Peer Revenue Impact factor: volatile

Papers

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

All of arXiv

Tuesday 22 September

Soft Condensed Matter

An interdimensional funnel enhances topologically protected oscillations in high-dimensional stochastic systems

Varsha Traynor and Jaime Agudo-Canalejo

Biochemical systems such as protein complexes often occupy configuration spaces whose dimensionality grows with the number of molecular components. Although usually viewed as an obstacle to coherent dynamics, here we show that high dimensionality can enhance topologically protected oscillations. In the topological regime, a stochastic system with a $D$-dimensional configuration space develops a steady-state current confined to a one-dimensional edge cycle. The intervening boundary faces between the $D$-dimensional bulk and the one-dimensional cycle form an interdimensional funnel, sequentially pushing the system towards lower dimensional faces, down to the one-dimensional cycle. As a consequence, increasing $D$ enhances the funnel, leading to improved one-dimensional localization and oscillatory coherence. A quantized biorthogonal Zak phase identifies the transition to the topological regime, which additionally exhibits a highly structured non-Hermitian skin effect. In contrast with previous realizations of topological states in stochastic systems, the phenomena we uncover have no counterpart in quantum condensed matter or active matter systems.
Read the full abstract

High Energy Physics (Phenomenology)

Tension between MiniBooNE and MicroBooNE within a 3+1 Sterile Neutrino Framework using Simulation-Based Inference

Julia P. Woodward, Austin Schneider, Joshua Villarreal et al. (+2 more)

The MiniBooNE low-energy excess and its subsequent exclusion by MicroBooNE provide an important test of sterile-neutrino explanations of short-baseline neutrino anomalies. Such a test is complicated by the fact that both experiments use different approaches to their analyses. In this contribution, we present a consistent approach to both experiments. We perform a joint fit of MiniBooNE and MicroBooNE data to a $3+1$ sterile-neutrino model omitting all data-driven factors, using MicroBooNE data from both the Booster Neutrino Beam (BNB) and Neutrinos at the Main Injector (NuMI) beamlines, and evaluate the parameter goodness-of-fit (PG) tension between the two experimental results. A rigorous frequentist treatment of both parameter fitting and PG tension is challenging. Existing methods require asymptotic assumptions known to be inaccurate for neutrino oscillation measurements or repeated likelihood optimization tasks. To this end, we use a previously developed frequentist fitting framework based on simulation-based inference (SBI) and introduce a new SBI-based method for evaluating PG tension that makes the required trial-based calibration computationally feasible. We find that the addition of MicroBooNE reduces the significance of the MiniBooNE preference for sterile-neutrino oscillations, with a trials-based preference for 3+1 remaining at $2.7 \sigma$. The two experiments exhibit a $\geq 2.5\sigma$ PG tension within the $3+1$ model.
Read the full abstract

All authors Julia P. Woodward, Austin Schneider, Joshua Villarreal, John Hardin, Janet Conrad

Theoretical Economics

Implicit Smooth Ambiguity: Perception, Attitude, and Identification

Kemal Ozbek

This paper develops a model of identifiable smooth ambiguity in a finite-state Anscombe-Aumann framework in which ambiguity perception and the ambiguity-attitude representation may depend on the act's endogenous certainty-equivalent utility level. The framework separates two sources of utility-level dependence that standard smooth-ambiguity models hold fixed. Neutral-transfer experiments within revealed cells recover conditional beliefs, while compensation experiments across cells recover cell weights and locally revealed ambiguity-attitude scales. These objects have distinct behavioral content. After the within-level structure is extended globally, ambiguity perception is identified at each certainty-equivalent level, whereas variation in admissible level-specific global attitude indices need not by itself reflect behaviorally identified variation in ambiguity attitude. Cross-level restrictions determine when the locally revealed attitude scales can instead be represented by one common global ambiguity-attitude index. The baseline characterization yields a doubly implicit smooth-ambiguity representation with level-specific perception and potentially level-specific global attitude indices. Requiring one common ambiguity-attitude index gives an intermediate stable-attitude model in which only perception may vary; imposing perception stability as well yields standard identifiable smooth ambiguity within the maintained full-support, neutrally separated, fixed-partition environment with at least three revealed cells. The contribution is therefore a behavioral decomposition of utility-level dependence in smooth ambiguity into separately revealed perception and attitude components.
Read the full abstract

Physics Education

Raytracing Black Holes with AI Accelerators

Thomas Fischbacher and Nicholas Kessler

We show how software infrastructure that was mainly built to power AI applications also can be used to great benefit for other purposes – by demonstrating a fully self-contained JAX-based black hole raytracer in under 180+200 lines of Python code plus documentation. Major motivations for showcasing such a construction are (a) to show to professional physicists how Machine Learning (ML) accelerators can greatly simplify building numerical applications for GPU accelerator hardware with very modest coding effort and programming expertise – even if the problems at hand do not involve ML per se, (b) to make a meaningful part of General Relativity (GR) accessible to computer science professionals who might not have realized that mathematical background they developed on Deep Learning problems may have brought this theory within easy reach for them, and (c) to inspire students and educators to explore the potential of an approach that represents physical law, i.e. relations between local differential quantities, not in terms of differentiable symbolic expressions, but in terms of differentiable algorithms.
Read the full abstract

