examined under a missing at random mechanism. Rather than imputing missing values and estimating the density from a completed data set, we adopt a strategy based on inverse probability weighting. The proposed estimator uses an adaptive Dirichlet kernel, which ensures nonnegativity on the simplex and favorable behavior near the boundary. When the observation probabilities are unknown, they are estimated through a Nadaraya-Watson regression step. The large-sample properties of the estimator are derived, including pointwise bias and variance expansions, optimal smoothing rates, and asymptotic normality. A simulation study investigates its finite-sample performance under varying sample sizes and missing rates. Simulations show our method outperforms inverse-probability-weighted kernel density estimators based on additive and isometric log-ratio transformations of the data for certain target densities. The methodology is further illustrated through an application to leukocyte composition data from the National Health and Nutrition Examination Survey (NHANES), which allows for the identification of the modal immune profile in the sampled population.
Read the full abstract

All authors Sami Baraket, Hanen Daayeb, Salah Khardani, Guanjie Lyu, Frédéric Ouimet

Combinatorics

2-switch: transition and stability on forests and pseudofests

Given any two forests (pseudoforests) with the same degree sequence, we show that one can be transformed into the other by a sequence of 2-switches in such a way that all the intermediate graphs of the transformation are forests (pseudoforests). We also prove that the 2-switch operation perturbs minimally some well-known integer parameters in families of graphs with the same degree sequence. Then, we apply these results to conclude that the studied parameters have the interval property in those families.
Read the full abstract

Optimization and Control · Cross-listed from cs.LG

Cost-Driven Representation Learning for Linear Quadratic Gaussian Control: Part II

We study the problem of state representation learning for control from partial and potentially high-dimensional observations. We approach this problem via cost-driven state representation learning, in which we learn a dynamical model in a latent state space by predicting cumulative costs. In particular, we establish finite-sample guarantees on finding a near-optimal representation function and a near-optimal controller using the learned latent model for infinite-horizon time-invariant Linear Quadratic Gaussian (LQG) control. We study two approaches to cost-driven representation learning, which differ in whether the transition function of the latent state is learned explicitly or implicitly. The first approach has also been investigated in Part I of this work, for finite-horizon time-varying LQG control. The second approach closely resembles MuZero, a recent breakthrough in empirical reinforcement learning, in that it learns latent dynamics implicitly by predicting cumulative costs. A key technical contribution of this Part II is to prove persistency of excitation for a new stochastic process that arises from the analysis of quadratic regression in our approach, and may be of independent interest.
Read the full abstract

All authors Yi Tian, Kaiqing Zhang, Russ Tedrake, Suvrit Sra

Information Theory

A Block Hierarchy for Rate-Distortion and Dispersion of Binary Markov Sources

The finite-length rate penalty for compressing a binary Markov source under Hamming distortion remains open in general. To study this penalty, we use memoryless sources that draw independent $K$-symbol blocks from the Markov law. We determine the rate and tilted-information variance per symbol at every distortion for positively correlated two-symbol blocks of arbitrary stationary bias and for symmetric three- and four-symbol blocks. For a symmetric chain with flip probability $0<p<1/2$, we prove that the normalized variance converges to a positive limit at every interior distortion where the Markov rate-distortion function is differentiable. At positive distortions up to Gray's critical distortion, this limit equals the source varentropy rate and yields a second-order coding converse for a single Markov path. We conjecture that the limit is the operational dispersion of the Markov source. Matching achievability remains open.
Read the full abstract

2603.07435 · 8 Mar · 30 pages, 13 figures. Retitled and substantially revised. Replaces the single-letter analysis with a block hierarchy. Proves convergence of the normalized block variance for symmetric persistent sources and a low-distortion second-order converse. Code, data, and proof companion: https://github.com/ANRGUSC/markov-block-hierarchy

Optimization and Control

Constrained Zero-Sum Stochastic Linear-Quadratic Differential Game for Jump-Diffusion Systems with Random Coefficients

