Number Theory
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Friday 4 September
We give a new Hecke-theoretic method to control the growth of monodromy in $p$-adic families of automorphic Galois representations for unitary groups and quasisplit classical groups.
Group Theory
The Codegree Isomorphism Conjecture for Finite Simple Groups
We prove the codegree isomorphism conjecture for finite simple groups. Thus every nonabelian finite simple group is determined, among finite groups, by its set of irreducible character codegrees.
Classical Analysis and ODEs
On the spectrality of the non-homogeneous golden-mean self-similar measure
We investigate the spectral properties of a class of inhomogeneous self-similar measures, which does not admit a non-trivial infinite convolution structure. A central example is the golden-mean self-similar measure $\mu$, for which the existence of an exponential orthonormal basis in the associated $L^2$-space has remained a long-standing open problem. The usual approach for homogeneous self-similar measures does not apply here, new methods are required. We establish several basic properties of the measure and then carry out a detailed numerical study of the zero set of its Fourier transform. Using a scanning and refinement algorithm that combines uniform grid sampling, quadratic interpolation, and golden-section search, we examine a wide range and find no real zeros of $\widehat{\mu}$, which provides concrete evidence that $\mu$ is very likely non-spectral, suggesting that inhomogeneity may serve as a natural obstruction to the existence of exponential orthonormal bases. To the best of our knowledge, our paper is the first attempt to study the spectrality of such measures through a combined analytic and numerical framework.
Read the full abstract
Signal Processing
Analytical Statistics of Vortex Beams in a Turbulent Channel for OAM-Multiplexed FSO Communications
Orbital angular momentum (OAM) multiplexing can increase the capacity of free-space optical (FSO) communications, whereas atmospheric turbulence causes modal crosstalk and irradiance fluctuations that degrade demultiplexing performance. Analytical modeling is therefore important for characterizing turbulence-induced propagation effects and the demultiplexed port-power statistics of OAM channels. In this paper, we first study the receiver-plane irradiance statistics of vortex beams after propagating through the turbulent channel. The average irradiance is derived using frequency-domain convolution, and a closed-form frequency-domain diffraction kernel is obtained based on extended Rytov theory to evaluate the scintillation index for moderate and strong turbulence. However, receiver-plane irradiance statistics alone are insufficient to describe the performance of OAM-multiplexed FSO communications. We therefore derive the demultiplexed port-power statistics. Specifically, we derive the average port power, modal crosstalk, port-power variance, and cross-port covariance based on the complex Gaussian expansion of general LG vortex fields and the extended Huygens-Fresnel framework. The demultiplexed port-power statistics are then used to evaluate the symbol-error rate (SER) of OAM-multiplexed FSO communications. Numerical results demonstrate that all derived statistics are consistent with those obtained by phase-screen simulations under different turbulence strengths and beam parameters. The resulting SER performance further shows that OAM-multiplexing performance is more sensitive to mode spacing for small receiver apertures than for large apertures.
Read the full abstract
Artificial Intelligence
Model Retirement Creates Reproducibility Risk in Biomedical AI Publications
Background. Large language models (LLMs) are being adopted in biomedical research at a rapid and accelerating pace, yet commercial services that host many widely used models operate under deprecation schedules that can complicate scientific reproducibility. Methods. We searched PubMed for original research articles from 2022 through March 2026 that applied a specific LLM to a biomedical task. An extraction agent identified model names from 61,077 article abstracts with human reviewers validating a subset for extraction accuracy. Extracted model names were normalized to canonical model identifiers. Lifecycle data (release date, retirement date, status) were compiled for the 50 most frequently used models. Results. We identified 8,931 paper-model mentions spanning 5,242 unique publications after restricting the analysis to the 50 most frequently used models. Among these mentions, 77.7% cited a commercial closed-weight model. Overall, 42% involved a model that was already retired by the time of official publication or is scheduled to retire within two years of publication. The median interval from publication to model retirement was 538 days. Conclusion. Many biomedical publications using LLMs are on a trajectory toward computational non-reproducibility after publication. Model deprecation should be treated as a core reporting and preservation issue for biomedical research.