Machine Learning

CRISP: Scalable Importance-Stratified Coresets for Imbalanced Tabular Learning

Hardhik Mohanty, Indrayana Rustandi and Mohamadreza Sheibani

Large imbalanced tabular datasets make repeated gradient-boosted tree training expensive. Existing coreset methods often lose accuracy when most majority examples are removed. We present CRISP (Coreset Reduction via Importance-Stratified Pruning), a linear-time method that allocates a negative-class budget across quantile strata of a proxy-model score. Sample weights account for unequal inclusion probabilities. At 95% negative-class reduction on a production fraud dataset, CRISP trains on approximately 1.70M of 25M rows and retains 99.7% of full-data Average Precision. This is a 93.2% reduction in total training rows. On public CriteoPrivateAds, CRISP has the highest mean Average Precision at each tested rate from 90% to 99.4% majority reduction. Sparkov results are mixed at lower rates, but CRISP has the highest mean at 99.2% and 99.4%. Ablations identify budget allocation and inverse-propensity weighting as the main sources of the production-dataset gain.
Read the full abstract

Metric Geometry

A Model for the space of convex quadrilaterals

Carl Eberhart

This paper describes a 'model' for the space of convex quadrilaterals, that is, a set $\mathbb{Q}$ of convex quadrilaterals in the plane which gives a 'cross-section' of the equivalence classes of similar convex quadrilaterals, in the sense that every convex quadrilateral is similar to exactly one member of $\mathbb{Q}$. It is also has the property that quadrilaterals which have nearly the same vertices are close to each other in the model in the Hausdorff metric. We also parameterize our model with the 4-cell and use that to describe and investigate various classes of quadrilaterals. For example, we show that the trapezoids form a 3-dimensional closed subset of the model which separates the model into 3 disjoint open sets, each homeomorphic with a 4-cell. We used SageMath to compose the latex for this note, and to make the Sage Cell Interacts to explore the model, available at sagelets.org.
Read the full abstract

Signal Processing

Untangling the Geometry and Speed for RF Sensing Spectrograms

Mert Torun, Darius Cuenca and Yasamin Mostofi

A fundamental challenge in RF sensing is that Doppler signatures observed by a link entangle the target's motion with the sensing geometry, resulting in limited applicability to unconstrained real-world settings. In this paper, we establish a new foundation for physically interpretable RF sensing that disentangles reflector speed from geometry, jointly recovering the speed, geometry factor, relative amplitude, and width of each dominant Doppler ridge. More specifically, we first develop a compact parametric representation of WiFi spectrograms and establish its low-dimensional structure through a systematic computer-vision analysis of a large and diverse human-activity dataset, thereby providing a tractable foundation for learning. Building on this representation, we then design a physics-informed autoencoder whose structured bottleneck and differentiable RF forward model enforce physically meaningful estimates of reflector speed and geometry. We further introduce a synthetic-to-real training framework, eliminating the need for real WiFi training data. We extensively validate the proposed framework under both known and time-varying geometries, using both independently generated synthetic test sets and 31 real WiFi experiments. The results demonstrate the superior performance in speed and geometry extraction, robustly recovering the underlying geometry, speeds, Doppler-ridge amplitudes, and ridge widths across all settings, while substantially outperforming the strongest baselines.
Read the full abstract

Machine Learning

Transfer Learning with Conformalized Quantile Regression for Solar PV Forecasting Under Load-Shedding-Driven Data Scarcity

Rakib Abdullah and K. M. Tahlil Mahfuz Faruk

Solar photovoltaic (PV) forecasting in regions affected by load shedding is challenging because reliable historical observations are scarce. This study proposes a transfer learning framework combined with Conformalized Quantile Regression (CQR) to improve PV power forecasting and provide reliable uncertainty estimates under severe data scarcity. A source-domain PV dataset from Alice Springs, Australia, is used to pretrain a temporal forecasting model, which is then adapted to simulated Bangladesh PV data representing different levels of historical availability. Experimental results show that transfer learning reduces RMSE by up to 23.7% when only one month of target-domain data is available and by 13.7% with three months of data. The proposed Transfer Learning plus CQR framework achieves 94.3% empirical coverage with three months of target data while producing prediction intervals that are 14% narrower than those obtained without transfer learning. These results demonstrate that combining transfer learning with conformal uncertainty quantification can improve both point forecasting accuracy and uncertainty reliability when target-domain PV data are severely limited.
Read the full abstract

Quantum Physics

Soft decoding for quantum LDPC codes with experimental validation

Arda Aydin, Edwin Tham, Nicolas Delfosse et al. (+1 more)

