Peer Revenue Play credits. Real papers.

Papers

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

All of arXiv

Thursday 10 September

Differential Geometry

Ricci-Yamabe solitons on the Lie group $\Sol \times \mathbb{R}^n$

In this article, we study Ricci-Yamabe solitons on the Lie group $\mathrm{Sol} \times \mathbb{R}^n$ equipped with a natural left-invariant Riemannian metric, explicitly determining the vector fields that characterize them. We then deduce that, in the case of a Ricci soliton, it is expanding, whereas in the case of a Yamabe soliton, it is shrinking. Finally, we show that if this Lie group is a gradient Ricci-Yamabe soliton, the vector field belongs to $\operatorname{Span}\{\partial_{t_1}, \dots, \partial_{t_n}\}$, and we explicitly provide the Perelman potential.
Read the full abstract

Software Engineering

Open Source Stewardship Communities: "We need you, but not your pull request"

Human-centric AI for software engineering means keeping humans responsible for work performed with AI. In Open Source Software (OSS), AI lowers the cost of implementing changes, but reviewing someone else's contribution remains comparatively expensive, so some projects now restrict who may contribute implementations while still welcoming other participation—not because the code is AI-generated, but because it no longer justifies the review cost. We call the resulting form a stewardship community: a small core retains implementation authority while a broader community continues to shape the software without writing code, and access to coding increasingly depends on approval rather than self-initiated contribution. This raises a broader question: what happens to the human community when coding agents let maintainers replace implementation work once supplied by external contributors? For human-centric software engineering, keeping humans in control of AI agents is not enough: AI can replace implementation labor while weakening how OSS communities renew themselves.
Read the full abstract

Spectral Theory

Payne's conjecture for buckling eigenvalues of odd index

In 1955, L. E. Payne conjectured that for a planar clamped plate, each buckling eigenvalue is at least as large as the Dirichlet Laplacian eigenvalue of the same domain with index shifted by one. We prove this conjecture for every odd buckling index.

Combinatorics

Extremal List Gaps and Inapproximability in Additive Graph Labeling

We study a vertex-labeling analogue of the $1$-$2$-$3$ problem and its list version. For a labeling $\ell:V(G)\to\mathbb N$, let $S_\ell(v)=\sum_{w\in N(v)}\ell(w)$. The additive number $\eta(G)$ is the least $k$ for which there exists $\ell:V(G)\to[k]$ such that $S_\ell(u)\ne S_\ell(v)$ for every $uv\in E(G)$, while the list additive number $\eta_\ell(G)$ is the least $k$ such that the same condition can be satisfied from every assignment of $k$-element lists $L(v)\subset\mathbb N$ with $\ell(v)\in L(v)$. We show that for every $k\ge2$, there is a graph $G$ with $\eta(G)=1$ and $\eta_\ell(G)\ge k$. The separation persists at the minimum possible ordinary value for positive-degree regular graphs: there is a regular graph $H$ with $\eta(H)=2$ and $\eta_\ell(H)\ge k$. We also determine a sharp lower bound for $\eta(G)$ in terms of the order and minimum degree of $G$, and show that the unbounded list gap persists at asymptotically extremal density. Finally, for every fixed $k\ge2$, it is NP-hard to distinguish $\eta(G)=2$ from $\eta(G)>k$, even on asymptotically extremal dense graphs. Consequently, $\eta(G)$ admits no polynomial-time constant-factor approximation unless $\mathrm P=\mathrm{NP}$. Together, these results reveal a robust gap phenomenon: the separation between ordinary and list additive labeling persists at the smallest possible ordinary values and even under asymptotically extremal density, while the ordinary parameter itself remains hard to approximate.
Read the full abstract

Accelerator Physics

Continuous-Wave Operation at 30 MV/m of a Superconducting Radio-Frequency Gun Cryomodule with an Exchangeable Cathode Plug

A prototype 185.7 MHz superconducting radio-frequency (SRF) electron gun cryomodule has been developed for the proposed LCLS-II-HE Low Emittance Injector at SLAC. It has an in-situ ultra-high-vacuum exchangeable cathode plug system compatible with semiconductor photocathodes. With a copper plug, the gun achieved continuous-wave RF operation at a cathode field of 30 MV/m. No measurable degradation was observed after plug removal and reinsertion with the cavity at 4.4 K. This result is a significant improvement over past continuous-wave SRF gun systems with exchangeable plugs.
Read the full abstract