Read the full abstract
Computer Vision
LookThere! Sparse Vision by Reinforced Selection
Vision transformers typically treat every image token as equally important, yet for most tasks in computer vision only a fraction are needed. Adaptive computation methods accelerate inference by choosing which tokens to process, but existing methods struggle at extreme sparsity and require heuristics that may not generalize like token diversity and attention scores. We address these limitations with LookThere, achieving a new pareto frontier in performance-compute trade-offs through an end-to-end reinforcement learning framework that jointly trains a shallow input selector and a deep representation extractor. The selector learns where to look and the extractor learns what to see, together saving computation by selecting only what is worth processing for a given task without relying on auxiliary signals. We show that LookThere only selects the task-specific input, excelling at sparse recognition in high-resolution settings (traffic signs, billiards), and maintaining accuracy with as little as 0.2% of the input. It generalizes across tasks and models, including global recognition (ImageNet classification), local recognition (ADE20K segmentation), zero-shot classification (by distillation), and regression (counting). Across all settings, LookThere surpasses state-of-the-art selection to provide a general and scalable framework for specialized and efficient adaptive computation.
Read the full abstract
Artificial Intelligence
SQL-Zero: Self-Evolving Text-to-SQL
Training a competitive Text-to-SQL agent usually depends on human-annotated natural-language/SQL pairs, which are expensive, domain-specific, and a bottleneck for scaling to new databases. We show it is possible to train a competitive solver with zero annotated pairs. We introduce SQL-Zero, a proposer-solver self-play in which a challenger and a solver start from the same base LLM and the only ground truth is execution against the database itself. The challenger generates SQL pairs calibrated to the solver's current difficulty (targeting "hard but solvable"), and both roles are updated with GRPO in alternating turns, with a template-level repetition penalty on the challenger to prevent diversity collapse. Training on BIRD databases with no labels, self-play improves over the zero-shot base on BIRD dev by 6.6 points at 3B and 7.3 points at 7B. It also scores higher than a matched control trained under the same recipe on human BIRD gold over the same databases, although an exact paired test does not resolve that margin. Transfer depends on scale: at 3B every iteration outperforms the base on unseen Spider databases and under lexical perturbation (Spider-Syn), where it also degrades less than the matched BIRD-gold control, whereas at 7B only the first iteration preserves transfer.
Read the full abstract
Chemical Physics
Voltage-embedded equivariant machine learning potential for open system simulations
Modeling electrochemical interfaces under operational non-equilibrium conditions is vital for energy technologies but remains bottlenecked by the expensive cost of ab initio methods. Current machine learning potentials, largely designed for closed systems under homogeneous electric fields, are limited in open quantum transport applications. To overcome this, we present an E(3)-equivariant graph neural network that embeds voltage bias for open-system simulations. Our approach decouples the system energy and forces into zero-bias and bias-dependent contributions, assigning distinct vector encodings to electrode and scattering-region atoms to capture non-equilibrium conditions. Trained on limited discrete bias data, the model achieves high predictive accuracy and robust extrapolation transferability. When applied to a lithium/water interface, our model successfully captures the field-induced dynamic reorientation of water molecules and reproduces asymmetric electrochemical behavior at different electrodes.
