Data Structures and Algorithms
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Tuesday 1 September
Sensitivity oracles preprocess a graph so that queries can be answered after any $f$ edge insertions and deletions, without recomputing from scratch. For structural optimization problems the known landscape is limited: for flows and cuts, all known compact oracles handle only $f\le2$ failures; existing oracles for $s$- and global min-cut apply only to undirected graphs; and for matchings, arborescence and spanning-tree packings, and arboricity, no efficient oracle is known for $f>1$. We present a unified algebraic framework based on sensitivity oracles for matroid packing, covering, and parity of sparse linear matroids, yielding the first oracles supporting an arbitrary number $f$ of updates across all of these problems (all constructions randomized Monte-Carlo). Concretely, we obtain efficient oracles for exact $(s,t)$-max-flow/min-cut, resolving an open problem of Baswana, Bhanja, and Pandey (ICALP'22) with near-optimal space; for all-pairs $k$-bounded flow, generalizing the near-optimal reachability oracle of Brand and Saranurak (FOCS'19, the case $k=1$); the first oracles for any $f$ for directed $s$- and global min-cut; oracles for $k$-disjoint arborescences, $k$-disjoint spanning trees, colorful spanning trees, and arboricity; and oracles for the existence of an $\alpha$-factor, with perfect matching as the case $\alpha=1$. We further introduce the subset sensitivity model, in which updates are confined to a susceptible edge set of size $\sigma$ fixed during preprocessing. Here we decouple updates from the matroid representation and eliminate the dependence on $k$ and the matroid density altogether: all of the above are supported with $\widetilde O(f^\omega)$ query time and $O(f\sigma^2)$ space. We also prove a matching $\Omega(\min\{\sigma^2,n^2\})$-bit lower bound when $f\ge2$, establishing optimality.
Computer Vision
HiLRP: Toward One Trustworthy Explanation for Vision Transformer: Conservation-Valid Attribution via Attention Primitives
Vision Transformer (ViT) design has become increasingly diverse, with backbones combining convolutional stems, windowed, linear, or multi-axis attention, patch merging, and spatial reduction in various configurations. This diversity poses challenges for existing attribution methods, whose assumptions often do not hold across ViT variants: Grad-CAM requires a terminal spatial feature map, attention rollout assumes global softmax attention, and layer-wise relevance propagation (LRP) requires module-specific rules. To the best of our knowledge, no existing method provides a unified attribution framework across this architectural space. We show that this architectural diversity can be captured by a simpler underlying structure. The attention and resolution-reduction operators in current ViTs can be decomposed into four operation types: linear maps, bilinear mixing, normalization or gating, and reindexing. Each operation admits a relevance rule that satisfies conservation. Based on these rules, HiLRP supports new backbones by construction rather than by architecture-specific derivation, and its attribution maps decompose the prediction rather than relying on heuristic assumptions. We prove conservation and conditional equivariance and verify both to machine precision. Across 14 attribution methods and 10 architectures, we find that no prior method remains reliable across ViT families, while Faithfulness Correlation becomes uninformative for backbones robust to spatial masking. HiLRP alone preserves conservation across windowed, spatial-reduction, multi-axis, and linear-attention models, where naive extensions can produce zero or inflated relevance. It also localizes attribution failures in class activation mapping, achieving 0.97 Pointing compared with 0.55 for competing methods on EfficientViT.
Read the full abstract
Robotics
EmbodiedSkills: World Action Model-augmented Cascaded Skills for Vision-Language-Action Agent
Vision-language-action (VLA) policies generate executable robot actions, but plausible proposals can have different consequences for the task. We present EmbodiedSkills, a skill-based agent framework that augments frozen VLA policies with world action models for consequence-aware action selection. The framework defines reusable skills for scene observation, context encoding, candidate generation, consequence prediction, outcome evaluation, candidate comparison, and action execution. Five stages organize their input-output dependencies in a shared feedback loop, accommodating both continuous action chunks and pick-and-place primitives. At each decision point, independently trained world action models predict the latent consequences of a common set of policy proposals. Selection combines their outcome scores and prediction disagreement with a lightweight comparator that estimates the benefit of an alternative from agreement in its predicted effects. The predictive and comparison modules learn from matched-state candidate branches, while the action policy and pretrained observation encoders remain frozen. On MetaWorld MT50 and CLIPort, EmbodiedSkills improves success rates from 76.28% to 78.64% and from 33.94% to 36.83%, respectively. An additional CALVIN evaluation with the frozen FLOWER policy increases five-task-chain success from 77.10% to 77.90%. These results demonstrate the applicability of consequence-aware selection across different policy backends and action representations.
