Machine Learning (Statistics)
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Monday 9 November
Historical decisions determine which outcomes are observed, potentially concealing a predictor's largest errors. Even unlimited data from the same observation process can therefore leave population tail risk unresolved. Here we derive sharp bounds on conditional value-at-risk, the average loss in a specified upper tail, for binary outcomes and bounded losses. These bounds span exactly the risks compatible with the observations and a stated limit on how outcome odds differ between observed and hidden cases at the same covariate values. Under stated regularity conditions, calibrated statistical certificates cover these ranges simultaneously across candidate predictors, supporting risk limits and model comparisons while separating statistical uncertainty from persistent ambiguity. For a fixed predictor and loss threshold, we derive a minimum-cost allocation for reducing excess-loss ambiguity by prioritizing hidden outcomes according to ambiguity removed per unit cost. This allocation treats audited population contributions as known and retains the original sensitivity bounds separately for audited and still-hidden outcomes. In controlled masking experiments with known loan repayments, recovering about 16% of hidden labels removes a median 61% of estimated fixed-threshold ambiguity, compared with 16% for random recovery. These results connect population tail-risk claims to the available evidence and identify where additional outcomes can sharpen their assessment.
Software Engineering
Will AI mess my RTL? A snapshot-based approach to verifying AI-written RTL
With the growing interest in using AI (Artificial Intelligence) for RTL (Register-Transfer Level) hardware development, robust and comprehensive verification has become more important than ever. As Large Language Models (LLMs) increasingly assist in creating and modifying RTL designs, ensuring that these changes preserve existing functionality is paramount for building trust in AI-driven workflows. We present SVApshot, a fully automated framework that leverages LLMs to generate and systematically attempt to correct SystemVerilog Assertions (SVA) for RTL modules, completely removing the human from the generation loop. SVApshot introduces a novel snapshot methodology that captures the current functionality of a design as a comprehensive set of formal assertions, creating a regression suite for validating future changes–whether made by humans or AI. The framework features automated duplicate detection, assertion set expansion, and iterative repair of failing assertions using LLM-guided debugging with formal verification feedback. Our experimental evaluation across diverse RTL modules demonstrates high coverage for complex modules, with successful detection of both manually-injected and AI-introduced bugs. Additionally, this framework opens the door for automatic generation of formal testbenches to benchmark LLMs for hardware design, moving beyond binary pass/fail metrics towards nuanced assertion-based scoring.
Read the full abstract
Nuclear Theory
On the Method of Projection onto Localized States for Describing Reactions and Clustering
The method of projection onto localized states intended for describing reactions and clustering is examined. In particular, the method offers a simple extension of the no-core shell model to continuum- and clusterized bound states. A critical analysis of a previous realization of the method is provided. A way to minimize the number of basis functions required for convergence is given. An easy way to obtain exactly unitary and symmetric $S$ matrix from calculations is pointed out. The method is tested in one- and two-channel $np$ scattering problems with the Daejeon $NN$ interaction. It is shown that excellent convergence can be reached even when only oscillator states with low number of excitation quanta are retained.
Read the full abstract
Software Engineering
WasmLens: Automatic Glass-Box Instrumentation for Composed WebAssembly Applications
The WebAssembly Component Model enables language-independent composition of independently developed components, but complicates observability. Existing tracing support flattens execution hierarchy by attributing every span created inside a composition to a single host-level parent context, so execution across component boundaries cannot be reconstructed. The problem is amplified because components are typically deployed as black boxes whose internals cannot be modified or inspected by the system integrator. To address these limitations, we propose a revised wasi:otel/tracing design in which the runtime maintains a per-component span stack, restoring parent-child relationships across component boundaries without changing component interfaces or source code. This change has since been adopted upstream. Building on this design, we present WasmLens, an automatic instrumentation tool that derives instrumentation points from a composition graph and injects tracing middleware into already composed binaries, yielding end-to-end traces for black-box compositions with no developer configuration. This automatic instrumentation raises mean request latency from $238$ $\mu$s to $454$ $\mu$s across six component boundaries. Raising the expected number of boundary crossings from 6 to 11 further increases latency to $514$ $\mu$s, an incremental cost of $12$ $\mu$s per crossing, of which $10.4$ $\mu$s is attributable to instrumentation. These results show that instrumentation overhead is additive in composition depth rather than proportional to baseline execution time.