All authors T. Konomi, T. Xu, Y. Choi, K. Elliott, B. Gower, B. Tousignant, J. Wenstrom, A. Taylor, C. Compton, W. Chang, X. Du, W. Hartung, S. H. Kim, S. Miller, D. Morris, M. Patil, L. Popielarski, J. Wei, Z. Yin, J. Smedley, J. Maniscalco, C. Adolphsen, M. Murphy, D. A. Gonnella, M. P. Kelly, T. Petersen, A. Arnold, S. Gatzmaga, R. Steinbrück, P. Murcek, R. Xiang, J. Teichert, J. Lewellen

Analysis of PDEs

Multi-Spectral QPAT with Frequency Averaged Measurements

We study some inverse problems in a frequency-averaged model for quantitative photoacoustic tomography, in which the optical parameters are modeled by a finite number of spectral functions and spatial coefficients, and the measurements are given by internal observables averaged over a spectral bandwidth. Using multiple measurements generated by suitable complex geometrical optics solutions, we establish uniqueness and Lipschitz-type stability for the associated linearized inverse problem for diffusion and absorption recovery. This linearized stability further yields local Hölder-type stability for the nonlinear inverse problem via an abstract linear-to-nonlinear argument.
Read the full abstract

Software Engineering

Learning to adapt GR(1) specifications through degradation

Reactive synthesis is a powerful tool for generating correct-by-construction controllers from formal specifications. GR(1) is an assume-guarantee specification framework that enables efficient synthesis, allowing synthesised controllers to be used in a wide array of applications. The limitation of such controllers is that, should they encounter environment behaviour unspecified in the assumptions of the specification, the specified system guarantees are no longer ensured. Our work proposes an approach based on oracle-guided inductive synthesis to adapt the specification to be consistent with the observed assumption violation, while degrading system guarantees as little as possible to maintain realisability. Our methodology discovers multiple potential solutions, so we propose a preference criteria, based on the ability of the specification to enable robustness under adaptation. Although our approach is capable of degrading the entire specification, for our case studies we successfully discover degradations that preserve the entire set of original guarantees.
Read the full abstract

Computation and Language

Repair Before Reinforce: Context-Augmented Knowledge Graph Reasoning for Multi-Hop Question Answering

Question-answering often requires reasoning across multiple connected facts rather than retrieving a single isolated relation. Knowledge graphs (KGs) provide a structured way to represent such facts, but training large language models (LLMs) only on isolated KG head-relation-tail triples may limit their ability to learn the surrounding context needed for multi-hop reasoning. In this work, we propose a context-augmented training framework for multi-hop question-answering. Although generally applicable, we validate the framework in the context of disease-specific KGs, extracted using a reliable KG extraction framework called GraphMERT, for Gastroparesis and Diabetes. For each primary KG triple, we attach supporting triples extracted from the same source text chunk to form a context graph (CG). This creates two supervision settings: KG-grounded supervision, which uses only the target KG triple or path, and CG-grounded supervision, which uses the target KG triple or path together with supporting context triples. We train the Qwen3-14B model using supervised fine-tuning (SFT) under both settings, producing KGModel and CGModel variants. To strengthen the lower-hop factual foundation of the models, we introduce an LLM-judged, history-aware adaptive repair pipeline that identifies unresolved one-hop failures, continually fine-tunes on targeted repair examples, and removes or quarantines problematic noisy triples. This repair stage enables the models to reach 100% accuracy on the cleaned retained one-hop validation sets. Finally, we employ reinforcement learning (RL) using lower-hop question-answer items and evaluate generalization on harder 3-hop, 4-hop, and 5-hop tasks. Across both diseases, context-augmented supervision consistently improves multi-hop performance over KG-only supervision. RL initialized from repaired SFT checkpoints yields larger and more stable gains.
Read the full abstract

Fluid Dynamics

Spreading dynamics of a drop of yield-stress fluid subject to vertical oscillations