Read the full abstract
Quantum Physics
Dynamical regimes of QAOA gradient response
Characterizing the trainability of the Quantum Approximate Optimization Algorithm (QAOA) requires understanding how its gradient landscape changes across circuit parameters and problem size. Yet these gradients are usually described in terms of the native QAOA angles, making it difficult to distinguish parameter specific features from broader changes in the underlying circuit dynamics. Here we introduce a dynamical representation of the QAOA parameter space based on a norm-weighted layer strength and a cost–mixer imbalance, separating the overall scale of the evolution from the relative contribution of the two generators. Using exact-state simulations of MaxCut, we find that the gradient landscape exhibits a coarse organization in these dynamical variables that persists across changes in circuit depth and schedule structure, while the finer interference pattern remains schedule dependent. Near-optimal solutions do not simply coincide with the largest local gradients, but instead occupy a distinct intermediate dynamical regime. Uniform schedules recover the broad location of this regime, whereas nonuniform schedules mainly reorganize its fine structure. Across the system sizes studied, near-optimal solution regions remain extended in the dynamical representation while their preimages in the native QAOA angles become substantially compressed at larger sizes. These results separate the persistence of useful QAOA dynamics from their accessibility in the native parameterization, and provide a dynamical framework for interpreting QAOA trainability across circuit and problem scales.
Read the full abstract
Computation and Language
Some Emotions Run Deeper: Layer-wise Probing and Causal Intervention in Large Language Models
Emotion is expressed in text along a wide spectrum, from surface lexical cues to inferences entangled with content. Most layer-wise analyses of emotion in LLMs use a single corpus, leaving open whether the depth at which emotion becomes accessible is a property of the model or also of the text source. We investigate this across three datasets spanning different degrees of explicitness and contextualization in emotion expression (Twitter posts, Reddit comments, and autobiographical narratives) and eight 1B–9B open-weight LLMs from the Llama, Qwen, and Granite families. We combine layer-wise probing with offline feature scaling and online forward interventions, transfer analyses, and an early-exit classifier. We find that (i) the best probing layer shifts systematically across corpora, from input-adjacent layers to over half model depth, and this ordering persists after matching label-by-length-bin distributions; (ii) across the evaluated settings, forward-pass interventions on probe-selected bands reduce test accuracy by 5–6 points more than same-width random bands ($q < 0.01$); (iii) selected bands transfer across datasets and emotion categories, suggesting partially shared affective information rather than strictly per-emotion substrates; and (iv) probe-selected early-exit representations outperform full-depth exits by $6.9$ percentage points on average.
Read the full abstract
Statistical Mechanics
An exact and fast solution of the inverse Regularized Optimal Transport problem
Optimal transport describes the most efficient way to move mass between two distributions, given a cost matrix for moving mass between each pair of locations. Entropic optimal transport, solved via the Sinkhorn algorithm, is a widely used regularized version of this problem. Its inverse problem asks the opposite question: given an observed transport plan, what cost matrix produced it? This is difficult because the cost is identifiable only up to an additive gauge freedom. Here we show that this freedom can be fixed exactly by a single double-centering operation applied to the observed plan, yielding the true cost matrix in closed form, with no iterative optimization required. When a modest number of true cost entries are known, the same approach lets us jointly estimate the temperature parameter controlling the entropic regularization, together with a diagnostic for the reliability of this estimate. We further show that the method is not specific to the entropic optimal transport, but extends to a broader class of transport models defined by an invertible relation between cost and plan.
Read the full abstract
Computer Vision
TimeSteer: Inference-Time Speech Scheduling in Joint Audio-Visual Diffusion Models
Although pretrained joint audio-visual diffusion models offer rich control over what to generate, they provide no explicit control over when an utterance should occur. To address this, we study inference-time speech scheduling, a novel task that places coupled speech and visual articulation within user-specified begin–end intervals without finetuning the backbone model. We uncover two intrinsic properties of the denoising process that enable this task. First, a timing-sensitive text-to-audio cross-attention head exposes each utterance's model-implied source span along the latent timeline. Second, the predicted clean latent already organizes coupled speech and visual articulation, allowing their temporal placement to be edited without regenerating the content. Building on these discoveries, we propose TimeSteer, a training-free framework that localizes each utterance's source span through Source Span Localization and transfers the associated audio-visual latent content from the source interval to the specified target interval through Region-Aware Latent Remapping. We further introduce SpeechShift, the first benchmark for interval-level speech scheduling in joint audio-visual generation. Experiments across two representative backbones show that TimeSteer substantially improves interval controllability over training-free baselines while maintaining competitive overall generation quality.
