Machine Learning
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Sunday 8 November
Diffusion models have demonstrated the effectiveness of score-based objectives for generative modeling, yet their potential for self-supervised representation learning remains underexplored. We introduce Entropy-Based Regularization (EBReg), a framework that uses latent score estimation to regularize the latent distribution during self-supervised training. EBReg jointly learns an encoder and a latent score model, and utilizes the learned score to provide a diversity-preserving signal that prevents representational collapse. Motivated by a mutual-information lower bound, we formulate the objective as view alignment with smoothed marginal entropy maximization and optimize the marginal entropy without explicit density evaluation. The same score model can additionally be used at test time to refine corrupted representations toward the training latent distribution. Empirically, EBReg learns competitive representations on ImageNet-1K, achieving 75.57% linear-probe accuracy with ViT-L. Across diverse architectures, score-guided test-time adaptation further improves robustness on ImageNet-C, yielding an average relative accuracy gain of approximately 5%.
Computer Vision
SegDis: Foreground Segmentation Followed by Type Discrimination for Salient and Camouflaged Object Detection
Salient and camouflaged object detection (SOD and COD) both recover foreground support, yet most joint methods output only a binary mask and therefore cannot distinguish the type. USC12K distinguishes the two types, but its baseline USCNet still binds foreground existence to the type within a single normalization, so the failure source cannot be attributed. Our cross-task study shows that foreground localization transfers between SOD and COD, whereas type discrimination does not: a COD-only model recovers foreground on SOD images but its type judgement follows the training distribution rather than the foreground-background relation, so the two decisions require separate mechanisms. SegDis factorizes the output into an explicit foreground probability and a conditional type probability given foreground: FLR provides high-quality task-agnostic support, and RTA refines the SOD/COD assignment from the foreground-background relation without a task identifier. Because support is explicit, each ground-truth foreground pixel is classified as correct-type, type-swapped, or missed, so foreground-recovery and type-assignment errors are diagnosed separately. With PVT-v2-B2, SegDis attains 68.98% FG-mIoU and 78.35% mIoU on USC12K at about one-eighth of USCNet's parameters; CSCS is 5.75.
Read the full abstract
Systems and Control
Scenario-Based Assessment of Transport Electrification for Low-Carbon Urban Planning: A Case Study of Phu Quoc, Vietnam
Urban net-zero strategies require a defensible emissions boundary and quantitative estimates of the contribution achievable through individual mitigation measures. This study assesses road-transport electrification within a planned residential development in Phu Quoc, Vietnam, using the assumptions and scenario data of an urban-planning study originally prepared in 2022. A vehicle-class inventory combines population-dependent fleet projections, assumed travel activity, gasoline and diesel consumption, fuel-specific carbon dioxide (CO_2) factors, and electricity-related emissions from battery-electric vehicles. The reference case projects annual road-transport emissions of 59.41, 73.86, and 87.23 tCO_2 for 2022, 2025, and 2030, respectively. Three year-specific substitution scenarios, corresponding to nominal electric-vehicle shares of 25%, 50%, and 100%, yield source-reported avoided emissions of 9.43, 13.64, and 55.41 tCO_2/year, or 15.9%, 18.5%, and 63.5% of their respective baselines. However, a cross-scenario audit identifies an inconsistency in the 2025 mixed-fleet results under time-invariant vehicle-class emission-intensity ratios: under that additional assumption, 2025 avoided emissions would be 23.45 rather than 13.64 tCO_2/year. The original detailed vehicle-energy input tables are not included in the available source document, preventing definitive reconciliation. The study therefore provides a transparent, preliminary planning assessment rather than evidence of citywide net-zero attainment. Its principal value is a reproducible accounting structure and an explicit identification of the data, electricity-system, and infrastructure constraints that must be resolved before policy or investment decisions.
Read the full abstract
Discrete Mathematics
Fractional distance-edge-monitoring in lexicographic products
We give an exact formula for the fractional distance-edge-monitoring number of $G\circ H$, where $G$ is connected of order $m\ge2$ and $H$ has order $n\ge2$. By the previously established endpoint-monitoring property, the fractional monitoring program reduces to fractional vertex cover, whose linear-programming dual is fractional matching. Writing $d_c(H)$ for the critical difference of $H$, we obtain $\operatorname{dem}_{f}(G\circ H)=(mn-D(G,H))/2$, where $D(G,H)=\max_{A\in\mathcal{I}(G)}\bigl(d_c(H)|A|-n|N_G(A)|\bigr)$ and $\mathcal{I}(G)$ denotes the family of all independent sets of $G$, including the empty set, while $N_G(A)$ is the open neighborhood of $A$. This formula characterizes the products attaining $mn/2$ and gives an equality criterion for the fractional matching lower bound of Ma, Cheng, Mao and Wang. Under the stated hypotheses, the lower bound of Ma et al. is attained if and only if either factor has a fractional perfect matching or $H$ is edgeless. Two outer trees show that the factor orders and fractional matching numbers do not determine the product value.