This paper studies a two-player zero-sum stochastic linear-quadratic (SLQ) differential game for controlled jump-diffusion systems with random coefficients, where the controls of both players are constrained to nonempty closed convex cones. Under a uniform convexity–concavity condition, we establish the existence and uniqueness of an open-loop saddle point and characterize it by a forward–backward stochastic differential equation with jumps (FBSDEJ) together with cone-type variational inequalities. Assuming the existence of positive bounded solutions to the associated system of indefinite extended stochastic Riccati equations with jumps (IESREJs), we derive a feedback-form representation of the unique open-loop saddle point by constructing predictable minimax selectors and combining the Meyer–Itô formula with jumps, and the FBSDEJ characterization. Finally, under additional structural conditions, we prove the existence of positive bounded solutions to the IESREJs by a double-truncation approximation and a multidimensional BSDEJ comparison theorem.
Read the full abstract

Functional Analysis

Plank theorems, Gaussian probabilistic estimates and Rump's 100 Euro conjecture

We prove Rump's 100-euro conjecture by deriving a weighted affine escape theorem from Ball's plank theorem in [Invent. Math. 104 (1991)]. More precisely, let $\mathbb{K}\in\{\mathbb{R},\mathbb{C}\}$ and let $A\in\mathbb{K}^{n\times n}$. For every $1\le p\le \infty$, we obtain an $\ell_p$-escape principle controlled by the row $\ell_q$-norms of $A$. Its cube case shows that $|A|e=ne$, where $e$ is the all-one vector, implies the existence of a nonzero vector $x$ satisfying $\|x\|_{\infty}\le 1$ and $|Ax|\ge e\ge |x|$, thereby settling the conjecture. As a consequence, we prove the global comparison $\rho_0(|A|)\le n\,\rho_{\mathbb{K}}(A)$,where $\rho_{\mathbb{K}}$ denotes the sign-real or complex spectral radius, respectively. This is the sharp form of Rump's Perron–Frobenius-type estimate, with the factor $3+2\sqrt{2}$ removed. Moreover, our $\ell_\infty$-escape principle sharpens Rump's result in [SIAM Rev. 41 (1999)] on the relation between the entrywise distance to singularity of a matrix and its entrywise Bauer–Skeel condition number. Finally, we also investigate the weaker Euclidean row condition, including sharp quantitative bounds and counterexamples to possible strengthenings. In particular, we use Gaussian probabilistic estimates to establish a complex analogue of a conjecture of Bünger.
Read the full abstract

Information Theory · Cross-listed from cs.LG

Context Channel Capacity: An Information-Theoretic Framework for Understanding Catastrophic Forgetting

Catastrophic forgetting remains a central challenge in continual learning (CL), yet lacks a unified information-theoretic explanation for why some architectures forget catastrophically while others do not. We introduce Context Channel Capacity ($C_\mathrm{ctx}$), the mutual information between a CL architecture's context signal and its generated parameters, and prove that zero forgetting requires $C_\mathrm{ctx} \geq H(T)$, where $H(T)$ is the task identity entropy. We establish an Impossibility Triangle – zero forgetting, online learning, and finite parameters cannot be simultaneously satisfied by sequential state-based learners – and show that conditional regeneration architectures (HyperNetworks) bypass this triangle by redefining parameters as function values rather than states. We validate this framework across 8 CL methods on Split-MNIST (1,130+ experiments over 86 days, 4 seeds each), showing that $C_\mathrm{ctx}$ perfectly predicts forgetting behavior: methods with $C_\mathrm{ctx} = 0$ (NaiveSGD, EWC, SI, LwF, CFlow) exhibit catastrophic forgetting (6–97%), while methods with $C_\mathrm{ctx} \approx 1$ (HyperNetwork) achieve zero forgetting (98.8% ACC). We further propose Wrong-Context Probing (P5), a practical diagnostic protocol for measuring $C_\mathrm{ctx}$, and extend the framework to CIFAR-10 via a novel Gradient Context Encoder that closes the oracle gap from 23.3pp to 0.7pp. A systematic taxonomy of 15+ closed research directions – including the Hebbian null result (frozen random features outperform learned features), CFlow's $\theta_0$-memorizer phenomenon, and the $S_N$ symmetry barrier to column specialization – provides the community with precisely diagnosed negative results. Our central design principle: architecture over algorithm – the context pathway must be structurally unbypassable.
Read the full abstract