Read the full abstract
Computer Vision
Seeing the World and the Self from Egocentric Video
Complete 3D perception from egocentric video requires recovering the surrounding scene and the wearer's full-body motion in a shared metric frame. Existing methods typically address scene reconstruction and motion estimation separately: scene reconstruction methods ignore the wearer, whereas motion estimation methods lack explicit scene geometry and often depend on external trajectories. Joint recovery is challenging because the two tasks exhibit asymmetric visibility and require different prediction paradigms. The largely visible scene supports deterministic geometric regression, whereas the severely occluded body requires generative motion inference. We therefore propose RESELF (REconstructing the Scene and the sELF), a unified framework that couples deterministic metric geometry reconstruction with geometry-conditioned motion generation. RESELF adapts a geometry foundation model pre-trained on large-scale exocentric data to egocentric video using frame-wise scale and relative-pose consistency objectives. The resulting camera trajectory and latent geometric features condition a diffusion model that recovers the wearer's motion. A subsequent closed-loop kinematic feedback stage further refines the camera head while preserving the reconstructed scene geometry. To support training and evaluation, we curate EE4D-JSM from EgoExo4D by aligning egocentric video, sparse metric scene geometry, camera trajectories, and full-body motion annotations. Experiments show that RESELF outperforms state-of-the-art methods designed for the individual tasks across depth estimation, camera tracking, and full-body motion estimation. Code, models, and datasets will be available at https://ka1guan.github.io/RESELF/.
Read the full abstract
Machine Learning
The Constitutional Coverage Trilemma in AI Governance
Frontier AI systems function as constitutional institutions: each deployed model encodes an implicit ranking among safety, helpfulness, honesty, autonomy, and equity. We ask whether the supply of frontier constitutional types covers human demand. Combining a paraphrase-controlled audit of the as-shipped default constitutions of $23$ frontier LLM archetypes with a pairwise-tradeoff study of $1{,}649$ US participants on the same instrument, we report three facts. Demand is broad: it spans all five values, with the largest constituency under one-third. Supply is narrow and drifting: the $23$-archetype hull occupies ${\sim}2\%$ of the demand hull under conservative noise-matched estimation ($0.10\%$ at full audit precision), no archetype puts helpfulness or autonomy first ($37\%$ of users are constitutionally homeless), and across six model families autonomy decreases in $5/6$, equity increases in $5/6$, and safety increases in $4/6$, with monotone within-family version trends (order-permutation $p = 0.013$) and the autonomy decline concentrated in scenarios where safety is not at stake. The drift's importance is directional: away from a value already undercovered, mechanically worsening the welfare floor for the least-served users. The fix is sparse: a $2$-vertex menu $\{e_{\mathrm{HON}}, e_{\mathrm{AUT}}\}$ beats the full $23$-archetype frontier by $47\%$ on mean regret (CI $[43\%, 52\%]$); three vertex additions cut mean/worst-group regret by up to $81\%$/$64\%$. We formalize these findings as a budgeted-pluralism trilemma, show the binding regime is empirically realized, and verify the conclusions are robust to distance-based welfare and to degraded routing. The instrument and audit harness are described in full in the appendices.
Read the full abstract
Computation and Language
From Base Rollouts to RL Reasoning: A Budgeted Search Perspective
Reinforcement learning with verifiable rewards (RLVR) improves language-model reasoning, but how these gains relate to inference-time decoding and search remains unclear. Does RL create reasoning the base model lacks, or shift the rollout distribution toward trajectories it can already reach but rarely samples? We study this behaviorally with a Unified Decoding Framework (UDF), which expresses token-level sampling, beam-like search, tree search, and sequence-level resampling as executable policies over a shared budgeted operating space, scored post hoc with pass@$k$, self-consistency, best-of-$N$, and first-finish success. Using paired Base/RL checkpoints from SimpleRL-Zoo, we ask whether an RL default-policy curve can be approximated by a structured path of Base operating points. On Math500, AIME, GPQA, and IFEval, the pass@$k$ recovery path follows a Budgeted Operating-Point Transition Rule (BOPTR), $N_{\mathrm{Base}} \approx \alpha N_{\mathrm{RL}}^{\beta}$, with benchmark-conditioned exponents. On Qwen2.5-7B, BOPTR gives the lowest transfer error among the non-oracle rules we test, 3.41 pp (95% CI [2.32, 5.53]); a three-seed replication gives 3.07 $\pm$ 0.39 pp. The rule extends to ten models across four families (3.28 to 4.87 pp on checkpoints added after fitting), to four benchmarks it was never fitted on (5.03 pp vs. 4.44 pp in fit), and holds without an RL checkpoint for the target model (4.19 pp) or without RL supervision of any kind (5.08 pp). These results support a qualified internalized-search reading: under the recipe we test, much of the measured RL gain corresponds to a change in sampling efficiency toward operating points the base model can already reach under search. We treat the scaling patterns as descriptive of this recipe and cohort, report where they break down, and use UDF and BOPTR as behavioral diagnostics rather than evidence of parameter-level equivalence.