The measurement of rheological properties from non-idealised flows is routinely used to characterise yield-stress materials. We extend the canonical slump test, which evaluates yield stress based on the spread of a large drop under gravity, by subjecting such a drop to vertical oscillation of its substrate. We test this flow configuration by direct comparison between experiments and numerical simulations of the time-dependent Cauchy equations coupled to the Saramito-Herschel-Bulkley (SHB) constitutive model. This provides a sensitive test of the SHB model near yield because the vibrated drop can releases stress exceeding the yield stress by adjusting its shape as the forcing acceleration is increased, and thus, remains close to the yield threshold for all forcing. We find that the SHB model systematically underpredicts the amplitude of the viscoelastic oscillatory response of the drop to forcing by a factor of five. However, spreading can be captured numerically by selecting a solvent viscosity that represents the appropriate post-yield rheology of the material at the expense of matching sub-yield properties. This highlights a fundamental challenge in simulating flows in which yielded and unyielded material coexist, such as those encountered in proto-rheological tests. We establish the vibrated drop as a powerful proto-rheological configuration for estimating the yield stress across drops of different rheology, size and shape, by establishing a proportional relationship between the threshold acceleration required to induce spreading relative to gravity and the yield number, the ratio of yield stress to gravitational stress acting at the centre of mass of the drop.
Read the full abstract

Atmospheric and Oceanic Physics

WIND-Bench: A Benchmark Dataset for In-Situ Near-Surface Wind Speed Observations Across the Conterminous United States

Accurate wind forecasts are essential for operational decision-making and public safety, yet forecasts tend to miss near-surface high wind speeds in complex terrain. In response, advances in machine learning (ML) weather prediction methods have demonstrated the ability to improve forecast skill beyond traditional numerical weather prediction (NWP) models. However, the absence of a benchmark dataset to evaluate NWP and ML models with sufficient, quality-controlled wind speed observations in complex terrain poses challenges to the development and intercomparison of high-quality surface wind forecasts across the Conterminous United States (CONUS). We develop the Wind IN-situ Data Benchmark (WIND-Bench), a benchmark dataset from in-situ observations in the Meteorological Assimilation Data Ingest System (MADIS) observational network. WIND-Bench integrates multiple sensor networks with quality control that distinguishes sensor failures from high-wind conditions, using a framework that validates observations against forecasts from the National Oceanic and Atmospheric Administration (NOAA) High-Resolution Rapid Refresh (HRRR) model. WIND-Bench provides a standardized benchmark for evaluating ML and NWP models and for quantifying forecast skill, accelerating the development, evaluation, and operational deployment of skilled near-surface wind forecasts.
Read the full abstract

All authors Kyla Bazlen, Grant Buster, Brandon Benton, Lauren North, Ansley Baring, David D. Turner, Emily Wells, Laura Vimmerstedt

Portfolio Management

Seasonal Trading in Commodity Futures: Evidence from Regression and Singular Spectrum Signals

Commodity futures are shaped by harvest cycles, weather shocks, storage conditions, and seasonal demand, but it remains unclear whether recurring patterns yield robust out-of-sample trading profits. Existing research documents return seasonality in commodity futures as well as more complex seasonal structure, while leaving less evidence on how alternative seasonal models compare under common implementation constraints. This article compares dummy-variable regression (DVR), Singular Spectrum Analysis (SSA), and robust low-rank SSA (RLSSA) within a unified trading framework, including a volatility-normalised specification. Using monthly delivery-avoidance returns for 15 liquid commodity futures, the models are estimated on rolling ten-year windows and evaluated from 2016 to 2024 with transaction costs, an equal-weight long benchmark, and Maximum Entropy Bootstrap (MEB) assessment. Across 500 MEB paths, the benchmark has the strongest average full-period profile, with a cumulative return of 16.81%, a Sharpe ratio of 0.191, and a maximum drawdown of -0.414. Classical SSA has the strongest average model outcomes, but negative median cumulative returns and deep drawdowns indicate substantial path sensitivity. Volatility normalisation reduces the average DVR short loss but generally weakens SSA-based portfolios. None of the 18 approximate paired MEB Sharpe tests rejects after within-family Holm adjustment. Because MEB preserves each contract's temporal rank ordering, the assessment is conditional on observed timing rather than a timing-randomised seasonal null. The evidence does not establish robust benchmark outperformance and shows that model performance varies materially across market subperiods.
Read the full abstract