The decoder is a critical component of a fault-tolerant quantum computer, computing corrections based on parity-check measurements performed throughout the computation. A soft decoder supplements its output with a confidence score which, when used alongside post-selection, can substantially improve logical performance. We introduce a soft beam search decoder for quantum low-density parity-check codes that uses internal decoder data as a confidence metric, removing the need for extra computation. We perform circuit-level simulations of five quantum LDPC codes relevant for superconducting and trapped ion architectures equipped with our global soft decoder and we obtain up to $580\times$ logical-error suppression at a physical error rate of $10^{-3}$ while rejecting only $0.1\%$ of shots. Then, we simulate an error detected measurement, which is a core logical operation of the walking cat architecture, using a streaming version of soft beam decoder and we achieve up to $210\times$ error suppression while increasing the rejection probability by only $0.5$ percentage points. Finally, we revisit recent quantum LDPC code memory experiments on trapped ions, demonstrating that our soft decoder doubles the logical qubit lifetimes at the price of a mean rejection rate of $2.6\%$–$5.6\%$ per syndrome round, bringing all five codes into the beyond-breakeven regime.
Read the full abstract

All authors Arda Aydin, Edwin Tham, Nicolas Delfosse, Min Ye

Strongly Correlated Electrons

Matrix Density Waves and Fractionally Charged Point Defects in Flavor Weyl Semimetals

Shantonu Mukherjee and Hridis K. Pal

A conventional interaction-driven Weyl density wave is a complex scalar and supports vortex lines, but no topologically stable point defects. We show that two Weyl flavors do not merely duplicate this order. Starting from the matrix-valued internode coherence, a local flavor-symmetric repulsion selects a traceless adjoint condensate within the density-wave sector, while the fermionic ground-state energy locks its complex components into the collinear form $\mathbf{\Delta}=\Phi_0\mathbf{n} e^{i\theta}$. The resulting order-parameter manifold, $(S^2\times S^1)/\mathbb Z_2$, supports both unit hedgehogs and half-quantum Alice strings of the same electronic mass that gaps the Weyl fermions. An elementary hedgehog binds a single normalizable zero mode whose empty and occupied sectors carry charges $-e/2$ and $+e/2$ at neutrality. An elementary Alice string carries a single chiral electronic mode despite its $\pi$ phase winding, and transporting a hedgehog around it reverses the hedgehog winding, $N\rightarrow-N$. Flavor therefore converts spontaneous translation breaking into a route to point-defect fractionalization and intertwines the point- and line-defect sectors of an interaction-generated electronic mass.
Read the full abstract

Sound

A Stem-Agnostic Approach to Hybrid AI Music Detection

Richa Namballa, François Rigaud and Romain Hennequin

The inclusion of generative audio in the music production process has led to an increase in hybrid music tracks that blend authentic human performances with AI-generated stems, challenging traditional AI music detectors which operate in a binary setting. In this work, we propose a stem-agnostic framework for identifying synthetic audio sources within hybrid musical mixtures. We introduce the inspectrogram, a novel time-frequency representation that maps localized probabilities of synthetic content across the audio spectrum. By combining the inspectrogram with a Wiener filter estimating target stem energy dominance, a single CNN model evaluates whether the specific stem is generated. Trained on rendered hybrid mixtures and evaluated across various stem classes, our model achieves strong performance on high-frequency sources such as vocals, drums, and guitar, but struggles on the low-frequency, narrow-band bass. We conclude that the quality of separation impacts the detection accuracy and identify source separation as a primary bottleneck and a crucial direction for future research.
Read the full abstract

Machine Learning

When Post-Processing Fairness Constraints Help and When They Harm: Evidence from Eight Cross-Domain Evaluations

Nithin Raghava Ramachandra Narla

Fairness audits in production ML typically occur once, at deployment, on a single domain. Both fail in practice: fairness can shift after retraining or a changing user base, and interventions validated on one dataset are rarely tested across the heterogeneous domains an organization deploys. We present FAPE (Fairness Auditing for Production Environments), a four-stage framework evaluating a single post-processing intervention, Fairlearn's ThresholdOptimizer, across eight domain evaluations: criminal justice, income prediction, legal admissions, credit lending, agricultural lending, a multi-domain benchmark corpus, healthcare, and education. Each is scored on demographic parity and equalized odds difference, plus disparate impact ratio and accuracy cost where computable. Intervention effectiveness tracks baseline disparity magnitude: across model-domain pairs the constraint improved disparity in 9 of 14 high-disparity cases and worsened it in 3 of 4 near-fair ones. Each of the five high-disparity exceptions reverses under one of two measurement checks, a minimum group size or thresholds fit on held-out data. A CUSUM monitor started at deployment, tested on a simulated shift, separates constrained models that never met a 0.1 parity convention from those that met it and later regressed. A single deployment-time audit is therefore an unreliable guide, which argues for baseline-disparity screening and continuous monitoring
Read the full abstract