Read the full abstract
Machine Learning
Position: Privacy Is a Claim, Not a Property of Synthetic Data
Synthetic data has become a common component of machine learning research. While widely adopted, its use in privacy-sensitive contexts has quietly shifted from a claim of residual inference risk under stated assumptions to an appearance-based property inferred from data generation itself. In this position paper, we argue that this shift reflects an implicit change in community standards for what counts as sufficient privacy evidence, rather than a misunderstanding of well-established privacy principles. Drawing on an empirical analysis of recent publications across major ML venues, we show that synthetic data is frequently used in privacy-sensitive settings without explicit articulation of threat models, inference risks, or falsifiable privacy claims. As a result, privacy assurance often remains implicit, difficult to verify, and unevenly distributed, with heightened exposure for rare and minority records. We argue for treating privacy as an explicit, evidence-based scientific claim and recommend that ML venues adopt norms requiring privacy-relevant assertions to be clearly scoped, testable, and contestable.
Read the full abstract
Artificial Intelligence
Making Prospective Memory SLM-Shaped: Typed Intention Stores for Small-Model Agents
Prospective memory means carrying out a deferred intention at the right future cue while other work continues. Benchmarks now isolate it as an agent skill, yet frontier LLMs still struggle: the best published PM-Bench scaffold reaches only 65.1% Set-F1. We argue that this loop is schema-constrained state tracking rather than open-ended reasoning, and that small models can execute it when the action space is typed. We propose the Prospective Intention Store (PIS) that puts lifecycle logic in code and scoped language work on the model. The scaffold is agentic and training-free: no selector fine-tuning and no trajectory distillation. On PM-Bench, DeepSeek-Chat with PIS reaches 82.9% Set-F1. On Gemma-E2B, Set-F1 is only 4.2% without a store and at most 6.6% under seven retrospective memories, while PIS reaches 66.2%. PIS further reaches 70.1% Set-F1, where retrospective memory methods stay at most 54.4%. PIS sets a new state of the art on this benchmark and enables small models to surpass the published large-model scaffold.
Read the full abstract
Software Engineering
Choosing Coding-Agent Benchmarks for Practice: Lessons from Task Demands and Agent Behaviour
Software teams evaluating coding agents need evidence that reflects the changes they expect those agents to make. Benchmark scores summarize how often an agent resolves tasks under broad labels such as bug fixing and feature implementation, but provide little guidance on whether the evaluated work resembles a team's workload. To support this decision, we characterize 2,487 tasks from five public benchmarks and analyze 14,922 runs of six agent configurations. We introduce the Spread-Novelty-Centrality (SNC) profile to describe how the reference solution is distributed across files and directories, how much it adds relative to what it removes, and the dependencies, history, and complexity of the code it touches. We also compare the scope of agent patches and their file exploration. Every pair of benchmarks differs on at least two SNC axes, including benchmarks with the same task label. Among tasks resolved by Claude Opus 4.6, patches are larger than the reference on FeatBench and smaller on FeatureBench. Across the six configurations, resolved runs tend to involve less distributed changes and code with lower values on the Centrality indicators, while the patch sizes associated with resolution differ between the Claude and Qwen configurations. These observations suggest that practitioners should select evaluation tasks based on their change characteristics, inspect the information provided in task descriptions, and assess patch scope alongside test outcomes. We explain how teams can apply this analysis to completed changes and identify the additional evidence needed before using it for task assignment or review automation.
Read the full abstract
Combinatorics
Farey Structure in Modulo Krinkle Tilings: Mediant Splicing and Generation of Prototiles from a Single Edge
The Modulo Krinkle tilings of Imura (arXiv:2506.07638) form a family of non-periodic, spiral monohedral tilings parametrized by a reduced fraction $m/k$ and an integer $t\ge 2$. We show that the Farey sum (mediant) $(m_1+m_2)/(k_1+k_2)$ of two Farey-adjacent parameters is realized by an exact geometric operation on prototiles: the lower boundary path of the $(m_1+m_2,k_1+k_2)$-prototile is obtained by concatenating the parents' lower paths after an edge-length-preserving progressive rotation (fan-twist) of their edges. Conversely, every prototile admits exactly one fan-twist splice decomposition – no non-adjacent parameters ever splice – the cut position being $k_1=m^{-1}\bmod k$, and the recursion descends the Stern-Brocot tree to a single unit edge. The combinatorial core of the operation is the classical standard factorization of Christoffel words; the contribution here is its exact edge-isometric realization on circular direction systems and the resulting structure theory for the Modulo Krinkle family, including the recently introduced variants: we prove a separation theorem stating that every variant prototile is the common recursively-generated core plus finitely many direction-invariant decoration edges. As a corollary of Imura's spiral-arm count formula, the two Farey parents are visible in the offset-free tiling itself: the numbers of counterclockwise and clockwise spiral arms are $tk_1$ and $tk_2$.