Read the full abstract
Computational Complexity
Linear Objective Problems: A Complexity Framework for Bilevel Optimization
Many bilevel and robust optimization problems are complete for the second level of the polynomial hierarchy, but such results are usually proven one problem at a time with problem-specific reductions. A recurring obstacle is that the lower level must select an optimal solution rather than merely a feasible one, whereas standard reductions preserve only whether a solution exists, not which solutions are optimal. We introduce linear objective problems (LOP), a framework for optimization problems whose solutions are explicit subsets of a ground set and whose objective is linear in these elements. We define tight solution-embedding reductions, which map the optimal solutions of one problem element by element onto the optimal solutions of another, without having to recognize optimal solutions, which is coNP-hard in general. This gives a notion of LOP-completeness, with Satisfiability as the canonical complete problem, even with unit weights. We show that many classical problems are LOP-complete, including Vertex Cover, Dominating Set, TSP, Knapsack, and Steiner Tree, and we introduce a duality that transfers LOP-completeness between minimization and maximization problems. As our main application, we prove a meta-theorem for min-max regret optimization under interval uncertainty: for every LOP-complete problem, the min-max regret version is $\Sigma_2^p$-complete. Instead of a separate proof for each nominal problem, it suffices to show that the nominal problem is LOP-complete. This recovers known results for the interval min-max regret versions of Knapsack and Maximum Benefit Set Cover and yields completeness for over 50 further problems.
Read the full abstract
Sound
Small Data Generative AI for Culturally Specific Artistic Creation: A Case Study on the Sonic Space of sarir
Large-scale generative AI (GenAI) models have become increasingly integrated into contemporary creative practice, yet they often reproduce cultural biases and fail to meaningfully engage with non-Western artistic traditions. In response, this paper explores a small-data approach that foregrounds artist-curated and culturally informed dataset creation and model training. Through a case study on sarir (the distinctive sound produced by the reed pen in Persian calligraphy), we demonstrate how small-data GenAI can operate within a highly situated cultural context that lies beyond the representational reach of large-scale GenAI models. We construct a bespoke sarir audio dataset and train two models, RAVE and StyleGAN2, adapting each to the sonic domain through audio synthesis and spectrogram generation. The resulting models learn the fine-grained timbral and gestural nuances of sarir, enabling both real-time interaction and extended generative output. By comparing these results with state-of-the-art text-to-audio models, we clearly demonstrate the disparity in their performance and highlight the expressive possibilities unlocked through the small data approach. This focused case study demonstrates the feasibility of small-data GenAI for culturally specific artistic creation and underscores the value of further investigations that extend this approach to other cultural materials and artistic domains.
Read the full abstract
Trading and Market Microstructure
Optimal execution in a fractional setting with reinforcement learning
We consider the problem of optimal liquidation with linear market impact, as in the seminal work of Almgren and Chriss, but with an asset driven by a fractional Brownian motion. For $H \neq 1/2$, we observe that the deterministic solution given by Almgren and Chriss is no longer optimal. We develop a strategy to exploit the information induced by the correlation between price increments. The strategy is derived from a policy gradient algorithm, parameterized by a neural network, trained to maximise the CARA utility, for which our method allows for significant improvements.
Read the full abstract
Machine Learning
LLM-based Multi-Criteria Deep Reinforcement Learning for Dynamic Efficient 5G UPF Selection
User plane function (UPF) selection in 5G networks is an active area of research, with various limited approaches proposed to enhance network efficiency and manage resources effectively. In this paper, we address this gap in the related literature by proposing a Deep Reinforcement Learning (DRL) UPF selection method that utilizes a dynamic, multi-objective reward function and validating it on a real-world 5G emulation platform. The proposed method is based on the large language models' (LLMs) translation of network operator intents and the network data analytics function (NWDAF) abnormality alerts. The experimental results validated in the real-world experimental setup demonstrate that our dynamic DRL solution achieves an effective balance between latency, power consumption, and traffic distribution, outperforming the heuristic baselines and fixed DRL by 16.89%. These findings highlight the real-world applicability of our approach for dynamic, intent-driven UPF selection.