Read the full abstract
Computation and Language
A local language model layered with rule-based and neural detectors for clinical text de-identification: an open-source toolkit and evaluation
Objectives: To measure whether a locally hosted large language model (LLM), stacked with deterministic and clinical named-entity recognition (NER) detectors, finds more protected health information (PHI) than open tools, and at what precision. Materials and Methods: DeidSharp, an open-source .NET toolkit, combines the findings of pluggable detectors (patterns, a clinical NER model, a local 26B mixture-of-experts LLM) and fails closed on any detector error. Configurations fixed on tuning halves were compared, under a prespecified protocol, with deid, Philter, the Stanford/Penn de-identifier, OpenAI Privacy Filter and GLiNER on reporting halves of PhysioNet's nursing notes (889 spans) and semi-synthetic MIMIC-IV-Note discharge summaries (29,023 spans), and on synthetic ASQ-PHI queries, with paired patient-level bootstrap intervals. Results: In the nursing notes, against Stanford/Penn, the recall-first stack found 98.9% of PHI (+2.8, 95% CI 1.4 to 4.4) at 9.3 points lower precision; the precision-first stack, without NER, found 96.5%, matching Stanford/Penn, at 19.6 points higher precision (83.2%). In the discharge summaries, both stacks led Stanford/Penn by more than 20 points (recall-first 95.3%, +21.5, 20.5 to 22.4), almost entirely on generated untyped names; on typed labels all three exceeded 99%. All four prespecified comparisons held. No configuration beat deid significantly on its home corpus; Philter found more discharge-summary spans than the recall-first stack at 23 points lower precision. Discussion: The LLM was the main detector and NER added place names; asked to verify its own findings, the LLM discarded a quarter of true PHI. Conclusion: A local LLM layered with deterministic detectors offers recall-first and precision-first operating points without sending text off-site.
Read the full abstract
Differential Geometry
The Szegő log-term and local rigidity of Kähler–Einstein polarizations of compact Hermitian symmetric spaces
We prove local rigidity of the vanishing Szegő log-term for integral Kähler–Einstein polarizations of a product of irreducible compact Hermitian symmetric spaces. Near the invariant metric, vanishing of the log-term forces the Kähler form to lie in its biholomorphic orbit. The argument combines the first variation of the Kempf distortion function with a central-positioning procedure, reducing rigidity to polynomiality of spectral multipliers of the Laplace operator and the Berezin transform. For integral Kähler–Einstein polarizations, a common-pole obstruction then rules out all infinitesimal deformations transverse to the automorphism orbit.
Read the full abstract
Combinatorics
On Symmetric Chain Decompositions for Bruhat–Chevalley Posets
We consider Stanley's problem on describing symmetric chain decompositions for the Bruhat–Chevalley posets of partial flag varieties and, more generally, Schubert varieties. These posets describe the containment relations of cell closures in the partial flag variety $G/P$ of a connected complex reductive algebraic group $G$. A famous instance of this problem asks whether the poset of partitions that fit inside a rectangle ordered by containment admits a symmetric chain decomposition (SCD), corresponding to the case of Grassmannian Schubert varieties in type $A$. Our first main result applies the “complete BP decomposition" of Richmond–Slofstra to show that the SCD problem for any rationally smooth Schubert variety reduces to the (generalized) Grassmannian case, in any semisimple type. We further study the SCDs provided by codes for Bruhat–Chevalley posets in type $A$. These SCDs can be viewed as a consequence of a particular form of complete BP decomposition. In particular, we show that a poset of permutation-labeled partitions in a rectangle corresponding to the quotient $\mathcal{S}_n/\mathcal{S}_{n-k}$ admits an SCD. Further, posets of weak compositions inside a rectangle are realized as Bruhat–Chevalley intervals with associated codes and SCDs.
Read the full abstract
Applied Physics
The Weight Is in the Reading: Hidden States in Memristors, Memlumors, and Electrochemical Transistors
Physical synapses are routinely described by a single number, the weight, that is identified with a conductance, a current, or a luminescence yield. This identification conflates two different objects: the internal state that the device stores, and the signal obtained when that state is read. Building on Chua's formulation of memristive systems, we separate the state variable, which evolves with the excitation history, from the observation map that converts it into a measured signal, and we define the weight as the outcome of a specified read protocol. Three platforms show the consequences: a memristor, in which the same voltage writes and reads; a memlumor, in which light both modifies the state and produces the photoluminescence that reports it; and an organic electrochemical transistor (OECT), in which the gate writes and the channel reads. In the transistor, two states prepared by different histories give the same channel current at a specific drain bias and become distinguishable again when the bias is changed, while the stored charge persists and the gate current continues to separate them. Published experiments provide precedents: a drain-bias-dependent cancellation of the OECT transient, and trapped charges in perovskite memlumors that are invisible to conventional time-resolved photoluminescence but revealed by a read-write-read sequence. A synaptic weight is therefore not a property of the material alone but of the material together with its readout, and reports of synaptic behavior should specify what is stored, what is measured, and how one is inferred from the other.
