Peer Revenue Returns subject to peer review

Papers

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

All of arXiv

Monday 31 August

Computer Vision

Where Should Experience Live? Hierarchical Hebbian Memory for Continual Vision Transformers

Vision Transformers provide strong visual representations but typically rely on slowly updated parameters, limiting their ability to organize newly acquired information across different memory timescales. This work proposes Hierarchical Hebbian Memory, a three-level memory architecture composed of rapid Working Memory, persistent Routed Episodic Memory, and slower Semantic Memory. A learned controller regulates memory contribution, read and write routing, plasticity, retention, and consolidation. A causal read-before-write lifecycle ensures that the current outcome cannot influence the prediction it supervises. The architecture is evaluated on Omniglot 5-way 1-shot recognition and CORe50 continual object recognition. With Swin-Tiny, the hierarchical model reaches 97.39% accuracy on Omniglot and 95.37% final accuracy on CORe50 when combined with experience replay. Learned multi-bank retrieval reaches 47.50% delayed-association accuracy, compared with 24.17% for a single persistent bank and 25.00% without memory. After intervening distractors, Episodic Memory retains approximately 0.96 cosine similarity with stored associations, while Working Memory falls to approximately 0.05. These results show that Hebbian association and learned memory routing can jointly organize online visual experience across rapid, persistent, and consolidated memory timescales within Vision Transformers.
Read the full abstract

Instrumentation and Methods for Astrophysics

The Canadian Galactic Emission Mapper (CGEM): A Cosmic Microwave Foreground Experiment

The Canadian Galactic Emission Mapper (CGEM) is a 4-meter single-dish radio telescope located at the Dominion Radio Astrophysical Observatory (DRAO) in Penticton, Canada. CGEM is designed to map polarized Galactic synchrotron emission across the entire northern sky at 8-10 GHz with 1 MHz frequency resolution and 0.5 degree angular resolution. Its goal is to obtain high-fidelity, low-noise maps of polarized Galactic synchrotron radiation, at frequencies where synchrotron dominates, in order to provide a reliable template for CMB B-mode foreground cleaning. We will show on sky performance of CGEM's single-pixel azimuthally-symmetric telescope, its coherent, dual-polarized radiometer, and show maps made from commissioning data from the first few months of operation.
Read the full abstract

All authors Parham Zarei, Pedro Villalba-González, Mandana Amiri, Charles L. Bennett, Guinevere Berg, Mark Halpern, Gary Hinshaw, Gordon Lacy, Joshua MacEachern, Thomas J. Rennie, Shuyu van Kerkwijk, Bruce Veidt, Janet Weiland, Don Wiebe, Edward J. Wollack

Artificial Intelligence

Vision Is Not Overhead: One-Pass Block Drafting for Lossless Speculative Decoding in Vision-Language Models

Speculative decoding accelerates generation without changing its output, but on vision-language models (VLMs) a self-reinforcing cycle holds it back. Because an autoregressive drafter pays a sequential pass for each drafted token, it must stay small and can ill afford to attend to the image at each pass. Prior work therefore compresses or hides the image, leaving the drafter weakest on the text the image determines. We present GLANCE, a one-pass block drafter that breaks this cycle on an unmodified VLM target. Its block-diffusion head drafts a whole block in one forward pass over the target's already fused vision-language states, reading the multimodal context once, however deep the draft. The target verifies a wide candidate tree in one pass and commits exactly its greedy output. In one production engine at a fixed round budget, GLANCE decodes up to 3.05 times faster than autoregressive decoding and outpaces the production EAGLE3-VL head on average and by about 11% on grounded tasks. An entropy law explains when drafting pays, predicting the longest accepted blocks on grounded tasks, where the target's next-token entropy is lowest. Our code is available at https://github.com/js-lee-AI/GLANCE.
Read the full abstract

All authors Jungseob Lee, Seongtae Hong, Dongyub Jude Lee, Chanjun Park, Jaehyung Seo, Sugyeong Eo, Heuiseok Lim

High Energy Physics (Theory)

Topologically induced deformations of differential forms

Motivated by deformations of spin structures and their geometric implications, we study a deformed exterior derivative, which is a third-order nilpotent operator. This higher-order differential structure induces an anholonomic rescaling of the local frame and leads to a bifurcation of the associated cohomological structure, with two distinct notions of closedness and exactness. We prove that the Poincaré Lemma fails in both sectors, so that closed deformed forms need not be exact even locally. We then apply the resulting framework to a four-dimensional effective field theory and show that a constant saturation of the deformation parameter can generate a hierarchy between the electroweak and fundamental Planck scales through a purely geometric mechanism, without introducing extra spatial dimensions.
Read the full abstract