Read the full abstract
High Energy Astrophysical Phenomena
A Differentiable Neural Surrogate for Photon Propagation in Neutrino Telescopes
Large-volume neutrino telescopes infer neutrino properties from Cherenkov light, but simulating the transport of billions of photons through highly scattering ice or water is computationally costly. We introduce candela, a differentiable SIREN neural field that learns the photon Green's function of the IceCube Neutrino Observatory, a cubic-kilometer detector embedded in Antarctic glacial ice. Given a point-like energy deposit and sensor, it predicts the expected photon yield and full arrival-time distribution at the sensor. Complete events are simulated by decomposing charged-particle energy deposits into point-like sources and superposing their predicted sensor responses. Trained on Monte-Carlo simulations, candela generates events $50$–$100\times$ faster than existing methods, with cost scaling only weakly with neutrino energy. It keeps median yields within $2\%$ of the MC expectation and timing distributions at the MC statistical floor across six photon-count decades. The model also provides end-to-end gradients with respect to event parameters and opens a path toward optimizing scattering-medium properties, which often dominate systematic uncertainties in neutrino telescopes.
Read the full abstract
Combinatorics
Stability of independence polynomials of spiders
For a graph $G$, let $i_k(G)$ denote the number of independent sets of cardinality $k$, and let \[ I(G,z)=\sum_{k\ge0} i_k(G)z^k \] be its independence polynomial. Following Brown and Cameron \cite{BrownCameron2018}, a graph is called stable if all zeros of its independence polynomial lie in the closed left half-plane. They proved that every star is stable, but also constructed nonstable trees. They then asked for a characterization of stable trees. In this paper, we extend and strengthen their result by proving that every spider, obtained from a star by arbitrary and possibly nonuniform subdivisions of its edges, has all its independence roots in the open left half-plane. Hence, every spider is stable.
Read the full abstract
Artificial Intelligence
Predicting Spatiotemporal Mobile Sensing-Based PM2.5 Concentrations Using Low-Rank Adapted Spatially Attentive Graph Neural Network
Urban air quality can vary significantly along transit corridors, necessitating high-resolution monitoring. This work introduces a novel mobile-sensing dataset from Surat, Gujarat, India, comprising PM$*{2.5}$ concentrations, meteorological variables (temperature, humidity, wind speed, wind direction), and land-use features. To represent the spatiotemporal data as a graph, two node-definition strategies were used: (i) uniform segmentation (200–400~m intervals) and (ii) DBSCAN clustering to adaptively group dense observations. For each node, rolling mean and standard deviation of meteorological variables were computed. To model this high-dimensional data, we propose a SA-GNN for fine-grained, short-term PM$*{2.5}$ forecasting and hotspot identification. We compared SA-GNN with LSTM, RNN, GRU, and ANN models. These models performed well on low-resolution data but had difficulty capturing rapidly changing patterns in urban air quality. SA-GNN employs cluster-specific GRUs to capture localized temporal dependencies and a Graph Attention Network to learn spatial heterogeneity. This hybrid architecture effectively models rapid fluctuations and complex spatial interactions. On our dataset, SA-GNN achieved $R^2 = 0.95$, RMSE $= 6.8$, and MAE $= 4.2~\si{\micro\gram\per\meter\cubed}$, outperforming all baseline models. Combining spatial clustering with adaptive attention significantly improves forecasting, enabling real-time, fine-grained monitoring and supporting personalized exposure tracking and timely alerts for healthier cities.
Read the full abstract
Physics and Society
Low-Dimensional Phase Diagram of Higher-Order Networked Systems
Higher-order networks exhibit rich critical phenomena that cannot be captured by traditional pairwise models. Here, we develop an analytical dimension-reduction framework that maps higher-order networked dynamics onto an effective low-dimensional system, allowing accurate prediction of tipping boundaries, bistability regions, and the nature of phase transitions. We demonstrate the power of this framework across a range of dynamical processes, revealing distinct effects of higher-order interactions on transition continuity and hysteresis. Furthermore, we find that system resilience exhibits a profound dependence on the alignment between pairwise and higher-order connectivity, with assortative mixing enhancing tipping toward active states. Our findings establish a general theory for understanding the critical transitions in higher-order networks, offering new insights for anticipating and managing systemic risk in complex systems.