Read the full abstract
Multiagent Systems
Multi-Agent Screening: What Agreement and Acceptance Cannot Tell You
When AI agents screen items for a costly human assessor, as in agentic search, e-discovery and AI reviewing of papers, only what they pass gets judged. We show what each kind of evidence can establish about their errors, even with unlimited data. Agreement shows that agents concur, not that they are right, however many there are. Acceptance by the assessor gives precision but only a lower bound on recall. Knowing which agents found which relevant items gives an upper bound on recall if agents' finds are not negatively correlated, and caps what one more agent can add if agents are interchangeable. Without assumptions about the agents, recall can be pinned down only by judging items they did not pass: a random sample of these estimates how many relevant items were missed, to any chosen accuracy, and the sample needed does not grow with the number of agents.
Read the full abstract
Instrumentation and Methods for Astrophysics
Ångström Exponent Calculation Using Wide-Field Stellar Photometry
The FRAM (F/Photometric Robotic Atmospheric Monitor) telescopes can measure the aerosol extinction at optical wavelengths using wide-field stellar photometry at the Cherenkov Telescope Array Observatory (CTAO) sites. The Angstrom exponent describes the wavelength dependence of aerosol extinction. In this work we extend the FRAM method from a single $B$-filter (440 nm) analysis using the Tycho-2 catalogue to a simultaneous three-filter analysis in $B$, $V$, and $R$ (440, 540, and 620 nm) using Gaia DR3 synthetic photometry, which provides photometry directly in the Johnson–Cousins system without passband transformations. The exponent is derived from the vertical aerosol optical depth (VAOD) for the B–V, V–R, and B–R filter pairs at three sites over 2021–2022. The multi-filter VAOD shows a systematic filter-dependent ordering that is inconsistent with physical expectation and produces negative exponents. It is not currently possible to determine which filter yields the correct VAOD, and none of the four processing configurations we tested resolves the ordering. We introduce an intercept correction as a practical remedy that yields physically plausible exponents. At the two southern Paranal sites (CTAO-S0 and CTAO-S1) the exponent distributions are unimodal, with mean $\alpha \approx 0.84$ (B–V) and $0.65$ (B–R) at CTAO-S0 and $0.60$ and $0.41$ at CTAO-S1. At CTAO-N (La Palma) the distributions are bimodal, with $\alpha \approx 0.2$ (B–V) during episodic Saharan dust intrusions and $\alpha \approx 1.3$ under clean background conditions. AERONET sun-photometer measurements on the dates common to both instruments show the same two populations, providing qualitative support for the FRAM method.
Read the full abstract
Cryptography and Security
Checking DER Commands Against Grid Context Catches Credentialed Attacks Static Checks Miss
Authorization of distributed energy resource (DER) commands judges a command by who sent it and whether its value has been seen before. An attacker holding a legitimate credential who writes ordinary values passes both tests. We add a check that judges each DER write against the grid's operating context: the bus voltages around the target, the hour, the day- ahead price, system load, and irradiance, compared with what operators wrote under similar conditions in the feeder's own benign history. We evaluate it live on the IEEE 123-bus and 8500-node feeders, driven by a measured 2024 year of price, load, and irradiance, against nine attack scenarios. A lookup trained on the same history catches none of the writes of the four scenarios that send values already familiar on the targeted device; the context check refuses 91.7% of insider-abuse writes and 64.4% of DER-aggregation writes while denying 1.28% of benign writes. In an offline test that replays recorded moments with curtailment setpoints drawn from the benign range, issued at a positive price, it refuses 47.3% and 50.5% on the two feeders, against 18.5% and 16.0% for the same check without context; live, the few setpoints tested were mostly shallow and were mostly allowed. The context check complements the lookup.