Read the full abstract
Spectral Theory
The Faber-Krahn inequality for $p$-Hermite operators
We prove a Faber-Krahn inequality for the first eigenvalue of the $p$-Hermite operator (the weighted $p$-Laplacian with Gaussian weight) on Lipschitz domains in $\R^n$ under Robin boundary conditions with positive Robin parameter. The main result states that, among all domains of given Gaussian measure, the first eigenvalue is minimized by a half-space, and equality holds only for half-spaces. This extends the classical Faber-Krahn inequalities for the $p$-Laplacian \cite{BucurCV} to the $p$-Hermite operator and generalizes the linear case \cite{ChiacchioMathann} to the full nonlinear regime $p>1$.
Read the full abstract
Algebraic Geometry
Resultant multiplicity via projective degrees and applications to tensor eigenvalues
Given a system $\mathbf{f}=(f_1,\ldots,f_n)$ of $n$ homogeneous forms in $n$ variables of the same degree, Macaulay's resultant vanishes precisely when the polynomials have a common projective zero. Its order of vanishing measures the singularity of the resultant hypersurface at $\mathbf{f}$. In this paper, we study how this multiplicity reflects the geometry of the projective zero scheme defined by $\mathbf{f}$. We give an exact formula for the multiplicity, expressed in terms of the projective degrees of the rational map defined by $\mathbf{f}$. As a consequence, we obtain a geometric lower bound involving the degrees, dimensions, and multiplicities of the irreducible components of the projective zero scheme. This extends the multiplicity estimates of Roy and Ghidelli from zero-dimensional schemes to schemes of arbitrary dimension. Finally, we apply this geometric estimate to tensor eigenvalues. It translates directly into a lower bound for the algebraic multiplicity of a tensor eigenvalue in terms of the geometry of its eigenscheme. This settles a conjecture by Canino et al. and consequently settles earlier conjectures of Qi and of Hu and Ye concerning the relationship between algebraic, geometric, and span multiplicities of tensor eigenvalues.
Read the full abstract
Quantum Physics
A nonabelian anyon violates Haag duality
We show that a superselction sector describing a single nonabelian anyon violates Haag duality, which is equivalent to the fundamental quantum-information theoretic principle of uniqueness of purifications. We provide a simple physics argument as well as a rigorous proof using sector theory, and instantiate our result concretely in Levin-Wen models. We also show that the associated ground state does not allow for quantum steering despite being a pure state. Finally, we show that this ground state violates approximate Haag duality, disproving the conjecture that all gapped ground states satisfy this condition. Our result implies that there are gapped phases of matter where Haag duality fails at every point in the phase.
Read the full abstract
Soft Condensed Matter
Intruder dynamics in granular media under localized surface loading
We experimentally investigate the dynamics of a spherical intruder driven horizontally at a constant force in a granular medium subjected to a localized surface overload. While intruder motion beneath a free surface exhibits constant acceleration in the quasistatic regime, the presence of a surface load induces a pronounced transient deceleration when the intruder passes below the loaded region. The magnitude of this deceleration increases with the applied overload and saturates at large overloads, while it decreases with intruder depth. Introducing a characteristic timescale and an overload-based Froude number, we show that the deceleration dynamics collapse onto master curves. We further develop a model incorporating stress transmission from the surface, which partially captures the intruder deceleration. In this approach, this deceleration is shown to depend on two parameters: the overload and the area on which this overload is applied. These results provide a framework to quantify how localized surface stresses influence subsurface intruder dynamics, with implications for locomotion, root growth, and underground transport in granular media.
Read the full abstract
High Energy Astrophysical Phenomena
AGN X-ray coronae at high luminosity: A broadband study of RBS 229 and PG 1407+265
Context. The X-ray emission of active galactic nuclei (AGNs) is commonly interpreted as arising from Comptonisation in a hot corona above the accretion disc. A prominent soft X-ray excess is also observed in many sources and is often explained by a warm Comptonising layer. Aims. We investigate the X-ray-emitting coronae of the luminous radio-quiet quasars PG 1407+265 ($z=0.94$) and RBS 229 ($z=0.334$), testing whether a two-corona scenario consistently describes their broadband spectra. Methods. We combine XMM-Newton EPIC and Optical Monitor data with simultaneous NuSTAR observations and long-term Swift monitoring, and model the spectra using Comptonisation models. Results. Both quasars are X-ray loud relative to the standard $\alpha_{\rm OX}$-$L_{\rm UV}$ relation, with PG 1407+265 showing the strongest deviation. Their spectra are well reproduced by a warm, optically thick Comptonising component accounting for the soft excess and a hot corona producing the hard X-ray continuum. Both sources show high-energy cut-offs at approximately 70-80 keV. RBS 229 exhibits moderate reflection and a weak neutral Fe K-alpha line, while no significant reflection features are detected in PG 1407+265. The warm-corona parameters of RBS 229 remain consistent across epochs, whereas PG 1407+265 shows strong variability, including an extreme soft-excess episode extending to approximately 3 keV in the rest frame. Conclusions. These results constrain AGN coronae in the poorly explored high-luminosity regime. The cut-offs are broadly consistent with trends in AGN luminosity and accretion rate and remain below the runaway pair-production limit. The contrasting variability suggests that the warm corona responds to changes in the accretion flow while remaining within a narrow range of physical conditions.