Econometrics

Manipulation Testing in Boundary Discontinuity Designs

We propose the first manipulation test designed for boundary discontinuity designs (BDDs) with general boundary shapes. A BDD is a multidimensional extension of the regression discontinuity design (RDD) in which treatment assignment is determined by whether the multidimensional running variable crosses a lower-dimensional boundary set. The test avoids multivariate density estimation and builds on the observation that, in the absence of manipulation, observations near the boundary should be approximately evenly split between treatment and control within arbitrary groups defined by their projections onto the boundary. We test this implication using a collection of binomial balance tests on observations near the boundary, with groups formed by k-means clustering. We establish the asymptotic validity of the test under suitable regularity conditions. We also evaluate finite-sample performance through Monte Carlo simulations and illustrate the test in three empirical applications.
Read the full abstract

Nuclear Theory

Symmetry-Reduced Variational Quantum Simulation of the $U(5)\rightarrow SU(3)$ Quantum Phase Transition in the Interacting Boson Model

The spherical-to-deformed quantum phase transition of the Interacting Boson Model (IBM) is investigated using the variational quantum eigensolver (VQE). The $U(5)$–$SU(3)$ transitional Hamiltonian is studied with $\chi=-\sqrt{7}/2$. We develop a symmetry-preserving, minimum-qubit VQE framework for collective nuclear models, achieving a substantial reduction in qubit requirements without compromising the finite-size quantum-phase-transition physics. The transition is characterized through the normalized $d$-boson occupation and ground-state energy derivatives. Finite-size results are found to approach the analytic critical point $\xi_c=8/17\simeq0.470588$, with an independent order-parameter extrapolation yielding $\xi_\infty=0.46986(61)$. The VQE reproduces ground-state energies and structural observables to numerical precision. These results demonstrate the potential of symmetry-reduced VQE for efficient quantum simulations of collective nuclear dynamics and quantum phase transitions.
Read the full abstract

Quantum Physics

Spatial correlations of photons interacting via transverse Rydberg blockade

We develop a theory of the transverse spatial dynamics of two interacting photons in a Rydberg nonlinear medium. Extending the well-studied one-dimensional case, we explore both longitudinal and transverse correlations of photons propagating as Rydberg polaritons. We identify distinct behaviors and scaling laws for these correlations, arising from fundamentally different mechanisms in the two directions: diffraction for transverse correlations and diffusion for longitudinal correlations. We develop a model incorporating a Gaussian optical beam and an inhomogeneous atomic density distribution, from which we derive quantitative predictions for correlation functions in both directions. We further show that the propagation equations can be reduced to a single Schrödinger-like equation, allowing approximate solutions that are supported by full numerical results. Our findings indicate that the transverse correlation length is determined primarily by the blockade radius, whereas the longitudinal correlation length is limited by bandwidth. These results establish transverse Rydberg blockade as a distinct and measurable correlation mechanism and show that spatial photon correlations provide a direct means of measuring the blockade radius.
Read the full abstract

All authors Bankim Chandra Das, Daniil Svirskiy, Matthias Metternich, Wolfgang Alt, Sebastian Hofferberth, Ofer Firstenberg

Quantum Algebra

Algebra objects in direct limit completions of compact Lie group duals and the classification of $c=1$ vertex operator algebras

Let $G$ be a complex reductive affine algebraic group with its symmetric tensor category $\mathcal{C}_G$ of finite-dimensional rational representations. We prove that every simple commutative algebra object with at most countable dimension in the direct limit completion $\operatorname{Ind}(\mathcal{C}_G)$ of $\mathcal{C}_G$ is isomorphic to the algebra $\mathcal{O}(G/H)$ of regular functions on the homogeneous space $G/H$, for some reductive algebraic subgroup $H$ of $G$. Then we apply this result to the theory of vertex operator algebra extensions. In particular, assuming the strong rationality of the $ \operatorname{A}_5$-orbifold of the vertex operator algebra $V_{\mathcal{L}_2}$ associated with the rank-one root lattice $\mathcal{L}_2:= \sqrt{2}\mathbb{Z}$, we classify all the not necessarily rational simple CFT type preunitary vertex operator algebra extensions of the simple unitary Virasoro vertex operator algebra $L(1,0)$ with central charge $c=1$ satisfying a certain spectrum condition. This result is the vertex operator algebra analogue of a conformal net result by Feng Xu. Every strongly rational preunitary vertex operator algebra extension of $L(1,0)$ satisfies the above spectrum condition because of the congruence subgroup modularity property of its characters. As a consequence, we get a complete classification result for strongly rational $c=1$ vertex operator algebras, up to the strong rationality of $V_{\mathcal{L}_2}^{\operatorname{A}_5}$.
Read the full abstract