Read the full abstract
Exactly Solvable and Integrable Systems
Complex singularities for Burgers' equation with piecewise-continuous initial conditions
There is a body of research devoted to understanding how complex singularities of solutions of nonlinear partial differential equations (pdes) spontaneously emerge at $t=0^+$ and propagate for $t>0$, and how their behaviour affects the solution on the real axis. Despite the importance of the small-time limit in these studies, there is still a lack of understanding of how complex singularities are born at $t=0^+$, including for initial conditions that are not analytic functions of the spatial variable. In this paper, we use Burgers' equation as a prototype nonlinear pde and study the complex-plane singularities for initial conditions that are piecewise smooth. Using matched asymptotic expansions, we show how infinitely many singularities emerge from points of discontinuity in a pattern that can be described using branches of the Lambert-$W$ function. For various initial conditions, we observe how these singularities rearrange themselves to align with the appropriate exactly-described long-time behaviour, including sigmoid-shaped travelling waves, constant-area (triangular wave) similarity solutions and $N$-wave solutions. In terms of Burgers' equation, our small-time asymptotic analysis of the singularity propagation for piecewise-continuous initial conditions illustrates the types of generic behaviours that arise for inner regions when diffusion dominates advection. More generally, this work is a step towards understanding complex-plane behaviour of solutions of nonlinear partial differential equations with non-analytic initial conditions.
Read the full abstract
Computer Vision
Enhancing Multimodal Emotion Recognition via Multi-Feature Encoding and Attention-Based Fusion
Multimodal emotion recognition has attracted growing interest due to its importance in human-computer interaction, remote education, and healthcare. This paper proposes a novel multimodal emotion recognition framework that integrates rich audio and visual feature extraction with an attention-based fusion strategy. For audio, we extract three complementary feature types: semantic embeddings from Wav2Vec2, MFCC features, and statistical acoustic descriptors such as pitch, energy, and rhythm. These are aligned and fused via a BiLSTM to capture temporal dependencies. For video, we propose a ResNet50-BiLSTM architecture that combines deep residual learning and sequential modeling to extract expressive spatiotemporal features from facial sequences. To enhance multimodal synergy, we introduce a feature-level fusion mechanism based on multi-head attention, allowing the model to adaptively weigh contributions across modalities. Experiments conducted on the MELD and IEMOCAP datasets demonstrate that our model significantly outperforms baselines in both accuracy and robustness. Furthermore, ablation studies show that the attention-based fusion strategy significantly improves performance in unbalanced data settings. Our findings suggest that the proposed framework effectively captures diverse emotional cues from speech and visual expressions, and offers a practical and generalizable approach for real-world multimodal emotion recognition tasks.
Read the full abstract
Computer Vision
Retinal OCTA Phenotyping with LLM Reporting for Alzheimer's Disease
Early identification of Alzheimer's disease (AD) remains challenging because established assessment methods can be costly, resource-intensive, or unsuitable for population-scale screening. Optical coherence tomography angiography (OCTA) provides non-invasive visualization of retinal microvasculature, but existing approaches often require diagnostic labels and provide limited measurement-level interpretation. We present an explainable OCTA pipeline that integrates annotation-aware vessel segmentation, layer-specific vascular biomarker extraction, label-free phenotyping, and measurement-grounded LLM reporting. Using 117 ROSE-1 images from 39 subjects, we apply annotation-matched segmentation models to superficial vascular complex (SVC), deep vascular complex (DVC), and combined SVC+DVC representations. The models achieve ROC-AUC values of 0.916-0.970 and Dice scores of 0.695-0.781. Six density and fractal-dimension biomarkers form subject-level profiles for exploratory clustering. Analysis of nine held-out subjects identifies an internally consistent lower-density, lower-fractal-dimension phenotype, although the absence of diagnostic labels prevents clinical interpretation. Reports generated using GPT, Gemini, and Llama are evaluated for measurement grounding, citation faithfulness, and diagnostic caution. Overall, the framework provides a transparent, non-diagnostic connection between retinal vascular measurements, exploratory phenotyping, and evidence-linked interpretation for Alzheimer's research.