Read the full abstract
Machine Learning
Fisher-Glass Learning: Classification Bounds and Subpopulation Audits
Accurate classifiers can fail when background cues change without altering the correct labels. Predictive performance alone does not quantify how much task information remains after accounting for such nuisance variation. Here we extend the Fisher-glass framework to connect nuisance-adjusted information with local classification difficulty and subpopulation auditing. We derive lower bounds on local minimax excess classification risk under explicit distinguishability and decision-separation conditions, alongside statistical audits accounting for estimation error and regularization. On Waterbirds, a feature-based diagnostic of task-background alignment selects an accurately classified group, with selection frozen before held-out evaluation. Across five training seeds, background replacements preserving the extracted bird foreground and labels yield 93.2% accuracy when background type is preserved and 40.6% when it changes, a controlled difference of 52.6 percentage points. Separately, search using task labels and a fixed nuisance direction identifies an independently evaluated population with 98.8% membership-weighted accuracy and low estimated projected predictive Fisher information. Training and acquisition experiments show that information gains and predictive gains can diverge. These findings support joint assessment of predictive performance, information under declared nuisance models and controlled sensitivity when diagnosing vulnerability and evaluating model improvements.
Read the full abstract
Mathematical Physics
One-Particle Resonances and Eigenstate Thermalization in Periodic Free-Fermion Lattices
We study eigenstate fluctuations of the particle number in the nearest-neighbour free-fermion model on the periodic cubic lattice $(\mathbb{Z}/L\mathbb{Z})^d$. For every fixed dimension, every particle number, every subvolume, and every normalized energy eigenstate, we prove a bound of order $NL^{d-1}$ for the particle-number variance. The corresponding eigenstate-thermalization estimate is of order $L^{d-1}/(N\eta^2)$ for fluctuations of relative size $\eta$. In particular, at positive particle density the bound decays as the inverse linear size. The proof decomposes the centered particle-number operator according to one-particle energy transfer and treats the zero- and nonzero-transfer components separately. The nonzero-transfer part is controlled by a sharp resonance bound, proved from Mann's theorem on relations among roots of unity, while the zero-transfer part is estimated by a Hilbert-Schmidt argument. These bounds are optimal in every dimension: the maximal nonzero-gap multiplicity has order $L^{d-1}$, and suitable many-body eigenstates realize matching variance lower bounds and show that the concentration bound cannot be improved in estimates uniform in the tolerance parameter.
Read the full abstract
Differential Geometry
Finiteness of fundamental groups for manifolds with sign-changing Ricci curvature
Let $M$ be a complete open $\kappa$ noncollapsed Riemannian manifold and have Ricci curvature bounded from below. If the negative part of its Ricci curvature goes to zero sufficiently rapidly at infinity in terms of the volume noncollapse constant and dimension $n$, then the fundamental group of $M$ is finite. When the Ricci curvature is nonnegative, the result agrees with the classical conclusion of [Peter Li, 1986].
Read the full abstract
Computer Vision
GeoScope: Training-Free Program Execution beyond the Counting Horizon for Ultra-High-Resolution Remote Sensing Question Answering
In ultra-high-resolution (UHR) remote sensing images, the objects that questions count in a region or relate by direction often span only a few pixels. Existing methods mainly decide where a vision-language model (VLM) should look and trust its answer in the chosen view. We show that this assumption fails for counting, and measure VLMs as one calibrates a sensor. All eight open VLMs tested fall increasingly short once a view holds more than a certain number of objects, about five for most on DOTA aerial images. This counting horizon narrows as views grow and objects appear smaller. An empirical power-law response model predicts the undercount a split recovers and the spurious counts it adds, and supports the rule of splitting only views beyond the horizon. In GeoScope, a text-only language model parses the question into a typed spatial program that fixes the region and its resolution. The VLM only grounds and counts within its operating range, and directions are computed explicitly from the grounded boxes. To our knowledge, it is the first UHR method to split views by a measured limit of the VLM. Without training, it improves Qwen3-VL-8B by 16.3 points on regional counting and 10.5 on spatial relations in XLRS-Bench, and has the highest reported accuracy, trained methods included, on regional counting and on the whole benchmark by question. On UHR-Micro it improves global and regional counting by 16.8 and 36.2 points, 15.6 of the former from splitting. On MME-RealWorld counting it gains 11.4 points, and we show that an option that is almost never correct yet often chosen distorts comparisons there. Project page: https://avalon-s.github.io/GeoScope/
Read the full abstract
High Energy Physics (Theory)
On Thermodynamic Role of the Gauss-Bonnet Topological Term
We revisit the thermodynamic implications of the Gauss-Bonnet topological term in four spacetime dimensions. It has been previously claimed in the literature that this term, although it does not modify the equations of motion, shifts the black hole entropy by a contribution proportional to the Euler characteristic of the horizon. We argue that this conclusion is sensitive to the choice of the boundary action on the horizon. We then identify a choice for which the Wald entropy expression naturally reproduces the Jacobson-Myers entropy, even for non-stationary black holes. We also show that there are no further contributions to conserved quantities from smooth outer boundaries, regardless of the choice of boundary conditions there and of the asymptotic behaviour of the spacetime. Moreover, we demonstrate that the topological Gauss-Bonnet term changes the critical temperature of the Hawking-Page phase transition, opening, in principle, a way for detecting its presence in the gravitational action. In fact, for a positive Gauss-Bonnet coupling constant, we observe a zeroth-order Hawking phase transition at high temperatures and pressures. While our discussion focuses on the case of the Gauss-Bonnet term in 4D, the results are more general and qualitatively apply to other Lovelock Lagrangians in their respective critical dimensions.