Logic

Asymptotic dimension of commutative monoid actions

Jinmin Wang, Jing Yu and Jingming Zhu

We prove that every bounded-to-one action of a finitely generated commutative monoid has asymptotic dimension at most the torsion-free rank of its group completion, with control uniform for a fixed monoid, generating set, and generator fiber bound. The estimate is sharp. Together with the equality of classical and Borel asymptotic dimensions for these actions, it gives the sharp Borel rank bound and recovers hyperfiniteness for bounded-to-one Borel actions of countable commutative monoids.
Read the full abstract

Networking

RIS-Enabled Integrated Access and Relay: Empowering Collaboration Among BSs

Hao Lin, Mustafa A. Kishk and Mohamed-Slim Alouini

The increasing number of Internet of Things (IoT) devices and applications leads to severe access congestion in conventional base station (BS) networks. Meanwhile, the low transmit power of IoT devices requires larger diversity gains from the system design. Therefore, low-cost traffic management and signal enhancement mechanisms become essential. Reconfigurable intelligent surfaces (RISs) are increasingly significant due to their flexibility and efficiency. Using reflection, refraction, and amplification, RISs can act as innovative relay nodes to overcome spatial limitations caused by blockages, and improve the coverage range of existing infrastructure. However, more extensive applications of RISs remain to be explored. In this paper, we propose to deploy co-sited RISs on BSs and use the integrated access and relay (IAR) architecture to enable wave-domain task-offloading between BSs, without requiring extra infrastructure or incurring extra decoding latency. Considering the fixed and dynamic sub-carrier schemes, we show the performance improvement in cell-free networks and cellular networks. In addition, we investigate the effect of RIS elements allocated to each IoT device in both cell-free and cellular networks. Finally, we discuss the opportunities and challenges of IAR BSs for future communications design.
Read the full abstract

Artificial Intelligence

Escaping Python Dependency Hell: A Hybrid Replay-and-Repair Pipeline for Python Dependency Resolution

Veronica Poweska, Ariana Oyanguren, Jessica Pourleyli et al. (+2 more)

Dependency conflicts in Python ecosystems arise from incompatible version constraints, missing packages, and undocumented compatibility relationships, causing many real-world code snippets to fail at execution. This paper presents PLLM+, a hybrid dependency-repair pipeline evaluated on the HG2.9K benchmark of 2,891 dependency-failing snippets. PLLM+ prioritizes inexpensive deterministic steps before invoking LLM-based repair: static AST-based interpreter inference, replay of historically successful dependency configurations from the competition-provided solutions database, and live PyPI validation of candidate package versions. When these steps do not resolve a case, the system falls back to a structured LLM-based repair loop with typed error classification and Proposer/Critic agents. On HG2.9K, PLLM+ solves 1,500 out of 2,891 snippets, compared with 1,169 solved by the PLLM baseline. It also reduces average runtime from 368.7 to 71.8 seconds per snippet. Most successful fixes come from replaying known configurations: 1,495 of the 1,500 successful fixes are produced by the solutions database, while the LLM fallback accounts for 5 additional fixes. These results suggest that, in this benchmark setting, deterministic reuse of previously validated dependency configurations is a simple and effective strategy, with LLM-based repair serving as a secondary fallback for cases not covered by prior solutions.
Read the full abstract

All authors Veronica Poweska, Ariana Oyanguren, Jessica Pourleyli, Sourena Khanzadeh, Manar Alalfi

Statistics Theory

Rolling Conformal Prediction in Sequential Model Training

Chen Cheng, Ruiting Liang and Rina Foygel Barber

We introduce Rolling Conformal Prediction (rolling-CP), a distribution-free predictive inference method for the setting of sequential model training. Specifically, given a data stream $(X_1,Y_1),(X_2,Y_2),\dots$, at each time $n$ the trained model may depend on the observed history $\{(X_i,Y_i)\}_{i<n}$. This setting arises naturally in modern sequential training, including one-pass training over massive datasets and continual fine-tuning or test-time adaptation of language models during deployment. Rolling-CP first calibrates each incoming observation against the current predictor and then rolls it into future training. In this way, we avoid the need for data splitting. Remarkably, although the models at times $n=1,2,\dots$ may have entirely different properties and accuracy levels, for exchangeable data it is nonetheless possible to establish a guarantee of marginal coverage, with a familiar universal factor-two guarantee (a worst case guarantee of $1-2\alpha$ coverage, as compared to the target level $1-\alpha$), without any assumptions of stability or any restrictions on the model training process. For i.i.d. data streams, we further prove high-probability training-conditional validity uniformly over time; under stability conditions, coverage guarantees sharpen towards $1-\alpha$. Numerical experiments on sequential regression, multiclass SGD, and one-pass neural-network training further demonstrate the practical effectiveness of rolling-CP.
Read the full abstract