Read the full abstract
Information Theory
Decoding Error Probability of the Random Code Ensemble over the Erasure Channel
In this paper, we provide explicit formulas for the average decoding error probabilities of the ensemble of all $(n,M)_q$ codes over the erasure channel under unambiguous, list and maximum-likelihood decoding principles. Moreover, we derive the asymptotic error exponents for the ensemble of all $(n,q^{Rn})_q$ codes.
Read the full abstract
Numerical Analysis
A class of high-order discontinuous-Galerkin methods satisfying infinitely many entropy conditions with provable error estimates and strong convergence for general nonlinear conservation laws
We propose a novel framework for deriving semi-discrete discontinuous-Galerkin (DG) methods using operator semigroups for scalar conservation laws, then apply it to construct a class of high-order OFDG-type schemes [13] satisfying infinitely many local entropy inequalities with general E-fluxes on non-uniform meshes. Such schemes are further generalized to systems of conservation laws in any number of space dimensions by using entropy stable numerical fluxes in the sense of [1]. Finally, we prove optimal error estimates for smooth solutions to nonlinear scalar conservation laws, and prove strong convergence for discontinuous solutions to strictly convex conservation laws via compensated compactness.
Read the full abstract
Combinatorics
Small graphs without power-of-two cycles: a lower bound of 24, a correction to a construction of Exoo, and explicit bounds for f(k)
The Erdos-Gyarfas conjecture states that every graph with minimum degree at least 3 contains a cycle whose length is a power of two. We prove by a SAT-based exhaustive search, certified by DRAT proofs, that every graph with minimum degree at least 3 on at most 23 vertices contains a cycle of length 4 or a cycle of length 8; consequently any counterexample has at least 24 vertices, improving the previously published bound of 16, and the smallest graph of minimum degree 3 with no 4-cycle and no 8-cycle has exactly 24 vertices. We show that the lemma underlying Exoo's 450-vertex construction for the bound f(5) at most 450 is false: the Tutte-Coxeter graph contains 8-cycles alternating between outer and chord edges, and the graph as specified contains 32-cycles. We repair the construction and verify the corrected graph, so the bound stands. We also give an exact window calculus for vertex-replacement constructions, prove that f(k) is at most 15 times the order of the smallest known cubic graph of girth 2 to the power (k-2) plus 1 for all k at least 4 (in particular f(6) is at most 32640, the first bound for f(6)), and show that Exoo's 78-vertex witness for f(4) at most 78 is optimal among gadget designs on bases with at most 12 vertices. All graphs, scripts and certificates are archived at doi:10.5281/zenodo.22180583.
Read the full abstract
Algebraic Geometry
Heisenberg Equivariant Compactifications of Rational Homogeneous Varieties
Let $G/P$ be a complex projective rational homogeneous variety of dimension $2m+1$. We prove that $G/P$ is an equivariant compactification of the Heisenberg group of dimension $2m+1$ if and only if it is isomorphic to either an adjoint variety, or the 3-dimensional smooth quadric $Q^3$, or a product $\mathbb{P}^{2m+1-d} \times Y$ with $1 \leq d=\dim Y \leq m$, where $Y$ is a product of cominuscule varieties.
Read the full abstract
Dynamical Systems
On the Hausdorff measure of projections of self-similar sets
For $0<d\leq1/4$, let $\mathcal{C}(d)$ be the four-corner Cantor set with Hausdorff dimension $s_d=\log 4/\log(1/d)$. Peres, Simon, and Solomyak, as well as Mattila, asked for which $d$ the measure $\Hau^{s_d}(p_\theta(\mathcal{C}(d)))$ is positive for almost every direction $\theta$. It was open for the range $1/9\leq d\leq1/6$. In this paper, we show that, for $\delta< d\leq1/6$, there is a set $\IP(d)$ of positive Lebesgue measure such that for almost every $\theta\in\IP(d)$, $\Hau^{s_d}(p_\theta(\mathcal{C}(d)))=0$, where $\delta=0.155124983896014\ldots$ is the unique zero in $(1/9,1/6)$ of the polynomial $P(d)=1-7d+3d^2+4d^3-2d^4-d^5$.