Read the full abstract
Optimization and Control
When is Polynomial Optimization SoS-Exact? Sufficient Regularization and Strict Gram Representations
Global minimization of multivariate polynomials is computationally difficult in general, and nonnegativity does not necessarily imply a sum-of-squares (SoS) representation. We show that adding sufficiently strong Euclidean regularization to a polynomial provides a general mechanism for overcoming such challenges. Namely, for any multivariate polynomial $P$ of degree at most $p$ with a nonzero linear coefficient, and any even integer $2d>p$, we prove that $m_\sigma(s)=P(s)+\frac{\sigma}{2d}\|s\|_2^{2d}$ is SoS exact for all sufficiently large $\sigma$. Furthermore, the shifted polynomial $m_\sigma-\min m_\sigma$ admits a positive-definite Gram representation on appropriately-shifted nonconstant monomials. Consequently, the global minimum and minimizer can be recovered exactly by a single degree-$d$ semidefinite programming (SDP) representation, whose size is polynomial in the dimension for fixed $d$, without requiring higher levels of the SoS hierarchy. The proof develops a degree-independent mechanism based on the Bregman remainder of the regularizer: its strict Gram margin absorbs the lower-degree polynomial perturbations. We further establish an analogous exact SDP regime for polynomial optimization over sufficiently small Euclidean balls, connecting large regularization with small-radius trust-region problems. Finally, we investigate how the algebraic structure of the regularizer governs the existence of a strict Gram margin, extending the mechanism beyond Euclidean norms while identifying limitations of coordinate-separable regularization.
Read the full abstract
Machine Learning
Generalized Adaptive Boosting and the Geometry of Mistakes
Not all mistakes are alike. For example, in medical diagnosis, false positive and false negative mistakes have a different impact and should not be aggregated. By revisiting the classical theory of boosting through this lens, we introduce the notion of $Z$-learner, which returns weak hypotheses achieving joint false-positive/false-negative guarantees $(\mathrm{FP},\mathrm{FN})$ from some set $Z \subseteq [0,1]^2$. This formulation generalizes the standard notion of $\gamma$-weak learning–recovered by taking $Z=\{(z_1,z_2): z_1+z_2<1/2-\gamma\}$–and allows for more flexible notions of weak learnability that account for the impact of the distribution on the achievable learning guarantees. This raises a natural question: which sets $Z$ are boostable? We answer this by giving a complete geometric characterization that unifies, as special cases, both the classical boosting dichotomy and the recent multi-objective extension of Bressan et al. (COLT 2025). We also present an adaptive algorithm for boosting $Z$-learners without relying on the prior knowledge of $Z$, generalizing AdaBoost's ability to handle unknown margin parameters. Finally, we demonstrate that $Z$-learners can achieve confidence amplification. However, the analysis is significantly more involved than in classical boosting and relies on the centerpoint theorem from discrete and computational geometry.
Read the full abstract
Machine Learning
TACTUS: A Robotic Palpation System for Spatially Heterogeneous Hardness Mapping of Woody Breast Myopathy
Woody breast (WB) myopathy is a spatially heterogeneous fibro-degenerative condition of the broiler Pectoralis major muscle. Existing mechanical reference methods rely on single-point or sequential compression measurements, limiting characterization of spatial hardness variability. This study presents TACTUS (Tactile Adaptive Conformal 3D Uniform Sensing), an automated tactile array that generates thickness-normalized, three-dimensional hardness maps at 20 anatomically distributed locations within approximately 30 s. TACTUS integrates 20 independently motorized tactuators with force-sensing resistors and encoder-driven linear actuators, alongside a LiDAR depth camera for surface profiling and adaptive compression. The system was evaluated using 100 broiler breast fillets across four WB severity classes (Normal, Mild, Moderate, Severe). TACTUS identified significant force differences among severity classes (ANOVA, p < 0.001), with the strongest differentiation in the cranial region and a significant caudal effect not detected by texture-analyzer measurements. Fillet-level TACTUS force showed moderate but significant correlation with texture-analyzer measurements (Pearson r = 0.441, p < 0.001). Semivariogram, local change entropy, and heterogeneity analyses revealed increasing and anatomically structured hardness variability with WB severity. These findings demonstrate TACTUS as an automated, operator-independent platform for spatially resolved mechanical characterization and ground-truth generation to support hyperspectral and near-infrared spectroscopy-based WB classification systems.