Probability

Geometric mean quantization via adaptive approximation

Marc Kesseböhmer and Aljoscha Niemann

Let $\nu$ be a compactly supported Borel probability measure on $\mathbb R^{d}$ with $\nu(B(x,r))\leq Cr^{a}$ for some $a>0$. Refine a dyadic cube exactly when its mass is at least $t$, and let $\mathcal{L}_{\nu}(t)$ be the mean depth at which this refinement stops. We show that the lower and upper geometric-mean quantization dimensions of $\nu$ are the lower and upper limits of $\log(1/t)/\mathcal{L}_{\nu}(t)$. The dimension exists precisely when $(q-1)\sum_{Q}\nu(Q)^{q}$, summed over all dyadic cubes, converges as $q\downarrow1$, and it is then determined by this limit. The mass-threshold formula yields harmonic integral bounds in terms of the local dimensions and encloses entropy and quantization dimensions in a common spectral interval. Convergence in law of the local information rates is equivalent to convergence of the rescaled spectra in a window of width $1/k$ around $q=1$; the two dimensions are then the arithmetic and the harmonic mean of the limit law, and we quantify their difference by sharp bounds and variance identities. Without any convergence assumption, vanishing threshold variance still forces equality of the corresponding lower and upper dimensions. Bernoulli mixtures realise every local-dimension law with compact support in $(0,1]$, and a regime-switching example separates convergence in law from almost-everywhere convergence.
Read the full abstract

Quantum Physics

Measurement-based feedback as a flexible tool to control non-Poissonian photon statistics

Ming Li, JunYan Luo and Georg Engelhardt

Non-Poissonian photon statistics are usually traced to nonlinear interactions, nonclassical states, or engineered reservoirs. Here we show that measurement-based feedback can be alternatively deployed to control the Poissonian character of the measurement statistics in an externally-driven linear cavity. Each photon detection triggers a weak displacement operation whose phase determines how the feedback interferes with the coherent cavity field, resulting in a significant cavity response. Full counting statistics and two-time photon correlations show that the feedback phase controls the sign and strength of the induced correlations, continuously tuning the emission statistics between super- and sub-Poissonian regimes. At generic phases, the Mandel parameter responds linearly to the feedback amplitude, whereas at specific phases, the leading response becomes quadratic. The same phase dependence is also found in higher-order cumulants of the emission statistics. These results establish measurement-based feedback as a route to controlling photon statistics in otherwise linear driven-dissipative systems.
Read the full abstract

Algebraic Geometry

Open extension of a genus zero Hurwitz-Frobenius manifold

Alexey Basalaev and Anton Rarovskii

Every Dubrovin–Frobenius manifold provides a solution to the WDVV equation, that was initially formulated in the study of the moduli space of curves. During the last two decades special attention was given to the "open" version of the moduli space of curves. Its genus zero intersection theory is governed by the system of equations called open WDVV equation, that extends "classical" WDVV equation. In this note we study Dubrovin–Frobenius manifold structures on the Hurwitz spaces of genus zero. We construct explicitly the solutions to open WDVV equation for these Dubrovin–Frobenius manifolds and show that these solutions appear as the restriction of the Dubrovin–Frobenius structure to its discriminant.
Read the full abstract

Applied Physics

A Green's-function method for vertical thermal boundary conductance in anisotropic multilayers

Dihui Wang, Troy Munro and Heng Ban

Vertical thermal interfaces occur in both engineered and natural materials. Their vertical thermal boundary conductance can differ from the horizontal counterpart, requiring dedicated characterization. Yet current thermal metrology resolves vertical thermal boundary conductance only in restricted geometries such as two bulk media, for lack of an efficient forward solution that admits anisotropy, multilayers, and depth-dependent vertical thermal boundary conductance together. We present a Green's-function boundary integral equation (GBIE) method that couples transfer-matrix Green's functions to an interface-only integral equation for depth-dependent $G_v(z)$, supporting dissimilar orthotropic multilayers ($k_x\neq k_y\neq k_z$) on either side and horizontal conductance $G_h$. For anisotropic film-on-substrate multilayers with films from $1~\mu\mathrm{m}$ to $100~\mathrm{nm}$, the GBIE agrees with three-dimensional finite element method (FEM) predictions to within one percent mean normalized phase and amplitude error, while running $29\text{--}210\times$ faster and reducing peak memory by factors of $120\text{--}450$ in single-core tests; a JIT-compiled JAX implementation reaches up to $4100\times$ on a matched 16-core comparison. The GBIE further reproduces a continuous film over a buried interface, representative of a thermoreflectance measurement, and a finite-depth interface with depth-dependent $G_v(z)$. The GBIE accommodates lateral-to-film-thickness ratios above $10^{5}$, where volumetric FEM can become computationally prohibitive. These results establish an efficient forward solution for vertical-interface heat transport in systems ranging from microelectronic device sidewalls to grain boundaries in polycrystalline solids.
Read the full abstract

Quantum Physics

Monolithic Quantum Error Correction in the Presence of Distinguishability

Shubham P. Jain, Eric R. Hudson, Wesley C. Campbell et al. (+2 more)