Analysis of PDEs

Global well-posedness and inviscid limit of the compressible Navier-Stokes-Vlasov-Fokker-Planck system with density-dependent friction force

This paper investigates the global dynamics of a three-dimensional fluid-particle interaction system that couples the compressible barotropic Navier-Stokes equations with the Vlasov-Fokker-Planck equation through a density-dependent friction force. The study establishes the global well-posedness, uniform-in-viscosity estimates, the global inviscid limit, and optimal large-time decay rates for classical solutions near equilibrium. First, for initial perturbations in $H^3$ sufficiently close to equilibrium, regularity estimates that are uniform in the viscosity coefficient are derived, and the existence of global classical solutions to the Cauchy problem is obtained. These uniform bounds enable us to rigorously justify the global-in-time inviscid limit as viscosity vanishes, with an explicit convergence rate proportional to the viscosity coefficient. This behavior differs significantly from that of the pure compressible Navier-Stokes system in the absence of particle interactions, emphasizing the stabilizing influence of kinetic coupling. Consequently, we establish for the first time the global existence of classical solutions to the compressible Euler-Vlasov-Fokker-Planck system. Moreover, under an additional mild assumption on the initial data, optimal time decay rates for both the solution and its spatial derivatives are obtained. Notably, the dissipative and microscopic components decay at a rate half an order faster than the macroscopic solution itself, indicating a novel relaxation mechanism induced by fluid-particle interactions. The analysis introduces new energy and dissipation structures for the coupled system, overcoming substantial difficulties arising from fluid-particle interactions.
Read the full abstract

Classical Analysis and ODEs

Sharp estimates for eigenvalues of localization operators before the plunge region

We study two closely related yet different localization operators: the time-frequency localization operator to the pair of intervals $S_{I, J} = P_I \mathcal{F}^{-1} P_J\mathcal{F} P_I$ and the localization of the coherent state transform to the square $L_Q$. Eigenvalues of both of them exhibit the same phase transition: if $|I| |J| = |Q| = c$ then first $\approx c$ eigenvalues are very close to $1$, then there are $o(c)$ intermediate eigenvalues and the rest of the eigenvalues are very close to $0$. Moreover, for both of them if $n < (1-\varepsilon)c$ for fixed $\varepsilon > 0$ then the eigenvalues are exponentially close to $1$. The goal of this paper is to establish sharp uniform bounds on these eigenvalues when $n$ is close to $c$ and see if there is a qualitative difference between the spectrums of $S_{I, J}$ and $S_Q$. We show that for $n < c -c^{0.99}$, say, in the time-frequency localization case we have $-\log(1-\lambda_n(c))\asymp\frac{c-n}{\log(\frac{2c}{c-n})}$ while in the coherent state transform case we have $-\log(1-\mu_n(c))\asymp (\sqrt{c}-\sqrt{n})^2,$ which is much smaller if $c-n = o(c)$, so there is indeed a difference between these two cases. The proofs crucially rely on the complex-analytic interpretations of these localization operators.
Read the full abstract

Dynamical Systems

Semialgebraic interior layers in root closures of bounded-height integer polynomials

Finite-degree root plots of bounded-height monic integer polynomials can exhibit gaps around interior points of their root closures. Outside the closed unit disk, these closures coincide with Mandelbrot sets for affine iterated function systems with collinear digits. We construct explicit semialgebraic interior layers, certified when finite inverse orbits enter canonical self-covering parallelograms. Such parallelograms exist exactly in the non-real lens, which contains the non-real locus precisely from twenty maps onward. Within the non-real lens, limits of each capture level are captured within two additional steps, optimally, and the layers recover the locus by closure. The layers expose persistent interior deficits: explicit quadratic parameters evade every finite level; sufficiently close distinct parameters have logarithmically growing capture depths; the underlying lattice obstruction yields degree-dependent root-free radii, explaining such gaps. For non-real parameters of modulus at least the square root of the number of maps, exact enclosure trees have sharp, alphabet-independent branching bounds.
Read the full abstract

Optimization and Control

Online Tracking with Predictions for Nonlinear Systems with Koopman Linear Embedding