Read the full abstract
Machine Learning
Agnostic Marginal-Nonuniform Learning
We study the problem of binary classification in the marginal-nonuniform learning framework in the agnostic setting. In this model, the constants in the learning rate may depend on the marginal distribution of the unlabeled data, but the guarantee must hold uniformly over all conditional distributions of the labels. We also study fine-grained rates. We consider two criteria. (1) The standard agnostic setting, where the learner competes with the best concept in the class. In this setting, an immediate consequence of standard agnostic PAC learning results gives us the optimal rate of $\tfrac{1}{\sqrt{n}}$ for every class $\mathcal{H}$ with finite $\mathrm{VC}(\mathcal{H})$ and shows that arbitrarily slow rates are necessary otherwise. This also gives us the optimal fine-grained rate of $\sqrt{\tfrac{\mathrm{VC}(\mathcal{H})}{n}}$. Therefore, the dependence on the marginal distribution does not help, which is in contrast to the realizable setting. (2) In the relaxed agnostic setting, where the learner competes with $(1+\lambda)$ times the error of the best concept for a fixed $\lambda \in (0,1]$, the picture from the realizable setting reappears. We obtain a trichotomy of the achievable optimal rates: exponential rates of the form $e^{-n}$ arise precisely when the class is finite; linear rates of the form $\tfrac{1}{n}$ arise when the class is infinite but $\mathrm{VC}(\mathcal{H})$ is finite; and arbitrarily slow rates occur when $\mathrm{VC}(\mathcal{H})=\infty$. We also obtain the optimal fine-grained rate of $\tfrac{\mathrm{uVCE}(\mathcal{H})}{\lambda n}$, where $\mathrm{uVCE}(\mathcal{H})$ is a recently introduced combinatorial complexity dimension called the uncentered $\mathrm{VC}$-eluder dimension.
Read the full abstract
High Energy Physics (Theory)
Deep Inelastic Scattering and Proton Structure Functions in AdS/QCD models
In this work we study the holographical fermionic deep inelastic scattering in a deformed AdS model. Imposing that the potentials that emerge from the Schrödinger-like equations for the fermionic chiral left and right modes are SUSY partners to preserve isospectrality, we implement a systematical truncation of such potentials. Then, we show that in the quadratic approximation of the deformed AdS model, the fermionic DIS and the proton structure functions $F_{1,2}(q^2,x)$ are identical with the results obtained from the soft wall model. Further, we obtain and compare high-order approximations for the fermionic potentials with the full solution of the deformed AdS model and show, as expected, that as greater the order of approximation, the better are the results for $F_{1,2}(q^2,x)$ when compared with experimental data explicitly for the Bjorken parameter values $x=0.85, 0.75$ and $0.65$.
Read the full abstract
Statistical Mechanics
Renewal Bias in Parallel Simulations of Absorbing Processes
Asynchronous parallel simulations of stochastic processes with absorbing states can exhibit systematic sampling bias when trajectory durations depend on their eventual absorbing outcomes. Workers that continuously restart upon absorption preferentially record rapidly terminating trajectories within a fixed computational budget, distorting the observed ensemble statistics. We investigate this effect using a minimal renewal model with two absorbing states and state-dependent exponential durations. Renewal theory yields exact expressions for the expected number and composition of completed trajectories, showing that the resulting bias decreases algebraically with execution time but persists as the number of parallel workers increases at fixed budget. We compare continuous sampling with First-Trajectory Filtering, which retains only the initial trajectory attempted by each worker. Although finite execution deadlines introduce residual censoring bias, this bias decreases exponentially with the computational budget. An analytical mean-squared error calculation reveals budget-dependent crossover worker allocations beyond which First-Trajectory Filtering becomes more accurate despite collecting fewer trajectories. We also examine the limitations of statistical inference under severe censoring when the absorbing-state probabilities and characteristic durations are unknown. Finally, simulations of the Axelrod model near its nonequilibrium transition demonstrate the preferential sampling of rapidly absorbing fragmented configurations, illustrating the relevance of the renewal mechanism in a spatially interacting system. Our results show that increasing computational throughput does not necessarily improve statistical accuracy and highlight the importance of accounting for trajectory-duration heterogeneity in parallel simulations of absorbing processes.