Quantum error-correcting codes require that the environment cannot distinguish relevant error processes acting on different codewords. In practice, residual Zeeman, Stark, or anharmonic interactions shift the underlying states, making errors distinguishable and weakening the conditions for perfect recovery. We study this effect with an analytically solvable four-level model and a canonical spontaneous-emission noise model for atoms and molecules. Distinguishability raises the effective Kraus rank of the noise channel, lowers the recovered fidelity, and sets an operational checking time that shortens as the frequency separation between the emitted photons increases. We find that rapid syndrome checking suppresses the buildup of distinguishability and quantify how fast one needs to check to restore near-perfect recovery. Error distinguishability is thus a practical limitation of monolithic error correction, and our analysis identifies fast syndrome extraction as a route to mitigating it.
Read the full abstract

All authors Shubham P. Jain, Eric R. Hudson, Wesley C. Campbell, Alexey V. Gorshkov, Victor V. Albert

Computation and Language

Classifying Interpretive Canons at the Sentence Level: A Benchmark from the German Federal Constitutional Court

Felix Ringe

Judicial reasoning remains challenging for large language models (LLMs) to analyze. This paper contributes a sentence-level benchmark for evaluating the ability of LLMs to classify interpretive canons as articulated by Larenz in the tradition of Savigny. Our contributions are threefold. First, we operationalize this conception of interpretation as classification criteria. Second, we provide a dataset of decisions of the German Federal Constitutional Court annotated at the sentence level. Third, we report baseline evaluations of four LLMs from three model families under expert hand-written prompts, compared against prompts optimized with Genetic-Pareto (GEPA). Mean F1 over the seven binary subtasks clusters between 70.4 and 79.2 across models, with grammatical interpretation usually the easiest canon to identify and systematic interpretation usually the hardest; under the tested configuration, GEPA-optimized prompts do not systematically outperform the hand-written ones, suggesting that the expert prompts provide a meaningful baseline.
Read the full abstract

High Energy Physics (Phenomenology)

Improved $B\to K^*_0\left(1430\right)$ transition form factors at next-to-leading order in light cone sum rules

Arslan Sikandar and M. Jamil Aslam

We present an improved determination of the $B\to K_0^*(1430)$ transition form factors using light-cone sum rules (LCSR) with an expansion near the light cone. We include the one-loop radiative corrections to the twist-2 and asymptotic twist-3 distribution amplitudes of the scalar meson and perform the calculation consistently in the $\overline{\mathrm{MS}}$ scheme. It allowed us to quantify the impact of radiative corrections on the form factors for each twist. A daughter sum rule is used from the $B\to K_0^*(1430)$ LCSR to ascertain Borel mass and threshold parameter simultaneously. We find that the dominant theoretical uncertainties arise from the $K_0^*(1430)$ distribution amplitudes and its decay constant. Our results are presented in a form that facilitates the incorporation of future improvements in these nonperturbative inputs, including those from lattice QCD and experimental measurements, and can be readily extended to other scalar mesons.
Read the full abstract

Soft Condensed Matter

Viscous Dissipation Governs Bubble Morphology and Failure in Soft Matter

Sushma SP, Anshul Shrivastava, Imnatoshi Jamir et al. (+2 more)

Cavitation, the growth of bubbles in fluids and soft matter, plays a central role in tissue damage, ultrasound therapies, and material failure, yet the influence of viscoelastic dissipation on bubble dynamics remains unclear. Here, we investigate cavitation in polyacrylamide hydrogels with identical elastic moduli but different viscous dissipation. We observe a striking symmetry breaking transition: elastic gels rapidly develop ellipsoidal cavities before rupture, whereas viscoelastic gels sustain large, nearly spherical bubbles. A modified Rayleigh Plesset framework shows that viscous stresses suppress shape instabilities and delay symmetry breaking, whereas elastic stresses accelerate cavity deformation. Finite element simulations independently validate these findings. The Deborah number (De) further captures the competition between material relaxation and cavity growth, with higher De associated with predominantly elastic growth and lower De with greater relaxation. Together, our results establish viscous dissipation as a key stabilizing mechanism that governs bubble morphology and failure in soft materials. These results have direct implications for understanding cavitation-mediated damage in biological tissues, optimizing therapeutic ultrasound, and advancing mechanical characterization of hydrogels for engineering and food processing applications.
Read the full abstract

All authors Sushma SP, Anshul Shrivastava, Imnatoshi Jamir, Aritra Chatterjee, Namrata Gundiah

Computation and Language

Recognized but Not Produced: A Generation Benchmark for Culturally Specific Kinship Terms

Sahil Pardasani and Madhusudan Singh

Current literature evaluates large language models (LLMs) on multilingual kinship understanding using multiple choice benchmarks, treating it as a recognition problem. We instead prompt five open weight LLMs to generate kinship terms in three non Western languages (Hindi, Tamil, and Korean) across two communicative tasks and pair this with a matched option-supported selection baseline. On identical relation language cells, GPT OSS120B selects the correct term in 90.67% of 75 valid cells but produces an accepted term in 36.00% of the corresponding attempts; Llama 3.370B shows the same pattern (77.92% versus 24.24%). Since the four-option condition displays the candidate terms and does not require script production, the difference is interpreted as an evaluation format gap rather than direct proof that lexical knowledge is intact. On explicitly specified L3 prompts, accuracy varies sharply, from GLM-5.1 at 72.29% to Llama-3.370B at 24.24%. The paternal-lineage advantage is language specific; it is large in Hindi but weak or reversed in Korean, while Tamil shared-term pairs provide a control for measurement variation. These results show that culturally specific kinship generation remains difficult even when the relationship is explicitly stated and motivate generation-based evaluation alongside multiple-choice testing.
Read the full abstract