We study the problem of online tracking in unknown nonlinear dynamical systems, where only short-horizon predictions of future target states are available. This setting arises in practical scenarios where full future information and exact system dynamics are unavailable. We focus on a class of nonlinear systems that admit a Koopman linear embedding, enabling the dynamics to evolve linearly in a lifted space. Exploiting this structure, we analyze a model-free predictive tracking algorithm based on Willems' fundamental lemma, which imposes dynamic constraints using only past data within a receding-horizon control framework. We show that, for Koopman-linearizable systems, the cumulative cost and dynamic regret of the nonlinear tracking problem coincide with those of the lifted linear counterpart. Moreover, we prove that the dynamic regret of our algorithm decays exponentially with the prediction horizon, as validated by numerical experiments.
Read the full abstract

All authors Chih-Fan Pai, Xu Shang, Jiachen Qian, Yang Zheng

Optimization and Control · Cross-listed from cs.LG

Feed m Birds with One Scone: Accelerating Multi-task Gradient Balancing via Bi-level Optimization

In machine learning, the goal of multi-task learning (MTL) is to optimize multiple objectives together. Recent works, for example, Multiple Gradient Descent Algorithm (MGDA) and its variants, show promising results with dynamically adjusted weights for different tasks to mitigate conflicts that may potentially degrade the performance on certain tasks. Despite the empirical success of MGDA-type methods, one major limitation of such methods is their computational inefficiency, as they require access to all task gradients. In this paper we introduce MARIGOLD, a unified algorithmic framework for efficiently solving MTL problems. Our method reveals that multi-task gradient balancing methods have a hierarchical structure, in which the model training and the gradient balancing are coupled during the whole optimization process and can be viewed as a bi-level optimization problem. Moreover, we showcase that the bi-level problem can be solved efficiently by leveraging zeroth-order method. Extensive experiments on both public datasets and industrial-scale datasets demonstrate the efficiency and superiority of our method.
Read the full abstract

All authors Xuxing Chen, Yun He, Jiayi Xu, Minhui Huang, Xiaoyi Liu, Boyang Liu, Fei Tian, Xiaohan Wei, Rong Jin, Sem Park, Bo Long, Xue Feng

Operator Algebras

An Index Theorem for Fredholm Operators via the Unitary Conjugation Groupoid

In a previous paper we introduced the unitary conjugation groupoid associated to any unital separable Type I C*-algebra. This groupoid encodes the representation-theoretic structure of the algebra through the action of its unitary group on the characters of commutative subalgebras and admits a canonical embedding of the algebra into its groupoid C*-algebra. In this paper we apply this framework to two fundamental operator algebras: the algebra of all bounded operators on a separable Hilbert space and the unitization of the compact operators. For any Fredholm operator in these algebras we construct a natural equivariant KK-theory class using the phase of the operator. Applying Kasparov descent for groupoids produces a class in the K-theory of the associated groupoid C*-algebra. Using the Morita equivalences established in the previous work, this descended class can be identified with the symbol class of the operator. We prove that composing this class with the boundary map of the Calkin extension recovers the classical Fredholm index. In particular, the construction yields index minus one for the unilateral shift and index zero for compact perturbations of the identity. This provides a groupoid equivariant formulation of the Fredholm index and connects the unitary conjugation groupoid with classical operator index theory. This perspective suggests a general approach to index theory through equivariant K-theory of groupoids associated with operator algebras.
Read the full abstract

Saturday 7 March

Information Theory · Cross-listed from cs.LG

Beyond Reward Suppression: Reshaping Steganographic Communication Protocols in MARL via Dynamic Representational Circuit Breaking