Read the full abstract
Machine Learning
D3Tree: Learning to Prune Block-Diffusion Draft Trees for Efficient Verification
Block-diffusion drafters propose a whole block of tokens in one forward pass, and draft trees built from their position-wise marginals increase the expected accepted length of speculative decoding. Every tree node occupies a query row in target verification, so under concurrent serving low-yield branches compete with other requests for verification capacity. Existing tree methods select verified nodes by draft probability under a fixed or cost-selected per-request budget, and thus always retain a prefix of the drafter's ranking. We present D3Tree, which learns which nodes of a block-diffusion draft tree are worth verifying. A tree-size head with 2H+2 parameters predicts each candidate's conditional survival from the draft state, the candidate embedding, and its draft log-probability. Thresholding the product of these scores along parent chains yields an ancestor-closed subtree whose size adapts to each request and decoding round. The head is trained on a frozen drafter with conditional labels on the full candidate tree; an optional joint mode also steers the drafter toward the retained reference path. At serving time, D3Tree verifies whole trees under light load and, under heavy load, allocates one batch-wide node budget by predicted survival, sized to the target's matrix-multiplication tiles. Retained trees share one ancestor-masked target pass, while construction, pruning, and scheduling run on the GPU and overlap with the host. Across six vision-language verifiers and 17 text and multimodal benchmarks, D3Tree is fastest in 70 of 72 load settings, with 3.2-4.0x speedup over autoregressive decoding for one request and 1.3-1.8x at 64 concurrent requests, where fixed-budget trees fall below autoregressive speed. Under a time-varying arrival trace, it has the lowest mean end-to-end latency in every arrival phase.
Read the full abstract
Distributed Computing
A Simple Phase-King Protocol with Predictions
Inspired by recent work (Ben-David, Dzulfikar, Ellen, and Gilbert, $\textit{Byzantine Agreement with Predictions}$, PODC 2025) we study synchronous binary Byzantine Agreement in the learning-augmented setting, where each process receives predictions about which processes are honest or Byzantine. We present a deterministic Phase King protocol for the anauthenticated setting, combining early stopping with a specifically designed prediction-aware king selection procedure. For at most $t<n/3$ Byzantine processes, the protocol handles arbitrarily many prediction errors, improving upon the $O(n^{3/2})$ bound of Ben-David et al., and terminates within $O(P)$ communication rounds, where $P=1+\min\{\mathbf{B}/n,f\}$, $\mathbf{B}$ is the total number of incorrect prediction bits held by honest processes, and $f$ is the actual number of Byzantine processes. Our protocol uses $O(n^2P)$ messages and $O(n^2P\log n)$ bits, avoiding the initial cubic bit cost of exchanging complete prediction vectors. In comparison to very recent independent work (Dufay, Dzulfikar, and Gilbert, $\textit{Predictions Can Only Help! Communication Efficient Byzantine Agreement with Predictions}$, DISC 2026), which achieves the same error tolerance and round complexity but better communication bounds, our construction provides a much simpler Phase King approach that completely avoids involved preprocessing constructions, rendering it more practical, especially for achieving agreement in time-evolving networks. We believe that the novel prediction-aware king selection technique of our scheme is of independent interest and can find further applications in learning-augmented distributed consensus.
Read the full abstract
Combinatorics
Tiling of Hyperbolic Surface by Maximal Number of Tiles
Tilings of a closed hyperbolic surface by congruent $n$-gons with $n\ge 7$ is a finite problem. One extreme of the finite problem is that the number of tiles is maximal. The problem is related to edge-to-edge tilings of the Poincaré disk ${\bb H}^2$ by congruent $n$-gons, such that all vertices have degree 3 and the vertex sets satisfy certain balance condition. For the special case $n=7$, we show that the heptagon in any such tiling belongs to one of the four families, and tilings of ${\bb H}^2$ by generic heptagons in each family is unique. As a consequence, for any closed hyperbolic surface of Euler characteristic $\chi$, we may classify all tilings of the surface by maximal number $-6\chi$ of congruent generic heptagons in the four families.