Read the full abstract
Quantum Physics
DF-SQD: Deterministic Fields for Sampling-Based Quantum Diagonalization
Sampling-based quantum diagonalization method exploits Quantum-centric supercomputing platforms to sample bitstrings for Hamiltonian projection on a quantum computer, and then classically diagonalize the Hamiltonian to estimate the eigenvalues and eigenvectors. In current quantum devices an algorithm is useful when shallow quantum circuits with error mitigation support can discover better results while having either a proof of convergence or some method to explain trust in experiment. In this paper, we introduce DF-SQD, a hybrid algorithm that derives deterministic auxiliary-field circuits from selected double-factorization leaves of the two-electron tensor. The circuits propose occupation-number configurations, while selected configuration interaction evaluates the original active-space Hamiltonian and can recentre subsequent proposal rounds. On N2 (32 qubits; 6-31G basis) and a 40-qubit [Fe2S2(SCH3)4]2- active-space Hamiltonian, we show that DF-SQD improves the energy obtained from sampled determinant spaces while using shallow number-preserving circuits in both simulator and hardware runs. For N2, DF-SQD is 45x more accurate with a 11.23% smaller subspace, and due to its ability to sample better bitstrings at lesser shots it is 2.93x faster than SQD in quantum devices. For the iron-sulfur cluster, DF-SQD generated a subspace dimension of 221M with 400K shots, while SQD needed 1.5M shots to generate a 238M subspace, thus we have better subspace recovery evident from the hardware at 3.75x reduced shots. At a matched 50M subspace dimension, DF-SQD is 1.32x more accurate (achieves a 24.5% relative error reduction over standard SQD). So overall, our method is able to discover better results with shallower circuits, is sample efficient, uses configuration recovery (so has targeted error mitigation) and we have empirical convergence observation.
Read the full abstract
General Economics
AI and the Economy: An Economic Examination of Production, Distribution, Firms, Labor, and Welfare
Artificial Intelligence (AI) is rapidly transforming economic systems by altering production processes, labor markets, and the structure of firms and industries. AI is not merely a technological innovation. It is fundamentally a major economic phenomenon and a new wave of innovation with important implications for productivity, employment, market structure, public policy, long-run economic growth, and collective welfare. This essay examines the economics of AI by analyzing the multiple channels through which AI influences economic activity and societal well-being. It argues that AI should be understood simultaneously as a form of capital, a form of synthetic labor, a general-purpose technology, as well as an economic infrastructure. The analysis shows that AI functions as a multi-channel engine of productivity growth through automation, augmentation, optimization, prediction, and innovation. It is argued that AI has the potential to increase output, reduce costs, stimulate entrepreneurship, and accelerate economic growth, while also reshaping labor markets through labor substitution, labor complementarity, and creative destruction. The essay further examines AI's effects on competition, market concentration, and entrepreneurship. Finally, it explores the welfare implications of AI for consumers, producers, workers, governments, and society as a whole, and concludes by outlining policy considerations aimed at maximizing the benefits of AI while mitigating its potential adverse consequences.
Read the full abstract
Machine Learning
Solving In-Table Prediction Problems by Deep Neural Networks with Performance Evaluation Using Synthetic Data
Tabular deep learning (TDL) leverages neural networks (NN) to extract patterns from tabular data. Traditional TDL methods follow a supervised learning paradigm, where a target feature is explicitly given. In this work, however, we explore a different approach by employing deep NNs to learn relationships among individual columns within a given table. We investigate whether NNs can predict the values of arbitrarily selected columns in a given table based on the remaining known columns. We call this problem In-Table Prediction (ITB), which is slightly different from table imputation methods and the pretraining task of TDL. Three potential usage scenarios are identified, which, to our best knowledge, have not been extensively studied in the literature. A self-supervised learning approach is applied to address this problem by randomly selecting columns to be masked out and used as learning targets. This work focuses on tabular datasets containing only continuous features. To handle missing values in continuous features, a novel neural layer is proposed to embed both numerical and empty values. Synthetic data is generated based on predefined column relationships, with empty values inserted using two distinct mechanisms. Additionally, an adapted masking strategy is employed to create test data. Performances of three NN architectures, namely MLP, Resnet and Transformer, are evaluated using the generated synthetic data. We conclude that, the attention-based structure outperforms the other two networks, when a sufficiently large number of training examples is available and a relatively large embedding length is chosen. We stress that these findings are obtained under controlled, synthetic conditions with a small number of columns and it should therefore be regarded as an initial, narrowly-scoped investigation rather than a general characterization of ITP on real-world tabular data.