Quantum Physics

Measurement-Only Dynamical Phase Transitions in Spin-1 Chains

Kemal Aziz, Haining Pan and J. H. Pixley

We study measurement-only dynamics in an interacting spin-1 chain, using forced, normalized local projections rather than Born-rule-sampled measurement outcomes. Starting from the nearest-neighbor (NN) Affleck-Kennedy-Lieb-Tasaki (AKLT) Haldane symmetry-protected topological state, we show that competing local projectors drive this order toward three distinct dynamical phases: a featureless topologically trivial product state, an explicitly dimerized valence-bond-solid state, and a topologically trivial next-nearest-neighbor (NNN) AKLT state consisting of two interwoven Haldane chains. Using matrix product state and exact statevector simulations, we identify continuous transitions among these phases and characterize them using nonlocal string order parameters, local dimer order, and entanglement scaling. For the NN-NNN AKLT model, finite-size signatures are consistent with an intervening extended critical regime rather than a single direct transition. These results extend studies of monitored topology to a bosonic SPT and show that local projections can control both topological and valence-bond-solid order.
Read the full abstract

Neural and Evolutionary Computing

The Computational Value of Sensory-Aligned Receptive Fields Depends on Neuronal Expressivity

Agnese Adorante, Aaron Spieler and Anna Levina

Biological sensory neurons have selective receptive fields organized along meaningful stimulus coordinates, such as frequency, motion direction, or retinotopic position. Such structure may arise from efficient coding and biological constraints on activity, connectivity, and wiring, as computational studies of simple neurons have shown across modalities. This raises a question: do structured receptive fields confer a computational advantage beyond resource efficiency itself, and does this advantage persist when individual neurons are highly expressive? We address this question in recurrent networks of Expressive Leaky Memory neurons, where we can independently vary neuronal complexity and the organization of feed-forward receptive fields. Across auditory and event-based visual classification tasks, receptive fields aligned with a task-relevant sensory coordinate improve test accuracy relative to budget-matched random receptive fields. This advantage disappears when sensory coordinates are scrambled, or when receptive fields follow task-irrelevant coordinates, showing that the benefit comes from alignment with task geometry rather than restricted connectivity alone. Increasing neuronal complexity reduces the performance advantage of structured receptive fields. Finally, generic synaptic sparsity regularization induces input selectivity and partially recovers performance, but remains substantially below explicitly structured receptive fields, suggesting that sparsity alone is insufficient to recover the full computational benefit of task-aligned receptive fields. Together, our results show that appropriate receptive fields can serve as a computational prior beyond sparsity itself, and that their value depends on the computational expressivity of individual neurons.
Read the full abstract

Medical Physics

Towards Modeling the Hemodynamic Impact of Mitral and Aortic Valve Repair in Patients with Left Ventricular Assist Devices

Mia Bonini, Michael Ferguson, Marc Hirshvogel et al. (+4 more)

Valve dysfunction is a major threat to long-term success in left ventricle assist device (LVAD) therapy, with direct implications for right heart performance. In this study, we apply a patient-specific, image-based computational modeling framework to evaluate the hemodynamic impact of simulated mitral and aortic valve repair in five LVAD-supported patients. Each patient was modeled under four conditions: (patient-specific LVAD-supported state), simulated mitral valve (MV) repair, simulated aortic valve (AV) repair, and simulated combined MV&AV repair. Because validation data for valve repair were unavailable, the simulated repair scenarios are exploratory in silico interventions based on clinically validated patient-specific models. The models integrate dynamic CT imaging, echocardiography, catheterization data, and device-specific LVAD parameters into a coupled 3D-0D simulation pipeline. Valve dynamics are governed by transvalvular pressure and flow, allowing physiological modeling of regurgitant lesions and surgical repair. The right ventricular (RV) was assessed using a combination of model-derived metrics, including right ventricular ejection fraction (RVEF), pulmonary artery pulsatility index (PAPi), and RV-PA coupling. In addition, we performed blood residence time (RT) analysis to evaluate blood stasis within the left heart and aorta. The simulations suggest that valve repair improved cardiac output, reduced pulmonary congestion, and enhanced right ventricular loading conditions. Notably, mitral valve repair restored aortic valve opening during the cardiac cycle, which improved sinus washout and reduced blood residence time within the aortic root-factors associated with lower thrombotic risk. Overall, these findings suggest a potential role for valve repair in LVAD-supported hearts, though larger, validated patient cohorts are needed to confirm these results.
Read the full abstract