Classical Analysis and ODEs

Trace-defect bounds for discrete Fourier concentration operators

Let $\Omega\subset\mathbb Z^d$ be finite and let $S\subset\mathbb T^d$ be measurable. We study the discrete Fourier concentration operator \[ T_{\Omega,S}=P_\Omega B_SP_\Omega, \] which describes simultaneous localization to a finite set of spatial indices and a prescribed spectral region. For independent exponents $\gamma,\eta\in(0,1]$, we introduce translation seminorms for $\Omega$ and $S$ and use them to bound the trace defect \[ \operatorname{tr}(T_{\Omega,S}-T_{\Omega,S}^2). \] The proof combines an exact trace-defect identity with dyadic Fourier estimates, leading to three regimes: when $\gamma=\eta$, the contributions from the dyadic scales are of the same order, and their sum produces a logarithmic factor, while for $\gamma\ne\eta$ the sum is controlled by one end of the scale range. For discretizations $$ \Omega_R=(RF)\cap\mathbb Z^d $$ of a bounded measurable set $F\subset\mathbb R^d$ satisfying an upper Minkowski-neighborhood estimate with exponent $\gamma$, we obtain trace-defect bounds of order $R^{d-\gamma}\log R$ in the critical case and $R^{d-\min\{\gamma,\eta\}}$ off the critical line. These estimates yield quantitative eigenvalue-counting and plunge-region bounds and recover the discrete Landau asymptotic. For a box model at $\gamma=\eta=1$, the logarithmic factor is sharp.
Read the full abstract

Machine Learning

PLSP (Pre-hoc Liminal Space Profiling): OOD Prediction over Detection – An Anticipatory Approach for Machine Learning Model Reliability

Out-of-Distribution (OOD) data poses a significant threat to machine learning models, often leading to model failure during deployment. All existing OOD detection methods are post-hoc, relying on evaluation metrics such as accuracy and AUC-ROC during inference to indirectly assess the model's response to OOD data by measuring deviations. In contrast to existing approaches, the proposed work shifts the paradigm from OOD detection to OOD prediction by proposing a pre-hoc anticipatory framework called PLSP for OOD prediction. We make several key contributions: (a) a dataset-independent metric called the CREDibility Score (CREDS) is proposed for OOD prediction; (b) credibility curves are introduced to study the maximum credibility a model can attain; and (c) credibility heat maps (and volume under surface) are introduced to characterize pre-hoc model behavior across different datasets. This work provides a novel perspective on signal processing under distributional shifts. Experiments across multiple datasets demonstrate that the proposed metric serves as a valuable measure for improving the robustness of machine learning models toward OOD prediction.
Read the full abstract

All authors Vipul Bansal, Himanshu Buckchash, Balasubramanian Raman, Deepak Dhungana

Machine Learning

Patient-Reported Survey Data Improve Prediction of Opioid Use Disorder

Electronic health records (EHRs) may incompletely capture patient-reported factors associated with opioid use disorder (OUD). We evaluated whether survey data improve prediction of a first recorded OUD diagnosis among 267,747 All of Us participants with documented opioid exposure, including 15,287 OUD cases. We compared EHR-only and EHR+survey models across 6-, 12-, and 24-month look-back windows using logistic regression, random forest, XGBoost, LightGBM, multilayer perceptron, LSTM, GRU, and Transformer. Survey augmentation improved PR-AUC across all 24 model-window combinations by 0.0087-0.0505; the best 24-month LightGBM model improved from 0.6219 to 0.6603. Survey coverage increased with longer windows and differed by OUD status (24 months: 21.7% OUD-positive vs. 60.7% OUD-negative). Permutation analysis ranked survey features as the second most important information domain at 24 months in both evaluated models. Patient-reported data provide complementary predictive signals beyond structured EHRs while highlighting the importance of survey availability.
Read the full abstract

All authors Xiyue Jiang, Zihan Ding, Grace Han, Yinan Liu, Richard N. Rosenthal, Fusheng Wang

Machine Learning

Predicting Collision Cross Sections with GRACE: Geometric Residual Adduct Conditioning via Early-fusion

Collision cross section (CCS), derived from ion mobility mass spectrometry, is a common descriptor for molecular annotation. Prediction is challenging for machine learning models because it reflects the size, shape, and ionization state of a gas-phase molecular ion. Most predictors either ignore explicit 3D structure or treat adduct identity as a late categorical feature, which limits their ability to capture adduct-dependent geometric effects. We present GRACE (Geometric Residual Adduct Conditioning via Early-fusion), a 3D CCS predictor that adapts a pretrained molecular geometry encoder using geometric residual adduct conditioning via early fusion. GRACE combines two inductive biases: a residual objective relative to an adduct-aware physical descriptor baseline and adduct conditioning within the encoder via a learned adduct token and low-rank attention adapters. We evaluate the model on a curated set of over 9,000 experimental molecule-adduct CCS records with random, scaffold, and adduct-sensitive splits designed to separate interpolation, scaffold generalization, and adduct-driven generalization. GRACE achieves the best mean percentage difference among the evaluated learned models on all three splits: 1.67% on the random split, 2.11% on the scaffold split, and 2.36% on the adduct-sensitive split. Diagnostic analyses suggest that residual learning stabilizes training by removing the dominant mass-CCS trend, while early fusion improves adduct-sensitive prediction relative to late fusion. Across four independent external test sets, GRACE shows consistently lower error than the other evaluated models. On a held-out set, GRACE also attains the lowest mean percent difference when compared with four previously reported physics-based workflows. These results support residual learning and encoder-level adduct conditioning as practical inductive biases for fast, accurate CCS prediction.
Read the full abstract