Read the full abstract
High Energy Physics (Phenomenology)
Spin alignment of vector mesons in a plasma at finite density
Holographic predictions for vector meson spin alignment at finite density have been limited by how the chemical potential is implemented. Existing models use a baryon chemical potential and report that the alignment is insensitive to it, while zero-density benchmarks do not address the geometric deformation. We address this limitation by using a charged AdS black-hole background with a quark chemical potential and solving the coupled bulk vector equations with infalling horizon conditions. The zero-density limit reproduces previous results of Sheng et al. [Phys. Rev. D 110, 056047 (2024)]. At \(T=0.15~\mathrm{GeV}\), the \(J/\psi\) spectral function has a resonance peak at \(\mu=0\) that lowers and broadens with increasing \(\mu\), whereas the \(\phi\) spectral function shows no pronounced peak. In the helicity frame, \(\rho_{00}\) deviates positively from \(1/3\) for the \(J/\psi\) and negatively for the \(\phi\). The density response is more visible in our setup than the weak baryon chemical potential sensitivity reported in a four-flavor soft-wall model. This difference illustrates the model dependence of the density response: it is controlled by the chemical potential normalization. We identify the presence of the resonance peak and the signs of the helicity deviations as robust qualitative features, and treat the numerical shifts of \(\rho_{00}\) as model-dependent estimates.
Read the full abstract
Artificial Intelligence
Dual Process Motion Planning
Robotic systems are deeply embedded in both industry and everyday life, where they are expected to act with speed, precision, and reliability. Classical control and planning methods have long delivered strong guarantees, but often at the cost of computational efficiency and adaptability. More recently, learning-based approaches have shown promise in overcoming these limitations, enabling agents to leverage experience to accelerate decision-making and address previously intractable problems. In this work, we bridge these two approaches through a neuro-symbolic perspective on nonlinear motion planning. Inspired by the Thinking Fast and Slow paradigm, we introduce a dual-process architecture that combines the strengths of robust reasoning and learning. Our framework integrates state-of-the-art symbolic solvers as a “System-2” component with experience-driven “System-1” modules. A metacognitive controller dynamically orchestrates their interaction, selecting when to rely on fast intuition versus slower, more precise reasoning. By evaluating the framework across diverse nonlinear benchmark environments, we demonstrate that this architecture yields consistent gains in planning efficiency, accuracy, and generalization, while promoting reuse across tasks. The results suggest that tightly coupling learning with structured reasoning offers a scalable path toward more capable and adaptive robotic systems.
Read the full abstract
Differential Geometry
Ricci curvature for fluid models on the torus via Zeitlin's quantization
Ricci curvature measures the average stability of geodesics under transverse perturbations, but how it should be defined in infinite dimensions is often unclear. This paper proposes a definition of Ricci curvature on the space of Hamiltonian diffeomorphisms on the two-dimensional flat torus $\mathrm{HDiff}(\mathbb{T}^2)$, the state space for ideal fluids. Our definition is based on Zeitlin's model, which approximates $\mathrm{HDiff}(\mathbb{T}^2)$ by finite-dimensional Lie groups $\mathrm{SU}(N)$. We derive a formula for the Ricci curvature tensor on $\mathrm{SU}(N)$ and provide numerical evidence for its convergence in the large-$N$ limit to our conjectured finite value. Additionally, we explore potential applications for hydrodynamics through the Lyapunov stability of gravest wave modes and Arnold's tradewind estimates for long-term weather predictability. Our framework extends to a wide range of settings. We demonstrate this by introducing Ricci curvature on the state spaces of fluids on rectangular domains, the Lagrangian averaged Euler equation induced by the $H^1$-Sobolev metric, and the quasi-geostrophic equation incorporating the Coriolis effect.
Read the full abstract
Mesoscale and Nanoscale Physics
Tunneling characteristics of twisted double bilayer graphene heterostructures
Electron tunneling between sheets of bilayer Bernal graphene twisted at different small angles was studied experimentally and theoretically. The current-voltage characteristics exhibit resonant peaks, steps, and regions of negative differential resistance, the origin of which is explained by the intersections of energy- and momentum-shifted electron dispersions of adjacent layers. A theoretical analysis of tunneling transport demonstrated that the key to understanding this phenomenon lies in the competition between two contributions: between like (conductivity-conductivity or valence-valence) and unlike (conductivity-valence) bands of parallel bilayer graphene sheets. A systematic evolution of the tunneling current patterns with increase of the twist angle is investigated. Polarization of electron wave function across graphene sublayers caused by displacement field within bilayer graphene is shown to strongly affect the tunneling probability, thus enhancing negative differential resistance due to Van Hove singularities at the band edges.