Read the full abstract
Information Theory
The Cross-Correlation Distribution of the Niho-Type Decimation $d=4(2^m-1)+1$
The cross-correlation problem is a classical problem in sequence design. In this paper, we determine the cross-correlation distribution of the Niho-type decimation $d=4(2^m-1)+1$ over $\mathbb F_{2^{2m}}$ for every positive integer $m$. With $q=2^m$, this is equivalent to determining the distribution of the number of roots in $U_{q+1}$ of $f_a(x)=x^7+ax^4+\bar a x^3+1$ for $a\in\mathbb F_{q^2}$, where $U_{q+1}=\{x\in\mathbb F_{q^2}:x^{q+1}=1\}$. The main difficulty is to count the four-element subsets of $U_{q+1}$ that may occur as root sets of $f_a$. We normalize such subsets by their product and study the resulting condition through the associated resolvent. This reduces the required enumeration to several equations over $\mathbb F_q$, which are evaluated by character sums and Kloosterman sums. The remaining mixed Kloosterman sum is related to a Kloosterman sum over $\mathbb F_{2^{2m}}$ and is evaluated using a theorem of Carlitz. Consequently, we obtain explicit formulas for all the frequencies in the cross-correlation distribution.
Read the full abstract
Classical Analysis and ODEs
Primes are Complete for a Class of Cantor Spectral Pair
We study the complete number problem for a class of self-similar spectral measures on the real line. We say that a real number $t$ is complete for a spectral pair $(\mu,\Lambda)$ if $(\mu,t\Lambda)$ is also a spectral pair. In this paper we consider the partial homogeneous Cantor measure $\mu_{N^r,\mathcal D}$ and its canonical spectrum. Our main result fully characterizes all prime powers $p^n$ that are complete for $p>N$. In particular, we obtain a necessary and sufficient condition for an arbitrary prime $p$ to be complete. {We further establish a connection between the problem and certain Diophantine equations}, including the Nagell–Ljunggren equation, and obtain complete number results for a class of related models. As applications, known solutions of the Nagell–Ljunggren equation provide complete and incomplete prime-power examples.
Read the full abstract
Software Engineering
Beyond Code Generation: Reliability, Verification, and Cost Economics in the Agentic Software Development Lifecycle
AI coding systems are moving from autocomplete and chat toward agents that can inspect repositories, edit multiple files, run tools, write tests, open pull requests, and work for long periods with limited supervision. This capability changes the bottleneck in software delivery. Recent field studies show meaningful gains in coding activity, but newer evidence also shows that those gains attenuate sharply between writing code and shipping reliable software. Review, integration, testing, security, deployment, and production operations remain constraining stages, while the economics are shifting from predictable per-seat licensing toward variable token, tool, sandbox, CI, and rework costs. This paper synthesizes peer-reviewed software-engineering research, university studies, benchmark audits, production reports from major technology companies, developer telemetry, and cost-management evidence released primarily from 2024 through September 2026. No new model experiment is claimed; numerical findings remain attributed to their original studies. The synthesis proposes four engineering concepts: the Agentic SDLC Throughput Paradox, Production-Qualified Change (PQC), the Verification Tax, and an Agentic SDLC Control Plane that allocates autonomy subject to cost, reliability, and human-attention budgets. An evidence-based horizon then maps today's supervised agents to future policy-bounded software factories. The central research question shifts from how much code an agent can generate to how much production-qualified value an engineering system can deliver per dollar, per reviewer-hour, and per unit of operational risk.