Read the full abstract
Computation and Language
A Lattice-Based HMM Approach to Aruz Scansion of Ottoman Poetry
Ottoman Divan poetry is composed in aruz, a quantitative meter based on the weights of syllables. The meter is sensitive to specific characters recorded in the original writing in the Arabic script. This work is concerned with automatic meter finding and scansion alignment for Ottoman Divan poetry, especially on modern Turkish spelling without the special diacritics that scholarly transliterations of Ottoman preserve. We define two tasks on this input, first identifying the meter used in the poem, and second aligning the meter's weights with the syllabification of the poem, identifying the flaws and freedoms the poet has taken along the way. We build a corpus of 52,000 lines split by poet, with scansion annotations for 500 test lines, and a robustness benchmark generated at the deviation rates measured on real web and OCR copies. Our model is a hidden Markov model over the positions of each meter, decoded over a lattice of every candidate path, and trained by EM from the printed meter, without scansion labels. On a held-out set it scans 77.6% of the test lines fully correctly, against 26.8% for the available rule-based scanner, 55.0% for a neural tagger trained with the same supervision and 63.6% for Claude Opus 5.5 few-shot. The code is open-sourced and the tool is released on the web.
Read the full abstract
Statistical Methodology
Average versus expected causal effects under interference: Modeling assumptions and estimation bottlenecks
Causal inference under treatment interference requires assumptions to reduce the exponentially large number of potential outcomes. We develop hierarchical finite-population models for potential outcomes by separating these assumptions into three layers: interference neighborhoods, exposure mappings, and outcome models. Given an interference neighborhood and exposure mapping, we show that potential outcomes admit a nonparametric decomposition into direct, spillover, and treatment–spillover interaction components. This decomposition provides a common coordinate system for nonparametric, semiparametric, and parametric models of finite-population potential outcomes. For a given potential-outcome model, we define two classes of estimands. Average causal effects (ACEs), such as direct and total treatment effects, contrast potential outcomes under fixed treatment–exposure conditions. Expected causal effects (ECEs), such as expected direct and indirect effects, contrast expectations of potential outcomes under a specified allocation policy. For both classes, we bound the variance of the canonical Horvitz–Thompson estimator by three terms: a diagonal weight term, a term from overlapping interference neighborhoods, and a residual design-dependence term. Our upper and lower bounds show that under graph-unaware designs, nonparametric design-unbiased estimation of ACEs can require sample sizes growing exponentially in the interference degree, whereas polynomial growth can suffice for design-aligned ECEs. Under neighborhood interference, we construct graph-aware independent-set designs with variance bounds polynomial in the maximum degree for direct and total treatment effects on every graph, whereas graph cluster randomization can incur worst-case variance exponential in the degree.
Read the full abstract
Machine Learning
CoilStellaration: A Dataset and Benchmark for Engineering-Aware Stellarator Coilset Generation
Stellarators are fusion devices that use externally generated, three-dimensional magnetic fields to confine hot plasma and offer a promising route to commercial fusion. Their electromagnetic coils pose a challenging design problem: they must reproduce the target magnetic field while satisfying engineering constraints and remaining practical and cost-effective to manufacture. Finding suitable geometries is a highly nonlinear inverse problem that generally has neither a closed-form nor a unique solution. This problem lies at the heart of stellarator design. We introduce CoilStellaration, a dataset and benchmark for learning the mapping between coil geometries from target plasma magnetic fields and engineering requirements. The dataset contains 177,000 optimized filamentary coilsets targeting plasma equilibria from ConStellaration, with sampled engineering requirements and achieved geometric and magnetic-field metrics. It spans multiple coil counts and field-period numbers, supporting studies of both specialized and shared models across configurations. The benchmark asks models to predict coils conditioned on a target equilibrium and requirements on coil-coil and coil-plasma spacing, curvature, and magnetic-field error. We show that magnetic-field accuracy remains the main challenge, with learned models rarely satisfying all requirements simultaneously. However, using learned predictions to initialize conventional coil optimization can accelerate early progress. We release baseline models and evaluation code to support reproducible research on both direct coil generation and learned optimization initializations. By lowering the barrier for machine learning and optimization researchers to engage with stellarator coil design, we aim to accelerate progress toward practical, cost-effective stellarator power plants.