Read the full abstract
Artificial Intelligence
BiVerify: Closing the Loop Between Backward and Forward Verification for Mathematical Self-Correction
Large language models (LLMs) have made substantial progress in mathematical reasoning but still produce incorrect solutions to complex problems, indicating the need for more reliable self-correction mechanisms. Mathematical self-correction typically identifies errors through verification, which can proceed backward or forward. The two directions provide complementary verification capabilities. Backward verification checks a candidate answer against the original constraints, bypassing complex intermediate reasoning, but cannot ensure correct reasoning or objective completeness. Forward verification follows the problem conditions and current solution, enabling it to locate reasoning errors and assess objective completeness, but may overlook key constraints in complex solution chains. We propose BiVerify, a dynamically scheduled framework for bidirectional mathematical verification. BiVerify begins with backward verification to screen the answer against the original constraints, followed by forward verification of the reasoning process; subsequent direction switching, state resetting, and re-verification are governed by verification outcomes. Upon failure, the system revises the candidate or verification flow until the same candidate passes both directions. We also introduce shared mathematical tools that separate logical reasoning from precise computation. We systematically evaluate BiVerify on five challenging mathematical reasoning benchmarks using multiple mainstream DeepSeek, GPT, and Gemini models. BiVerify exceeds the strongest representative public baseline evaluated in this work by 13.5 percentage points in average accuracy, with the margin reaching 31.2 percentage points on GPT. Further analysis shows that BiVerify's overall performance exceeds that of either verification direction alone.
Read the full abstract
High Energy Physics (Phenomenology)
Probing Exotic Electron-Electron Interactions via Adiabatic Spin Inversion in MRFM
Light bosons beyond the Standard Model can mediate exotic spin-dependent interactions between electrons, motivating sensitive searches at short distances. We propose a magnetic resonance force microscopy (MRFM) scheme that exploits oscillating-cantilever-driven adiabatic reversals (OSCAR) of electron spins to convert these interactions into measurable mechanical signals. The ferromagnetic tip serves as both a dense polarized-spin source and a source of magnetic-field gradients, enabling force detection without magnetic shielding between the source and sensor. Within a linearized treatment of the coupled spin-mechanical dynamics, we derive resonance-frequency shifts associated with the $V_2$ and $V_3$ interactions and a second-harmonic force signal arising from the velocity-dependent $V_8$ interaction. We further investigate interrupted-OSCAR differential detection and distance-dependent measurements to discriminate exotic signals from conventional magnetic backgrounds. Using representative parameters from ensemble and single-spin MRFM experiments, we estimate the projected sensitivity under specified noise and statistical assumptions. The results suggest that the proposed scheme could provide complementary sensitivity to short-range spin-dependent interactions, with the potential to improve existing constraints on $V_2$ over selected submicrometer interaction ranges. Our study establishes adiabatic spin-inversion MRFM as a promising approach to probing exotic electron-electron forces.
Read the full abstract
General Relativity and Quantum Cosmology
Exceptional Points in Scalar Quasinormal Modes of Rastall String-Fluid Black Holes
In this paper, we study exceptional points (EPs) in scalar quasinormal modes(QNMs) of Rastall string-fluid black holes by adding a localized Pöschl–Teller perturbation to the scalar potential, with the bump position $a$ and amplitude $\eb$ serving as independent controls. We solve the hyperboloidal problem with two-domain Chebyshev collocation and arbitrary-precision linear algebra, and use the complex discriminant, eigenvector coalescence, Jordan-chain residuals, square-root splitting and closed-loop mode exchange as complementary tests of defectiveness. For the regular $A=1$, $\eR=-1$ branch, the two-parameter search locates a resolution-stable second-order EP from $N=24$ to $N=60$, while constrained semidiscrete evolution gives the Jordan response $(C+Dt)e^{\ii\omega_{\rm EP}t}$. For the logarithmic $A=2$ branch, defective points occur in each finite truncation but their coordinates do not converge, so we make no continuum claim for this branch.