Read the full abstract
Software Engineering
How Developers Discuss Generative AI: A Longitudinal Study of the Visual Studio Code Community
Generative AI tools such as GitHub Copilot, ChatGPT, and coding agents have rapidly become part of everyday software development, yet little is known about how mainstream open source communities discuss them in practice. This paper presents a longitudinal analysis of generative-AI-related discussions in the Visual Studio Code (VS Code) GitHub repository, using 43,806 candidate issues created between January 2021 and June 2026. To improve corpus quality, we combined keyword retrieval with semantic relevance filtering, yielding a filtered corpus of 25,227 AI-related issues. We applied BERTopic to the retrieved corpus to identify discussion topics, using the filtered corpus for theme validation and a robustness re-clustering, and analyzed their evolution over time using monthly prevalence and Mann-Kendall trend tests. The results show that developer discussions are dominated by practical concerns regarding the operation of AI-assisted development environments, including agent management, configuration, reliability, authentication, and billing, whereas risks frequently emphasized in survey-based studies, such as hallucination and licensing, rarely surface in this venue. This suggests that discussions of generative AI in the VS Code issue tracker primarily focus on operational aspects of AI-assisted software development. Furthermore, discussions evolved from AI-assisted code completion toward conversational and agent-based development, reflecting the increasing integration of generative AI into software development workflows. These findings suggest that GitHub Issues provide a practical, workflow-oriented perspective on generative AI that complements survey-based studies of developer perceptions.
Read the full abstract
Human-Computer Interaction
Beyond Prompt-to-App: Accountable Translation in Teacher-Facing Agentic Authoring
Natural-language app builders let domain experts create software, but their pipelines transform professional intent across compilation, generation, checking, and approval. We report a bounded trace study of a teacher-facing agentic authoring system. Evidence comprises six eligible build attempts across three accounts; a separate corpus of 37 workshop units from 23 display names contextualizes commitments without person-level linkage. Compiled specifications added governance requirements, while downstream representations sometimes normalized case-specific learning relations. Two drafts met a stored package/security threshold despite analyzer reservations and unresolved correspondence to their briefs; four attempts in one account produced no usable payload, and repair messages did not translate internal terms into domain-legible revisions. We develop accountable translation as an analytic framework for making consequential changes attributable, inspectable, scoped in validation, and contestable. It extends HCI accounts of traceability and end-user debugging by locating professional authority and repair rights across heterogeneous technical and organizational handoffs.
Read the full abstract
Artificial Intelligence
Train What You Deploy:Token-Faithful Post-Training of a Production Coding
Existing post-training pipelines for coding and terminal agents suffer severe token and control fidelity errors: simplified training environments mismatch production deployments, and offline token reconstruction from agent logs distorts original prompts and conflates policy calls with background model operations. We present a fidelity-aware training coupling framework that retains trainer-side sampling over original prompts, eliminates spurious model calls via a negotiated training protocol, and restricts loss computation to verifiable token spans with closed-failure guarantees. We further propose Certified Divergence Proximal Policy Optimization (C-DPPO), which establishes tight two-sided TV certification bounds, adaptive-K rules, budget-aware sequence guarantees, and error-robust policy masking atop standard DPPO. Evaluated on matched Baize5B and Baize10B models with identical training and test protocols on TMax-100, C-DPPO yields a consistent +3.0-point performance gain over standard DPPO across model scales. Certificate audits validate the reliability and full operational coverage of our certified training pipeline.
Read the full abstract
Combinatorics
On the exact quantum chromatic number of generalized Johnson graphs
The quantum chromatic number is a fundamental parameter in the study of nonlocal games, capturing the extent to which entanglement can improve performance in distributed tasks. In this paper, we investigate the quantum chromatic number of generalized Johnson graphs. By constructing modulus-one orthogonal representations, we obtain general upper bounds on their quantum chromatic numbers. We further analyze the smallest eigenvalue of these graphs. Combining the resulting Hoffman-type lower bounds with the upper bounds obtained from orthogonal representations, we determine the exact quantum chromatic numbers of two infinite families of generalized Johnson graphs. Finally, applying a forbidden-distance theorem for binary codes, we show that the classical chromatic numbers of these families grow exponentially with $n$, whereas their quantum chromatic numbers grow linearly. These families exhibit an exponential separation between the classical and quantum chromatic numbers.