All authors Parthasarathy Suryanarayanan, Susanta Das, Shreyans Sethi, Kenneth M. Merz, Jr., Joseph A. Morrone

General Relativity and Quantum Cosmology

Classical and quantum perspectives on temperature in accelerated frames

We study the notion of temperature in uniformly accelerated frames within a relativistic thermodynamic framework. Considering a perfect fluid at rest in a non-inertial (Rindler) frame, we derive the condition for thermal equilibrium from energy–momentum conservation. This leads to a position-dependent temperature profile consistent with the Tolman–Ehrenfest relation. We emphasize that this temperature characterizes the local thermodynamic equilibrium of the fluid in the accelerated frame and does not arise from a transformation law between different observers. The physical interpretation of the result and its relation to acceleration-induced effects are briefly discussed.
Read the full abstract

Robotics

Chain-SLAM: Globally Consistent Backend for Multi-Session LiDAR SLAM via Chained Loop Closure

Maintaining consistency over long spatial and temporal horizons remains a fundamental challenge in large-scale LiDAR SLAM, particularly when integrating maps collected across multiple sessions. We present Chain-SLAM, a LiDAR SLAM backend enabling online multi-session map alignment and reuse with global consistency at large scale. We implement a chained loop closure mechanism that efficiently propagates geometric constraints across inter-session keyframes through an adjacency graph, enabling robust long-horizon consistency triggered by reliable short-horizon loop closures. The system initializes inter-session alignment with GNSS-proximity place recognition, then performs on-the-fly loop closure detections and joint optimization of loaded maps and newly acquired trajectories within a unified factor graph, maintaining both inter- and intra-session geometric consistency without dynamic object removal, and cross-platform robustness with minimal hyperparameter tuning. Experimental results show improved trajectory accuracy and robust multi-session integration on large-scale datasets. We release our source code to support reproducible research in large-scale multi-session LiDAR SLAM. Project site: https://ai4ce.github.io/Chain-SLAM/
Read the full abstract

All authors Zhiheng Li, Xinhao Liu, Juexiao Zhang, Yongqing Liang, Chen Feng

Game Theory

On the Fragility of Worst-Case Nash Equilibria in Atomic Congestion Games

Aggregate performance in smart mobility systems depends heavily on the emergent behavior of selfish, resource-sharing agents that participate within the systems. As a result, recent work has focused on how a system designer can leverage incentives to influence behavior so that system cost (e.g., traffic congestion) is minimized. These results show that worst-case equilibria can be quite inefficient compared to system-optimal allocations. However, it is unclear to what extent agents are “satisfied” with their decisions in these worst-case scenarios. We demonstrate that in any incentivized atomic congestion game, agents' aggregate satisfaction at equilibrium (relative to their actions in an optimal allocation) is correlated with the efficiency of the corresponding system cost, in the sense that if agents are very satisfied with their equilibrium choices, the equilibrium must be relatively efficient. Further, we show that worst-case Nash equilibria are fragile, as every agent is indifferent between their action in a worst-case equilibrium and their action in a system-optimal allocation. In summary, at equilibrium, either agents are highly satisfied with their decisions or their decisions are highly inefficient, but both cannot be true simultaneously. This work adds to recent results for other classes of games which indicate that worst-case equilibrium efficiency guarantees only occur when agents are indifferent about their decisions.
Read the full abstract