Read the full abstract
Machine Learning
MG-Cache: Multi-Granularity Caching for Accelerating Diffusion Transformer Inference
Diffusion Transformer (DiT) models, while having become a powerful backbone for generative synthesis, suffer from high inference cost on generating high-resolution images. Diffusion caching has emerged as a promising acceleration technique to reduce inference cost. However, existing caching strategies fail to accurately identify the varying importance of different computational modules and rarely coordinate cache reusing across the timestep and layer granularities, making it difficult to control error accumulation. To address these limitations, we propose multi-granularity caching (MG-Cache) for DiTs, a novel and fully training-free approach to exploit computational redundancy both on timesteps and Transformer layers. Our core innovation contains a two-stage caching strategy: an offline timestep profiling analysis that identifies redundant timesteps for cache reuse, followed by an adaptive layer-wise caching strategy that combines input stability with importance of each Transformer layer to selectively cache stable Attention and MLP submodules. Extensive experiments on mainstream open-source DiT models validate the effectiveness of our approach, achieving significant acceleration while maintaining high fidelity to the uncached output, with up to a 4.18x acceleration on HunyuanVideo.
Read the full abstract
Distributed Computing
OOM: Offload-Optimized Full-Parameter MoE Training via Memory-Execution Co-Design
Offloading model state to host memory and streaming layer weights let a single multi-GPU node train large mixture-of-experts (MoE) models. GPU memory then becomes a performance budget, shared by resident expert state, retained activations, and execution worksets. The best allocation depends on the expert execution path, batch size, and sequence length. We present OOM (Offload-Optimized MoE), a full-parameter MoE training system that co-designs expert execution and memory allocation. For expert-parallel execution, OOM updates resident experts on their owning GPUs and jointly plans state residency and MoE activation retention according to pipeline dependencies. For local execution, OOM bounds GPU worksets through dependency-aware partitioning; in hybrid models, it also retains attention results on the host to avoid costly recomputation and selectively regenerates layer boundaries through cheaper layers to make room. Calibrated cost models guide joint configuration within GPU and host budgets to minimize step time; OOM selects among the configured paths through short executions. On four H100s, OOM trains with 1M-token contexts, whereas FSDP2, ZeRO-Offload, and MegaTrain exhaust GPU memory at 512K. It also trains a 122B-parameter MoE model, on which all three exhaust host memory. On four 24 GB RTX 4090s, it trains 512K-token contexts and 1,024 1K-token sequences per step without gradient accumulation. Across matched-global-batch comparisons, OOM delivers 1.58x and 4.08x geometric-mean (up to 2.14x and 7.77x) throughput over CPU-offloaded FSDP2 on the two platforms while using the least host memory among the evaluated systems – as little as 1/3.0 of FSDP2's.
Read the full abstract
Representation Theory
Sections of the Kottwitz homomorphism and affine pinnings
Let $G$ be a connected reductive group split over $\mathbb{Z}$. The Kottwitz homomorphism identifies the group of connected components of the loop group of $G$ with the group $\Omega$ of length zero elements of the extended affine Weyl group. We show that this homomorphism always has a homomorphic section, and that the section can be chosen inside the normalizer of $T$ in $G(\mathbb{Z}[t,t^{-1}])$. We then ask whether the section can be chosen to preserve an affine pinning. This is possible if no simple factor of $G$ is of type $B_{2m}$ or $C_{2m}$. It is impossible for $(SL_2\times Sp_{2n})/\mu_2$ and $(SL_2\times Spin_{2n+1})/\mu_2$ when $n$ is even. We show that the obstructions for $G$ and for its Langlands dual group are dual to each other, and we compute them by a lattice computation with a Coxeter element. As an application, we show that the component group of the centralizer of any semisimple element lifts to the centralizer. Under the same condition on the simple factors, the lift can be chosen to preserve a pinning of the neutral component.
Read the full abstract
Algebraic Topology
Fundamental Groups of Permutation Complexes
We present an elementary argument showing that fundamental groups of permutation complexes are free. Hence, $\mathbb{R} P^2$ cannot be realized by a permutation complex, answering a question of Chachólski, Levi and Meshulam.
Classical Analysis and ODEs
Uncertainty principle on $\mathbb{T}^m \times \mathbb{R}^n$
This paper establishes an uncertainty principle on the mixed space $\mathbb{T}^m \times \mathbb{R}^n$ for $n \ge 2$. Leveraging the paraboloid restriction theorem on $\mathbb{Z}^m \times \mathbb{R}^n$ derived from Barron's global-in-time Strichartz estimates for the semi-periodic Schrodinger equation, we construct measurable sets $S \subset \mathbb{T}^m \times \mathbb{R}^n$ and $A(R,\delta) \subset \mathbb{Z}^m \times \mathbb{R}^n$ that form a strong $L^2$ annihilating pair, satisfying the uniform inequality $\|f\|_{L^2(\mathbb{T}^m \times \mathbb{R}^n)} \le 3(\|f\|_{L^2((\mathbb{T}^m \times \mathbb{R}^n)\setminus S)} + \|\widehat{f}\|_{L^2((\mathbb{Z}^m \times \mathbb{R}^n)\setminus A(R,\delta))})$ for all $f \in L^2(\mathbb{T}^m \times \mathbb{R}^n)$, where the measures of $S$ and $A(R,\delta)$ can be made arbitrarily large while the constant 3 remains unchanged. As an application, we derive a two-time observability inequality for the semi-periodic Schrodinger equation under a support condition on the phase-modified solution at the first observation time.