Machine Learning

Do LLMs Know Your Neighborhood? Auditing LLM Priors for Neighborhood-Level Mobility Prediction and Structural Alignment

Human mobility is central to urban planning, transportation, public health, and emergency response, yet fine-grained trajectory data are often proprietary, restricted, and privacy-sensitive. Large language models (LLMs) offer a potential alternative by generating plausible mobility traces and predicting individual movement, but their ability to infer aggregate neighborhood-level mobility remains unclear. We evaluate zero-shot LLMs on Census Block Group-level mobility prediction across four U.S. metropolitan areas using anonymized Cuebiq data to construct point-level, trajectory-level, and temporal mobility outcomes, paired with sociodemographic and built-environment predictors. We compare LLM predictions with supervised baselines and introduce a directional alignment analysis to test whether LLM-implied predictor effects agree with empirical OLS and Jonckheere-Terpstra trends. Supervised models achieve 0.580 average accuracy, compared with 0.435 for the best LLM, with spatial extent outcomes showing the strongest predictability but also the largest LLM-baseline gaps. Directional analysis shows that LLMs often rely on coarse, stable predictor-level priors that remain similar across outcomes and cities, including asymmetric treatment of protected-group predictors. Overall, LLMs can partially recover aggregate mobility patterns from urban context, but their predictions should not be treated as structurally grounded without auditing empirical alignment and potential bias.
Read the full abstract

All authors Saad Mohammad Abrar, Eesha Kurella, Arnav Dadarya, Naman Awasthi, Kazi Tasnim Zinat, Vanessa Frias-Martinez

Computation and Language

From Tool Use to Technological Agency: LoopCAT as a Local-First, Open-Source Tool for Translation Technology Education

Translation students need to learn both how to use translation technologies and how to judge the choices those technologies make available. This article presents LoopCAT, an Apache-2.0-licensed, local-first computer-assisted translation environment co-created with OpenAI Codex using GPT-5.5 and GPT-5.6, and proposes a framework connecting workflow competence, evaluative judgement, and technological agency. The account draws on repository history, implementation inspection, and the verification records of an identified development build. LoopCAT combines local project storage, translation memories, terminology, quality assurance, document exchange, and optional connections to local or hosted AI services. Its English, Catalan, and Turkish interface catalogs also make the application itself available as teaching material: students can translate English UI strings into another language, review the existing automatically generated target drafts, import their revisions, and test the interface. We organize these opportunities around four forms of participation: operating a workflow, evaluating outputs, inspecting and configuring mechanisms, and making or defending a bounded intervention. A six-session sequence, a UI-localization assignment, a placeholder example, and an assessment rubric specify how teachers could use the framework. The paper separates implemented capabilities from proposed educational benefits; it reports no new student-learning outcomes. It distinguishes the latest package checks from earlier regression evidence and sets out a protocol for classroom evaluation. LoopCAT provides an inspectable setting for teaching how translation decisions interact with data, interfaces, and software rules. Whether these activities improve judgement, transfer, or participation remains an empirical question.
Read the full abstract

Quantum Physics

Generation of multicomponent Schrödinger cat states in schemes with measurement of Gaussian states

In this work, we propose a scheme for generating superpositions of Fock states whose numbers differ by four or eight. The proposed protocol is based on photon-number-resolving measurements on multimode Gaussian states. We compare of the generated states with multicomponent Schrödinger cat states. We evaluate the fidelity and determine how it scales with the number of detected particles. Furthermore, we derive the optimal configuration for the generation such states.
Read the full abstract

Artificial Intelligence

SlideBank: A Persistent Hierarchical Evidence Bank for Consistent Whole-Slide Reasoning

Whole-slide images (WSIs) are challenging for vision-language reasoning because diagnostically relevant morphology is sparse, heterogeneous, and distributed across gigapixel-scale images and multiple spatial resolutions. Existing WSI models and pathology agents can aggregate slide features or actively acquire evidence, but the information retained after exploration is often difficult to access semantically while preserving its connection to the original visual evidence. We introduce SlideBank, a training-free framework that represents each WSI as a persistent, concept-indexed, and spatially grounded evidence bank. SlideBank performs question-independent coarse-to-fine exploration to identify informative regions and multi-scale views, converts them into explicit morphological observations, and grounds pathology signals to their supporting patches and WSI coordinates. At inference time, questions are routed to relevant signals and evidence scales, and the linked global, regional, and patch evidence is integrated through confidence-based cross-level consensus. Experiments on WSI-VQA and SlideBench-BCNB show that with Patho-R1, SlideBank reaches 52.77% on WSI-VQA and with Quilt-LLaVA, it reaches 50.92% average accuracy on SlideBench-BCNB, while structured signal-guided retrieval consistently outperforms random evidence sampling. Reusing the same bank across repeated queries further achieves over 99% rephrasing consistency and substantially reduces amortized inference cost through persistent evidence reuse.
Read the full abstract