Quantum Physics

Learning Sparse Quantum States

We study the problem of tomography for $k$-sparse quantum states. In contrast to classical distribution learning, where tight sample and time complexity bounds in terms of support size are well understood, no non-trivial bounds were previously shown for this problem. We give the first near optimal algorithm for learning $n$-qubit $k$-sparse pure quantum states, obtaining fidelity at least $1-\varepsilon$ with high probability using $\tilde{O}(k/\varepsilon)$ copies of the state and $\tilde{O}(kn/\varepsilon)$ time. Both bounds are optimal up to polylogarithmic factors. As an implication, we also obtain an algorithm with near optimal $\tilde{O}(kr/\varepsilon)$ sample complexity for learning $k$-sparse rank-$r$ mixed states, via the random purification channel technique. Obtaining time complexity nearly matching the sample complexity, for $r>1$, remains an important open question.
Read the full abstract

Human-Computer Interaction

BRIDGE-EEG: Bridging Self-Supervised Pretraining and Efficient Deployment for Cross-Dataset EEG Classification

The growing use of electroencephalography (EEG) motivates automated analysis that is accurate, transferable, and deployable on constrained hardware. Recent EEG foundation models learn general representations from large-scale pretraining, but their size and computational cost limit edge and wearable deployment. We introduce BRIDGE-EEG, an efficient multi-task EEG classification pipeline that preserves the benefits of pretraining while reducing model size. A unified preprocessing scheme maps heterogeneous recordings with different channel counts, montages, and sampling rates to a device-agnostic 62-channel time–frequency representation. We pretrain an SE-ResNet18 teacher (11.84 M parameters) with SimCLR on unlabeled EEG from five heterogeneous datasets, then compress it into SE-ResNet8 (1.56 M) and SE-ResNet4 (0.48 M) students using task-agnostic and task-specific distillation. We evaluate six benchmarks spanning abnormality detection, motor imagery, and emotion recognition. For abnormality detection and emotion recognition, the students achieve accuracy comparable to or better than several recent EEG foundation models with 10–1,000$\times$ more parameters. Motor imagery shows a remaining representation gap, highlighting the importance of pretraining diversity. Inference profiling on a server GPU, desktop CPU, and NVIDIA Jetson Orin Nano shows up to 3.0$\times$ lower edge energy per inference (15.64 mJ vs. 46.67 mJ). The compact models further support future deployment on MCU-class wearables.
Read the full abstract

All authors Meghna Roy Chowdhury, Chengwei Zhou, Haotian Yu, Gourav Datta, Shreyas Sen

Geometric Topology

Holomorphic realizations of pairs of foliations on Riemann surfaces

Let $X$ be a hyperbolic Riemann surface and let $\mu$ and $\nu$ be laminations on $X$ homotopic to measured foliations with finite Dirichlet integral. We prove that $\mu$ and $\nu$ are filling if and only if there exists a homeomorphism to another Riemann surface $f:X\to Y$ and an integrable holomorphic quadratic differential $q$ on $Y$, unique up to the natural equivalence, such that the push-forward laminations are homotopic to the horizontal and vertical foliations of $q$ respectively. This extends a classical theorem of Gardiner-Masur from closed surfaces to arbitrary surfaces. As well, the dual $\mathbb{R}$-tree interpretation yields the solution of an asymptotic Plateau problem for minimal surfaces in a product of two $\mathbb{R}$-trees. We construct examples such that $f:X\to Y$ is not homotopic to a quasiconformal map, and we present sufficient conditions that ensure it is. We deduce applications to main inequalities for locally quasiconformal maps, harmonic maps between surfaces, and big mapping class groups.
Read the full abstract

Robotics

Guardrailed Meta-Agent Loops: Stress-Testing Policy Pinning, Budget Bounds, and Crash Recovery

Self-improving agent workflows create an audit problem when the same controller can change both its behavior and the conditions under which that behavior is judged. We present GuardrailLoop, a simulation-based testbed that makes three operational contracts jointly testable: preservation of human-defined policy, compute accounting at every recorded execution prefix, and recovery of a specified scientific state after crashes. A hash-pinned policy fixes goals, scope, evaluation identity, budget, and release conditions; machine-directed evolution is restricted to a code-owned feature catalog and bounded knobs. The contribution is an executable boundary and an evaluation protocol that separates useful adaptation, state recovery, and repeated execution. In a paired 50-seed 2 x 2 study, round-stage growth changes target attainment by +1.00 and restricted mean compute to target by -56.97 simulated GPU-hours (95% paired-bootstrap interval [-58.91,-54.70]); idle growth has zero measured utility effect. Across 240 enumerated crash injections, all runs recover the defined outcome, but only 210 preserve the normalized trace: 30 pre-commit crashes repeat a planner call. Resource-drift, kill-switch, integrity, and output-guard matrices satisfy their specified checks. These findings show why successful outcome recovery is insufficient evidence of exactly-once execution. They establish conformance within one calibrated deterministic testbed, rather than general safety or real-world self-improvement.
Read the full abstract