Read the full abstract
Analysis of PDEs
Uniqueness of Axisymmetric Vortices in the Magnetic Ginzburg-Landau Equations
We study vortex solutions of the magnetic Ginzburg-Landau model (or the Abelian-Higgs model). We establish uniqueness of axisymmetric vortex solutions for every winding number $n\geq1$ and coupling constant $\lambda>0$. Our proof uses a lattice inequality and constrained minimization to construct a positive radial solution lying above any two given radial solutions. A comparison argument based on the virial identity then establishes uniqueness. We further prove that the unique radial solution depends real analytically on $\lambda$ and that both radial profiles are strictly increasing with respect to $\lambda$.
Read the full abstract
Computer Vision
NuJEV: Structured Visual Judgments for Consistent Nuclei Quantification
Nuclear quantification requires locating and typing nuclei, beyond generating a count. Inspired by Jev's structured decisions in place of free-form generation, we introduce NuJEV for nuclear perception in Qwen3-VL-2B. We jointly train visual LoRA and a parallel head to predict nuclear presence, centers, and types. One image encoding feeds both the nuclear head and original language decoder. The decoder interprets region and type requests, with counts computed from the predicted nuclei. On 2,732 PanNuke test images, NuJEV achieves a five-query mean absolute error ($\mathrm{MAE}_5$) of 1.844 overall and 5.268 on dense images, with localization/typed F1 of 0.816/0.682. A matched encoder ablation shows 41.0% lower dense error with visual adaptation. For language requests, NuJEV reduces request MAE by 76.2% relative to direct generation. Our code will be released as open source.
Read the full abstract
Algebraic Topology
Persistence landscape velocity: Closed formulas and a topological probe of quantum criticality
We introduce the landscape velocity $\xi$, a scalar measure of variation of a parameter-dependent persistence landscape. For families of finite barcodes whose bars are all born at the fixed filtration value $0$, we show that $\xi$ is the $L^2$ speed of the landscape curve where the sorted death times are differentiable and derive an exact closed formula that requires no persistence landscape at all: the death times of the barcodes, as functions of the parameter, suffice. For the velocity of the empirical mean of several landscapes we also prove a closed formula which, again, strikingly requires neither the empirical mean landscape nor any of the individual landscapes, only the individual barcodes. As an application, we revisit Olsthoorn's persistent homology analysis of the transverse-field Ising chain. Whereas his approach uses different homology degrees and summaries depending on the boundary conditions, ours uses $H_0$ only and a single rule: the position of the maximum of $\xi$. This rule locates the quantum critical point for both periodic and open boundary conditions, and for both mutual-information and correlator-based dissimilarities. The position of the maximum converges to the critical point faster than all the estimators we can extract from Olsthoorn's work. In the 4 scenarios considered (different boundary conditions and different dissimilarity matrices), the peak of the normalized velocity becomes increasingly pronounced with system size, with observed logarithmic growth. This logarithmic growth is proved for periodic boundaries with correlator dissimilarity under stated hypotheses, identifying landscape velocity as a singular topological response to quantum criticality.
Read the full abstract
Commutative Algebra
A divisorial domain with non-Prüfer normalization
We answer a 1968 question of Heinzer by constructing a two-dimensional local divisorial domain whose normalization is not Prüfer.