Read the full abstract
Classical Analysis and ODEs
Second-Order Linear $q$-Difference Equations
Second order $q$ difference equations arise in various areas such as basic hypergeometric series, orthogonal polynomials and the analytic theory of special functions. When the independent variable is scaled by powers of $q$ ($0<q<1$), such equations display asymptotic behaviour that differs markedly from their differential or classical difference counterparts. Our purpose is to give a complete, self contained treatment of linear, second order $q$ difference equations with divergent formal solutions, including explicit error bounds, exact resummations and Stokes phenomena.
Read the full abstract
Human-Computer Interaction
reComposer: Editable Planning for Controllable Image Generation
Prompt-based image generation makes it easy to produce visual assets, but offers little support for engaging with the creative process. Prompting collapses interdependent creative decisions into a monolithic instruction and returns a finished image, leaving control limited to revising the prompt or correcting the model's output. We present reComposer, a process-oriented image-generation interface built around editable planning, which decomposes users' evolving intent into meaningful creative decisions - such as subject, composition, or palette - and collaborates with users to explore underspecified goals and alternatives. reComposer supports conversational and direct manipulation of the plan, moving between concept-level intent and visual-level control. We evaluated reComposer with 8 design and non-design professionals. Compared to prompting, participants felt greater control and ownership while developing visual assets, exploring alternatives, and iterating on details. Our findings suggest how shifting generative AI interaction from specifying outputs to shaping process can support creative thinking alongside asset production.
Read the full abstract
Computer Vision
Differentiable Inverse Rendering Recovers Intrinsic Properties from Intravascular Optical Coherence Tomography
Medical images encode intrinsic tissue properties that indicate disease progression, yet these properties remain entangled with the system-dependent physics of image formation. Differentiable inverse rendering asks whether they are recoverable by inverting the imaging system's forward physics rather than through post-hoc correction. We address this in optical coherence tomography (OCT), focusing on the optical attenuation coefficient, a per-pixel marker of pathology and cellular composition. Conventional recovery relies on closed-form inversions limited to simplified light-tissue models, or learned methods coupled to specific imaging configurations. We instead recover per-pixel attenuation from a single OCT image by gradient descent through a differentiable physics-based forward model. We verified the method against analytical solutions, validated it on five in vitro phantoms, and applied it without modification to 29{,}269 clinical intravascular OCT images from 128 patients. The inversion needs no training data or segmentations and is portable across platforms and retrospective archives. Replacing the forward model or recovering another optical parameter leaves the optimization procedure unchanged. We demonstrate per-pixel recovery is tractable because gradient information scales to this high-dimensional problem, where gradient-free methods become intractable. Our work supports quantitative tissue characterization from OCT images and provides a framework for extension across imaging modalities and organ systems.
Read the full abstract
Classical Analysis and ODEs
Sharpness of the Martin-O'Bryant autoconvolution inequality
We show that there does not exist a constant $c<1$ such that ${\lVert f\ast f\rVert}_{L^{2}}^2\le c\,{\lVert f\ast f\rVert}_{L^{1}}{\lVert f\ast f\rVert}_{L^{\infty}}$ holds for every nonnegative $f\in L^1(\mathbb{R})\cap L^2(\mathbb{R})$.
Systems and Control
Hierarchical Nonlinear Control for Swarm Encirclement Around a Moving Target UAV
This paper presents a hierarchical inner-outer loop nonlinear control strategy using a geometric Lyapunov-based framework for the cooperative encirclement of a target by multi-ple unmanned aerial vehicles (UAVs) in three-dimensional space. The encirclement problem is formulated on a two-sphere with arbitrary radius and center, denoted Gamma-2, by projecting agent positions from R3 onto a spherical manifold centered at the target, being the unit two-sphere at the origin S2 used for intermediate calculations. Desired encirclement configurations are generated via a point separation method on Gamma-2, while stable convergence from R3 to these points is guaranteed using a Lyapunov-based reduced-attitude control law inspired by geometric control on the two-sphere. Formal stability results are derived using Lyapunov analysis and the framework is validated in simulation using a nonlinear quadrotor model and controller. Results demonstrate stable encirclement and highlight the suitability of the proposed approach for three-dimensional multi-agent aerial systems.