Solar and Stellar Astrophysics

Plasma Heating and Energization in Hot-Onset Flare Precursor Events

Recent observations of solar flares have revealed a distinct 'hot onset' phase preceding the impulsive phase, characterized by elevated plasma temperature (>10 MK), a steady increase in emission measure, and the absence of detectable nonthermal emission. The standard flare model attributes the formation of hot flare plasma to chromospheric evaporation driven by energy deposition from nonthermal electrons, and therefore does not naturally explain why thermal heating precedes nonthermal emission. Motivated by recent observations of guide-field evolution in two-ribbon flares and by large-scale MHD simulations, we investigate hot-onset energization using simulations of magnetic reconnection with the kinetic macroscale model kglobal initialized with a spatially non-uniform guide field. Because the efficiency of electron non-thermal energy production from Fermi acceleration decreases with guide-field strength, the evolving guide field produces two distinct energization regimes. During the initial high-guide-field phase, the bulk electron temperature rises to approximately twice its initial value, while non-thermal particle production remains weak. Upon entering the low-guide-field regime, energetic particle production is significantly enhanced, with the population of high-energy electrons growing rapidly. These findings provide a self-consistent physical explanation for the hot onset, demonstrating that the macroscopic evolution of the guide field naturally bridges the hot-onset and impulsive phases of solar flares without requiring a separate trigger mechanism.
Read the full abstract

Combinatorics

On the Critical Window for Adaptable 2-Colorability

We determine a sharp threshold for the adaptable 2-colorability of a random graph equipped with a uniformly random, not necessarily proper, red/blue coloring of the edges. To accomplish this, we characterize a family of subgraphs along with edge colorings whose inclusion or exclusion determines adaptable $2$-colorability. We further show that above the threshold, a long path with alternating edge colors is formed. We use this path to prove the existence of such a subgraph in the supercritical regime. We then provide and prove symmetric bounds on the critical window for $2$-adaptable colorability. Particularly, we prove bounds matching that of the critical windows for the giant component in the Erd$\H{o}$s-R$\'e$nyi random graph model as well as the satisfiability of a random $2$-SAT instance. Finally, we show that below the critical window, the solution space of adaptable $2$-colorings remains connected, that is one can travel from one adaptable $2$-coloring to another by a sequence of $2$-colorings which differ on $O(\log{n})$ many vertices.
Read the full abstract

Algebraic Geometry

Siebert's formula for virtual pullbacks

For a DM-type morphism of algebraic stacks $f \colon X \to Y$ with a relative perfect obstruction theory $\mathbb{E} \to \mathbb{L}_{X/Y}$ having a global resolution, we give a formula for the induced virtual pullback in terms of $\mathbb{E}$ and $f$. We apply our formula to prove functoriality properties of the virtual fundamental class in Gromov-Witten theory.
Read the full abstract

Optics

Frequency-Multiplexed Photonic Reservoir Computing with a Synchronously Pulse-Driven Optical Cavity

In this work, we introduce a new platform for frequency-multiplexed reservoir computing based on an optical cavity driven by pulses whose repetition period is matched to the cavity roundtrip time. The input symbols are encoded by modulating either the amplitude or the phase of the driving pulses. Our numerical results show that the proposed synchronously pulse-driven cavity operates in a weakly nonlinear regime under anomalous dispersion, while under normal dispersion it exhibits optical bistability with a high-peak-power upper-branch state. By isolating the key physical parameters, we quantify their individual contributions to information-processing performance. We further show that spectral symmetry breaking induced by third-order dispersion and Raman scattering almost doubles the information-processing capacity. Finally, we demonstrate that modulation-induced branch switching within the bistable regime limits stable reservoir operation.
Read the full abstract