All authors Mia Bonini, Michael Ferguson, Marc Hirshvogel, Maximilian Balmus, Paul C. Tang, Francis Pagani, David Nordsletten

Instrumentation and Methods for Astrophysics

The Simons Observatory: Commissioning of the Mid-Frequency Small Aperture Telescopes

Kathleen Harrington, Remington Gerras, Nicholas Galitzki et al. (+88 more)

The Simons Observatory (SO) is a cosmic microwave background (CMB) survey experiment observing from an altitude of 5,200 m in the Atacama Desert in Chile. The SO Small Aperture Telescopes (SATs) will cover six spectral bands between 30 and 280 GHz primarily to search for evidence of primordial gravitational waves in the large angular scale $B$-modes of the CMB polarization. The first two Mid-Frequency (MF) SATs, operating in the 90 and 150 GHz frequency bands, began commissioning observations in late 2023, and we present the results of those observations here. Each MF SAT operates with approximately 9,400 biased transition-edge sensors (TES) at 1 mm of precipitable water vapor (PWV), representing a typical 80% yield. Planet observations confirm the telescope optics performance, calibrate the instrument response, and measure the end-to-end optical efficiency of the SATs. The realized instrument efficiencies are within the range expected based on instrument model predictions. In nominal observing conditions, the noise equivalent temperatures from both telescopes combined are $2.4~\mathrm{\mu K_{CMB}\sqrt{s}}$ at 90 GHz and $3.2~\mathrm{\mu K_{CMB}\sqrt{s}}$ at 150 GHz, substantially outperforming the baseline instantaneous sensitivity targets for the observatory.
Read the full abstract

All authors Kathleen Harrington, Remington Gerras, Nicholas Galitzki, Samuel Day-Weiss, Yoshinori Sueno, Thomas Alford, Michael J. Randall, Kyohei Yamada, Maximiliano Silva-Feaver, Kevin Crowley, Shunsuke Adachi, Alexandre E. Adler, Prakamya Agrawal, Simone Aiola, David Alonso, Kam Arnold, Susanna Azzoni, Andrew Bazarko, Sanah Bhimani, Simon Biquard, Bryce Bixler, Josh Borrow, Michael L. Brown, Erminia Calabrese, Yuji Chinone, Steve K. Choi, Nadia Dachlythra, Mark Devlin, Simon Dicker, Peter N. Dow, Shannon M. Duff, Jo Dunkley, Rolando Dunner, Daniel Dutcher, Hamza El Bouhargani, Josquin Errard, Allen Foster, Ken Ganga, John Groh, Yilun Guan, Jon E. Gudmundsson, Saianeesh Haridas, Masaya Hasegawa, Matthew Hasselfield, Shawn W. Henderson, Carlos Hervías-Caimapo, Adam D. Hincks, Renée Hložek, Johannes Hubmayr, Bradley R. Johnson, Yutaro Kasai, Brian Keating, Ben Keller, Theodore Kisner, Brian J. Koopman, Akito Kusaka, Adrian Lee, JB Lloyd, Anto Lonappan, Marius Lungu, Aashrita Mangu, Michael McCrackan, Jeff McMahon, Jenna Moore, Thomas W. Morris, Hironobu Nakata, Federico Nati, Laura Newburgh, David V. Nguyen, John Orlowski-Scherer, Lyman Page, Ioannis Paraskevakos, Tristan Pinsonneault-Marotte, Erik Rosenberg, Yuki Sakurai, Goureesankar Sathyanathan, Thomas P. Satterthwaite, Yudai Seino, Elle Shaw, Sara M. Simon, Suzanne T. Staggs, Junna Sugiyama, Junya Suzuki, Satoru Takakura, Osamu Tajima, Alex Thomas, Daniel B. Thomas, Tran Tsan, Eve Vavagiakis, Yuhan Wang, John Wilson

Audio and Speech Processing

A Temporal-Envelope Frontend with Learnable Per-Channel Energy Normalization for Whisper-Based Children's ASR

Edem Ahadzi, Ruchi Pandey and Tomi H. Kinnunen

Temporal envelopes carry cues critical to speech intelligibility, yet ASR frontends based on log-mel spectrograms do not explicitly model continuous sub-band envelope structure. This limitation is particularly acute for children's speech, where high acoustic variability demands robust feature representations. We propose a modular time-domain frontend that decomposes speech into sub-band envelopes using mel-spaced windowed-sinc filters and the Hilbert transform, with learnable per-channel energy normalization (PCEN) jointly optimized with the Whisper model. On the MyST children's speech corpus, systematic ablations identify full-band windowed-sinc filters, Hilbert envelopes, a 25 Hz smoothing cutoff, and learnable PCEN as the best configuration. Under the same Whisper-small fine-tuning setup, the frontend reduces WER from 13.16% to 11.08%, a 15.8% relative reduction over the log-mel baseline, and outperforms the evaluated Kid-Whisper checkpoint on the same cleaned test split. These results show that temporal-envelope representations and learnable frontend normalization are effective complements to backend adaptation for children's ASR.
Read the full abstract