Read the full abstract
Computation and Language
Controlling and Assessing Appropriate Persona Use in LLM-based Dialogue Generation
In persona-based dialogue generation (PDG), LLMs often overuse persona attributes by incorporating them regardless of dialogue context, resulting in unnatural responses. Despite its practical significance, the underlying causes remain unexplored, with no method to mitigate this problem or metric to assess the appropriateness of persona use. To address these issues, we first conduct a comprehensive analysis of LLM-based PDG, revealing that LLMs exhibit a systematic bias to incorporate all given persona attributes, and that existing metrics fail to capture contextual appropriateness. Building on these findings, we propose Self-CONtrastive Persona Overuse Suppression (SCONPOS) to mitigate overuse by directly intervening in LLMs' internal representations at the prompt encoding stage, without requiring any response generation. We further propose the Persona Appropriateness Score (PAS), a novel metric that penalizes both overuse and underuse. Experimental results demonstrate that SCONPOS systematically reduces overuse, and PAS captures the contextual appropriateness of persona use.
Read the full abstract
Differential Geometry
A Bernstein problem for translating solutions to the mean curvature flow
We study entire graphical translating solutions of the mean curvature flow, \[ \operatorname{div}\!\left(\frac{\nabla G}{\sqrt{1+|\nabla G|^2}}\right) =\frac{1}{\sqrt{1+|\nabla G|^2}} \qquad\text{in }\mathbb R^N. \] Every such graph is mean-convex, since its mean curvature is the vertical component of its unit normal. In dimension two, mean-convex translating solitons are convex, and an entire graphical translator is therefore the rotationally symmetric bowl soliton. In higher dimensions Wang constructed non-rotational entire convex translating graphs. We prove that a further loss of rigidity occurs at the Bernstein dimension: for every $N\ge 8$ there exists a one-parameter family of entire graphical translators that are mean-convex but not convex. The construction starts from the Bombieri–De Giorgi–Giusti (BDG) entire minimal graph in $\mathbb R^8$ and develops a singular perturbation theory for the translator equation around it. The main new feature is a transition layer near Simons' cone: the translating term breaks the odd symmetry of the minimal graph, and after a suitable recentering the matching problem is governed by a parabolic inner equation. A detailed analysis of this layer, together with weighted Jacobi theory on the BDG graph and global barriers, yields the desired entire solutions.
Read the full abstract
Quantum Physics
QMClaw: A Scalable General-purpose Framework for Quantum Measurement and Control
As quantum computing continues to scale, quantum measurement and control (QMC) are increasingly constrained by calibration workflow complexity and by requirements for low-latency execution, robust exception handling, and traceable workflow governance. Existing frameworks for QMC are specialized and task-specific, while language-model-based agents for QMC suffer from excessive latency and cannot satisfy the strict timing and control-density demands of large-scale quantum systems. Here we propose QMClaw, a general, workflow-oriented framework for QMC built, featuring a local-first, tool-governed, robust architecture. At its core is a RuleEngine-centered control layer that processes structured context, performs rule-based state transitions, and generates execution plans for typical calibration workflows. Language models are used only for natural-language interaction, high-level task understanding, and exception support, keeping the critical fast path efficient. We implement a single qubit tune-up workflow as a demonstration and validation using real quantum device dataset. We also prove that the framework achieves quantitatively acceptable levels in terms of resource cost, LLM calling times and decision latency, enabling its practical deployment in large-scale quantum qubit measurement and control scenarios. This work presents a general workflow-oriented framework for QMC and provides evidence that rule-centered architectures are a promising design choice for scalable quantum-system calibration.