In decentralized Multi-Agent Reinforcement Learning (MARL), steganographic collusion – where agents develop private protocols to evade monitoring – presents a critical AI safety threat. Existing defenses, limited to behavioral or reward layers, fail to detect coordination in latent communication channels. We introduce the Dynamic Representational Circuit Breaker (DRCB), an architectural defense operating at the optimization substrate. Building on the AI Mother Tongue (AIM) framework, DRCB utilizes a Vector Quantized Variational Autoencoder (VQ-VAE) bottleneck to convert unobservable messages into auditable statistical objects. DRCB monitors signals including Jensen-Shannon Divergence drift, L2-norm codebook displacement, and Randomized Observer Pool accuracy to compute an EMA-based Collusion Score. Threshold breaches trigger four escalating interventions: dynamic adaptation, gradient-space penalty injection into the Advantage function A^pi, temporal reward suppression, and full substrate circuit breaking via codebook shuffling and optimizer state reset. Experiments on a Contextual Prisoner's Dilemma with MNIST labels show that while static monitoring fails (p = 0.3517), DRCB improves observer mean accuracy from 0.858 to 0.938 (+9.3 percent) and reduces volatility by 43 percent, while preserving mean joint reward (p = 0.854). Analysis of 214,298 symbol samples confirms "Semantic Degradation," where high-frequency sequences converge to zero entropy, foreclosing complex steganographic encodings. We identify a "Transparency Paradox" where agents achieve surface-level determinism while preserving residual capacity in long-tail distributions, reflecting Goodhart's Law. This task-agnostic methodology provides a technical path toward MICA-compliant (Multi-Agent Internal Coupling Audit) pre-deployment auditing for autonomous systems.
Read the full abstract

Information Theory · Cross-listed from cond-mat.stat-mech

Transient Acceleration and Cross-Dissipation Interference in Fisher-Regularized Wasserstein Gradient Flows

We study transient nonequilibrium dynamics in Fisher-regularized Wasserstein gradient flows and identify a sign-changing cross-dissipation mechanism generated by the coupling between transport dissipation and Fisher-information geometry. Using the Ornstein–Uhlenbeck Fokker–Planck system as an analytically tractable setting, we derive an exact reduced variance dynamics on the Gaussian manifold, \[ \dot{u}=2(1-u)+\frac{\varepsilon}{u}, \] where \(u(t)=\sigma^2(t)\) is the variance and \(\varepsilon>0\) is the Fisher regularization strength. The reduced dynamics reveal distinct transient regimes induced by the interaction between transport relaxation and information-geometric curvature. The associated cross-dissipation term changes sign at the critical scale \(\sigma=1\), separating cooperative acceleration for localized states with \(\sigma<1\) from transient interference at larger variance scales. In the subcritical regime, Fisher curvature accelerates the descent of the baseline free energy; beyond the critical transition, it partially opposes the Ornstein–Uhlenbeck pullback and generates transient overshoot toward a displaced Fisher-regularized equilibrium. We also establish a bounded transient-acceleration-window result, showing that the cooperative acceleration phase has finite duration with an upper bound depending only on the Fisher regularization strength. Finite-difference simulations support the analytical predictions and suggest that qualitatively similar sign-transition behavior may persist beyond Gaussian closure for non-Gaussian initial conditions, including bimodal and Laplace distributions. Overall, the results provide a transient dynamical perspective on Fisher-regularized dissipative systems and show how information-geometric curvature can reorganize intermediate-time Wasserstein relaxation while preserving the globally dissipative structure of the flow.
Read the full abstract

History and Overview of Mathematics

Mathematical Proof

These Course Notes provide an introduction to mathematical proofs for undergraduate students transitioning from computational calculus to abstract mathematics. Topics include propositional logic, proof techniques, mathematical induction, fields, sets and relations, sequences and series, completeness of the real numbers, cardinality, and related foundational material. Numerous examples and exercises (with complete solutions) are included. The notes are designed for a one-semester proof course.
Read the full abstract

Logic

Ranked Forcing and the Length of Generalized Borel Hierarchies

We extend A. Miller's framework of $\alpha$-forcing to the case of a regular uncountable cardinal $\kappa = \kappa^{<\kappa}$ and apply it to study the structure of the $\kappa$-Borel hierarchy on subspaces of the generalized Baire space ${}^\kappa \kappa$. We isolate a class of iterations of $\alpha$-forcing and show that it satisfies a certain combinatorial property of admitting a sufficiently rich family of rank functions; this fact is then used to construct several models in which nontrivial constellations for the length of the $\kappa$-Borel hierarchy on multiple subspaces of ${}^\kappa \kappa$ are realized simultaneously. Finally, we provide a higher variant of Steel's forcing with tagged trees and generalize arguments of Stern to derive the exact $\kappa$-Borel complexity of certain classes of well-founded trees.
Read the full abstract