Quantum Physics
What an Observer Reads off a Holographic Screen
What does an observer read off a holographic screen that is an array of qubits rather than a surface? In the generalized holographic principle the interaction between two separable systems acts on such a screen as a weighted sum, up to an overall prefactor, of commuting observables with values plus and minus one. When the weights are nonnegative and the prefactor is one and the joint eigenspaces are lines, everything a probe reads off the spectrum depends only on the list of weights, which we call the profile. On the screen with equal weights summing to one the running dimension read by a diffusion probe falls short of two at unit time by at most twice the reciprocal of the number of qubits; a geometric profile gives instead, through its level count, a plateau of any positive height. An observer correlates sets of qubits; its readout carries a profile at a given weighting when the levels it reads, counted with multiplicity, are those of a screen with one weight per set. We prove that a readout carries a profile at every weighting exactly when the distinct nonzero traces of its qubits are linearly independent over the field with two elements, the trace of a qubit being the list of sets that contain it; such a readout is a partition of the qubits it covers up to an invertible recoding, and a dependent one carries a profile only on finitely many proper linear subspaces of weightings. Refining a grouping without end builds a tower whose boundary is a profinite set and hence a condensed set, on which a strictly positive compatible profile separates the locally constant functions by a distance. Applications certify probabilistic and strong contextuality and give the leading shortfall of the distance between the poles of the fuzzy sphere; in a model of monogamous entanglement the pairs that weight-one threads representing operational ER=EPR can occupy at once are exactly the partial matchings.
Read the full abstract
Cryptography and Security
Attacking Slide Puzzle CAPTCHAs Based on Time and Space Generative Adversarial Networks
CAPTCHAs can protect websites from malicious attacks. As a typical representative of behavior CAPTCHAs, slide puzzle CAPTCHAs have become increasingly popular in recent years due to their user-friendliness and high security. Previous attack methods on slide puzzle CAPTCHA mainly face challenges in the following four aspects. (1) The defense mechanisms of previous slide puzzle CAPTCHAs were relatively simple. The current slide puzzle CAPTCHAs have added complex interferences, detected slide trajectory features in the two-dimensional space, and added other defense mechanisms. (2) The previous notch detection methods were difficult to accurately locate the notch positions. (3) Previous attack trajectories were limited to the horizontal dimension, making them impossible to pass the existing detection. (4) The previous attack trajectories were generated by fitting functions, which were relatively simple and could not well simulate the slide characteristics of human users. In response to the above challenges, firstly, we systematically analyze the defense strategies of six most popular slide puzzle CAPTCHAs: Capy, Dingxiang, Geetest, Kyger, Shumei and Tencent. Secondly, for each type of CAPTCHA, we construct a notch detection dataset and a slide trajectory dataset separately. Thirdly, we use advanced YOLOv11 to detect the notch positions and train generative adversarial networks (GAN) to generate attack trajectories. As far as we know, we are the first to generate attack trajectories by directly learning human slide trajectory characteristic. According to the experimental results, this method cracked six most popular slide puzzle CAPTCHAs with success rates of over 90%, and the attack time is only about 5 seconds, indicating that the security of existing slide puzzle CAPTCHAs is facing challenge. At the same time, we propose various methods to enhance the security of slide puzzle CAPTCHAs.
Read the full abstract
Optimization and Control
An Adaptive Short-Step Variant of BroxGD with Linear Convergence under Quadratic Growth
Broximal Gradient Descent (BroxGD), recently analyzed by Richtárik et al. (arXiv:2605.08850v1), is a projection-free sister method to Projected Gradient Descent (ProjGD) for solving $\min_{x\in\mathcal X}f(x)$, where $f:\mathbb R^d\to\mathbb R$ and $\mathcal X\subseteq\mathbb R^d$ is nonempty, closed and convex. The method performs local linear minimization over the feasible set intersected with a Euclidean ball and carefully prescribed radius. Like ProjGD, BroxGD enjoys linear convergence in the smooth strongly convex setting, with the same condition-number dependence. However, the known radius rule leading to linear convergence requires unavailable information and is therefore impractical. We develop an adaptive short-step variant of BroxGD which converges linearly for smooth convex functions under quadratic growth (generalizing strong convexity), without requiring unavailable information. We use $$z_k\in\arg\min_{z\in\mathcal X\cap\mathbb B(x_k,t_k)}\langle\nabla f(x_k),z\rangle,\quad x_{k+1}=\mathrm{Proj}_{S_k}(x_k-\nabla f(x_k)/L),$$ where $L$ is the smoothness constant of $f$, $\mathbb B(x,t)$ is the Euclidean ball centered at $x$ of radius $t$, and $S_k$ is the line segment connecting $x_k$ and $z_k$. The radius is updated by $t_{k+1}=\max\{\|x_{k+1}-x_k\|,t_k/2\}$. Projection onto $S_k$ requires only scalar clipping. Assuming an exact local linear minimization oracle, each update uses one oracle call. For any positive initial radius, we bound all oracle calls to reach objective error at most $\varepsilon$ and obtain $O((L/\alpha)\log(1/\varepsilon))$ for fixed initialization, where $\alpha=\min\{\alpha_0,L\}$ and $\alpha_0>0$ is the quadratic-growth constant.