Data Structures and Algorithms

Exponential Lower Bounds for Integer-Weighted Shortest-Paths Preservers of DAGs

We study a graph simplification problem introduced by Bernstein, Bodwin, and Wein [ITCS'24]. We start with a graph with arbitrarily large positive edge weights and the goal is to reweight the edges to small aspect ratio (ratio between largest and smallest weight) while preserving the shortest paths structure (the sequence of vertices and edges along shortest paths). They studied whether polynomial aspect ratio is always possible. They proved that for general graphs, both directed and undirected, it is not: there exist graphs for which any shortest-paths preserving reweighting requires exponential aspect ratio. In contrast, they showed that every DAG (directed acyclic graph) admits a reweighting with linear aspect ratio. However, the resulting edge weights are not integers. This motivated them to pose the open question of whether all DAGs admit a reweighting with polynomially-bounded integer edge weights. Our main result is to answer this question in the negative: we prove that there exist DAGs for which any shortest-paths preserving integer reweighting requires weights of size $2^{\Omega(n)}$. In fact, this is even true when the DAG has very simple structure: 3 layers of vertices with only 3 vertices in the middle layer. In contrast, we show that if the number of vertices in the middle layer is decreased to 2, then a linear upper bound is possible. We extend our exponential lower bound to the approximate version of the problem where only a single $\alpha$-approximate shortest path in the original graph must be preserved as an exact shortest path in the reweighted graph. Our exponential lower bound holds even for any finite approximation ratio $\alpha>1$.
Read the full abstract

Theoretical Economics

Informational Effects in Conflict Escalation

We study a war of attrition in which each party chooses how far to escalate before conceding. If neither concedes before escalation reaches a random disaster threshold, both receive nothing; otherwise the confrontation ends peacefully. When the parties are ex-ante symmetric and their benefits are independent, improving both parties' private information about their own stakes raises both the probability of peace and each party's ex-ante equilibrium utility. The result holds for every nontrivial mean-preserving spread of posterior values—each party's expected benefit from prevailing given its private information. Higher stakes work in the opposite direction: a first-order stochastic increase in posterior values induces greater persistence, lowers the probability of peace, and can reduce both parties' ex-ante utilities. Starting from symmetry, a sufficiently small improvement in only one party's information also promotes peace; its first-order effect is exactly half that of improving both parties' information. At an asymmetric starting point, however, even a small one-sided improvement can instead lower the probability of peace.
Read the full abstract

Formal Languages and Automata

Stochastic Hybrid Automata for Power Profile Modeling in Energy Systems

Power profiles are widely used to describe power demand and production in energy systems. Yet, real-world usage often involves uncertainty, making it challenging to determine, e.g., whether a battery can reliably meet a given profile. Formal modeling enables the computation of probabilities that a battery will satisfy such demands. Given the combination of discrete and continuous dynamics, along with uncertain usage patterns, stochastic hybrid automata (SHA) offer a suitable framework for such modeling and analysis. In this paper, we formalize two orthogonal approaches to specify the (uncertain) power profiles that a specific battery can serve as well as a storage model. For both types of power profiles, we provide formal construction rules for SHA reflecting the behavior of the storage model given a specific power profile. We analyze how design choices, such as the power profile, storage model, and representation of uncertainty, affect the complexity of the resulting SHA. Finally, we demonstrate how these models quantify feasibility using the tools Modest and realyst, and assess how the analysis scales with profile size and the number of stochastic components.
Read the full abstract

2609.12209 · 10 Sep · This paper is accepted for publication (without appendix) in the Proceedings of the 2026 International Colloquium on Theoretical Aspects of Computing (ICTAC). The appendix was part of the submission and provides additional material which is not included in the ICTAC publication

Hardware Architecture

Vortex: Bridging Extreme Compression and Efficient LLM Inference

Extreme compression techniques, including vector quantization (VQ) and input-dependent sparsity, can significantly reduce the memory footprint of large language models (LLMs). However, a key challenge remains in translating such compression into practical efficiency. On conventional systolic-array-based accelerators, VQ incurs high dequantization overhead, while the irregular patterns of input-dependent sparsity are difficult to exploit. In this study, we address these challenges with Vortex, an architecture compatible with systolic-array-based accelerators with minimal hardware overhead, bridging the gap between extreme compression and efficient inference. Vortex adopts a bi-flow execution strategy that efficiently supports vector-quantized models across both prefill and decoding workloads, and we further optimize it through systematic design space exploration. On the algorithm side, we propose codebook-wise contextual sparsity to align with VQ execution. Across end-to-end workloads, Vortex achieves $8.03\times$–$23.7\times$ speedup and $5.68\times$–$12.5\times$ energy reduction over state-of-the-art accelerators.
Read the full abstract

All authors Haoxuan Shan, Cong Guo, Bowen Duan, Chiyue Wei, Feng Cheng, Yuzhe Fu, Yintao He, Hai "Helen" Li, Yiran Chen