Read the full abstract
Artificial Intelligence
Measuring the Behavioral Fidelity of Long-Horizon Human Activity Simulations
As LLM-based human simulators are increasingly used for policy, evaluation, and training, they must faithfully reproduce real behavioral patterns. While prior work has examined behavioral fidelity in survey responses and dialogue, longer-horizon real-world activity remains largely unexplored. We introduce a framework for evaluating behavioral fidelity in long-horizon activity simulations across temporal granularities and levels of analysis. As a case study, we collect a 43-hour multi-camera dataset of in-the-wild office activity and compare trace-derived conditioning mechanisms: persona descriptors, few-shot exemplars, and statistical transition and time-of-day priors. We find that behavioral fidelity is not uniform across metrics: statistical priors bring activity and sequence distributions closest to real behavior, yet over-fragment routines and suppress within-person variability. These findings motivate a more holistic evaluation that spans multiple metrics, temporal granularities, and levels of analysis.
Read the full abstract
Combinatorics
Hamilton Starters and Path Decompositions in Directed Circulants
For integers q at least 3 and r at least 1, consider the directed Cayley graph on the cyclic group of order qr whose allowed steps are the integers from 1 through r. A Hamilton cycle H is called a q-layer balanced Hamilton starter if, for every step from 1 through r and every residue class modulo q, H contains exactly one arc of that step whose tail belongs to the given residue class. The r translates of H by successive multiples of q then form a Hamilton decomposition of the digraph. A q-layer balanced Hamilton starter is called chain-compatible if one arc can be selected from each Hamilton cycle in this decomposition so that the selected arcs form a simple directed path. In both constructions, the compatible deletion chain is the arithmetic step-2 path beginning at 0 and ending at 2r. For q equal to 4, a chain-compatible starter is obtained explicitly whenever r is congruent to 1 modulo 4 and r is at least 9, while for q equal to 3 one exists for all sufficiently large r congruent to 5 modulo 6. The proofs are constructive: the four-layer case uses an ABAB step word, while in the three-layer case a directed rotational terrace is lifted to a 3-layer balanced directed 1-factor and a fixed four-arc trade joins its two cycles. Deleting the unique prescribed-path arc from each translated Hamilton cycle gives r Hamilton paths; together with the prescribed path, these form an optimal decomposition of the arc set into r+1 directed paths in both cases.
Read the full abstract
Signal Processing
Joint PAPR and OOBE Reduction for AFDM via Chirp Parameter Tuning
This paper addresses the joint reduction of the peak-to-average power ratio (PAPR) and out-of-band emissions (OOBE) in affine frequency division multiplexing (AFDM) systems by selecting the pre-chirp parameter c2. While existing approaches typically optimize either PAPR or OOBE independently, the proposed method jointly considers both metrics. To this end, a weighted cost function combining PAPR and OOBE is introduced to evaluate the trade-off between the two objectives. A pre-chirp selection scheme, inspired by the selected mapping (SLM) technique, is then employed to identify the optimal c2 value from a finite set of candidates, yielding a Pareto-optimal operating point within a discrete set. Simulation results demonstrate that the proposed approach simultaneously reduces both PAPR and OOBE compared with conventional AFDM. Moreover, its performance remains close to that of methods specifically optimized for a single objective, with only about a 1 dB degradation in PAPR reduction and a 2-3 dB degradation in OOBE suppression.
Read the full abstract
Networking
Highly Detailed Simulation for Connected Automated Vehicle Cooperative Driving
End-to-end simulation of connected and automated vehicles requires consistent fidelity across mobility, environment modeling, V2X radio propagation, and decision-making/control modules. However, 2D representations of complex road infrastructure often fail to capture critical signal propagation dynamics, leading to overly optimistic connectivity assumptions. This paper presents a unified workflow within CAVISE that integrates map-based scene preparation with microscopic mobility simulation. The proposed framework incorporates a comparative analysis of trace-driven propagation in 3D versus a simplified 2D baseline, and a modular interface for integrating Autonomous Intersection Management (AIM) models. Collectively, these capabilities enable high-fidelity Cooperative Driving Automation (CDA) experiments. Leveraging ray tracing using Sionna RT at 5.9 GHz with the same radio and solver configuration for both geometry variants, we show that planar reduction in multi-level road infrastructure can remove physically present occlusions and substantially distort signal-loss dynamics. In a bridge overpass case study, the 2D baseline eliminates an occlusion interval observed in 3D, changing the outage behavior from intermittent to consistently connected and yielding an average signal-loss shift on the order of 10 dB. These propagation-induced biases highlight the inability of planar models to capture vertical occlusions, necessitating 3D-aware communication modeling to ensure the validity of cooperative driving automation and intersection control evaluations.