Read the full abstract
Materials Science
Crystal and magnetic structures of MnSi in zero magnetic field: Neutron diffraction and symmetry analysis
Combining high-resolution wide-angle neutron powder diffraction with magnetic-superspace symmetry analysis, we refine the crystal structure and present, to our knowledge, the first systematic refinement of cubic $P2_13$ MnSi's zero-field magnetic structure across multiple magnetic satellite families. We compare single-arm rhombohedral 3+1D and cubic four-arm 3+4D magnetic superspace-group (MSSG) descriptions with propagation vectors along $\langle111\rangle$. Simultaneous extinction of the two satellite groups around $\{100\}$ dictates a transverse circular-helix model. This suppression and refinements of 18 separately varied satellite-group intensities support this ideal limit with only one adjustable magnetic amplitude. In the ideal single-arm model with MSSG $R3.1'(0,0,g)ts$, the moment has constant magnitude 0.535(3) $\mu_B$ per Mn. An explicit four-arm 3+4D cubic superspace construction presented here combines four interpenetrating helices and, for the illustrated phase choice, has moments ranging from zero to 0.874(5) $\mu_B$ per Mn, with the same root-mean-square moment as the 3+1D structure. Their identical powder profiles prevent the present measurements from distinguishing them. The average crystal structure remains cubic within experimental resolution, providing structural motivation for the cubic magnetic model. The mean lattice parameter, determined with formal relative precision $\sigma_a/a \lesssim 10^{-5}$, exhibits nonmonotonic temperature dependence below the Néel temperature.
Read the full abstract
Combinatorics
Zero forcing is not unimodal
The zero forcing polynomial of a graph counts the number of distinct zero forcing sets of every possible size. In 2018, Boyer et al. asked whether the zero forcing polynomial is always unimodal, and provided some evidence that it is. We answer this question by constructing an infinite family of graphs whose zero forcing polynomials are not unimodal. Our construction also shows that the graph polynomials associated with connected forcing and total forcing are also not unimodal in general.
Read the full abstract
High Energy Astrophysical Phenomena
The Fate of Collisional Neutrino Flavor Conversion: Fixed Points, Period Doubling, and Chaos
Core-collapse supernovae and neutron star mergers are among the most energetic astrophysical events and promising sources of multimessenger signals. Their dynamics and nucleosynthesis depend sensitively on neutrino transport, yet the impact of neutrino flavor evolution in these dense environments remains poorly understood. Predicting the outcome of collisional flavor instability (CFI) remains a challenge, despite progress in local-relaxation methods for fast flavor conversion. Here we construct homogeneous collective-precession states with stationary flavor populations in two-flavor quantum kinetics with collisional relaxation. The method applies to arbitrary energy spectra and energy-dependent collision rates. Our numerical examples exhibit persistent oscillations and a period-doubling route to chaos. Even in these regimes, the stationary populations estimate the time-averaged flavor populations to within $10\%$ in our tests. These fixed points provide a physically motivated input for incorporating CFI into astrophysical simulations, with implications for neutrino heating and ejecta nucleosynthesis.
Read the full abstract
Statistics Theory
Financial Connectedness: Heavy Tails andResponse Attribution
A financial market can appear mainly dependent on its own shocks under a variance account and mainly dependent on external shocks under a cumulative-response account. We establish how these opposite diagnoses arise from the same propagation system. For a given cumulative magnitude, squaring gives greater weight to responses concentrated across sources and time. Source priorities can therefore reverse even in stable Gaussian systems with known parameters. Heavy tails raise a separate question: whether the reported shares have a fixed population basis. One innovation with an infinite second moment removes the full network's finite covariance input. We characterize when normalized shares nevertheless stabilize and construct rank-amplitude connectedness, a cumulative-deviation target without finite first or second moments. It can be consistently estimated under local dependence and identification conditions. With fitted dynamics and rank scatter fixed, changing attribution replaces 3.14 names in the daily Top 10 of forty stocks, selecting recipients with greater external persistence relative to their own adjustment. In five futures contracts, mean external share rises from 28.0% to 55.9%, and 70.9% of recipient-dates switch from own to external dominance.