All authors Beidi Zhao, Gexin Huang, Ciro Zhang, Anqi Li, Yusheng Tan, Chen Zhou, Gang Wang, Zu-hua Gao, Xiaoxiao Li

Materials Science

Multi-image Overlap Stitching and Automatic Image Construction for Coherent X-ray Imaging

Direct-space and real-time coherent X-ray imaging (direct-CXI) enables visualization of magnetic-domain structures and dynamics over length scales exceeding the field of view of a single image. However, large-area measurements typically require raster scanning, producing hundreds of partially overlapping images that must be accurately aligned and combined before quantitative analysis can be performed. Here, we present Multi-image Overlap Stitching and Automatic Image Construction for coherent X-ray imaging (MOSAICX), an automated stitching workflow for large-area direct-space coherent X-ray imaging. The workflow consists of image centering, masking, trimming, binarization, hierarchical stitching, and post-processing. To demonstrate the method, we apply it to a dataset comprising 434 direct-CXI images of the antiferromagnetic topological insulator MnBi$_2$Te$_4$. The images are first combined into column reconstructions and subsequently stitched into a single large-area composite image. The resulting reconstruction reveals the complete magnetic-domain and domain-wall landscape over the scanned region while suppressing imaging artifacts and detector defects. The presented workflow provides an efficient approach for processing large direct-CXI datasets and enables visualization and analysis of magnetic-domain structures beyond the field of view of individual measurements.
Read the full abstract

All authors Starr Boney, Umeshika Dissanayaka, Lillian Rutowski, Aaron George, Min Gyu Kim

Cryptography and Security

OreProof: Verifiable Provenance with Limited Disclosure for Critical-Minerals Supply Chains Using Zero-Knowledge Proofs

Critical-minerals supply chains face a structural tension: regulators and buyers demand verifiable provenance, yet upstream actors are hesitant to disclose supplier identities, assay grades/yields, and prices that verification appears to require. We report a design science account of OreProof, a prototypical traceability platform addressing this verifiability-disclosure trade-off. Instantiated for gold, OreProof combines a hybrid on-chain/off-chain data model, Groth16 zero-knowledge proofs for selective disclosure, a Merkle-batched anchoring pipeline, and UNTP-aligned verifiable credentials on a public zkEVM testnet. Against a transparent baseline, directly inferable confidential attributes fell from three of four categories to none under a defined attacker model, while batched anchoring substantially improved throughput. Our contributions are the artifact prototype as well as four nascent design principles: prove over committed data rather than exposing it; credential only verifiable origin and flag unknown inputs for blended commodities; emit standards-aligned credentials from the outset; and partition disclosure by supply-chain role.
Read the full abstract

All authors Oleksandr Hrabar, Hossein Arshadi Soufiani, Henry M. Kim, Chien-Chih Chen, Ali Vazirizadeh, Hjalmar Turesson

General Relativity and Quantum Cosmology

A Fast and Scalable Transformer Pipeline for Binary Black Hole Detection

With the projected increase in the detection rate of compact-binary coalescences in the coming decade, there is critical need to develop fast, robust, and scalable alternatives to matched filtering for gravitational-wave searches. Transformer models have revolutionized natural language and audio processing but their application to gravitational-wave astronomy is still largely unexplored. In this work, we introduce \castor, a transformer-based coincident search pipeline for detecting binary black hole gravitational-wave signals from Advanced LIGO detectors. One of the major features of our model is that it allows the false-alarm rate to be estimated via time slides cheaply without requiring repeated evaluations of the neural network. We evaluate \castor\ on datasets from the Machine-Learning Gravitational-Wave Search Challenge (MLGWSC-1) and on approximately five months of real O3b observing strain. When tested on benchmark datasets, \castor\ ranks among the most sensitive machine-learning pipelines and successfully recovers the majority of confident events from the GWTC-3 catalog that lie within its training range. We also benchmark \castor\ against another transformer architecture, GW-Whisper, a domain-adaptation of OpenAI's audio foundation model. We find that \castor\ substantially outperforms the repurposed audio model in sensitivity and also reduces the computational cost of background estimation by a factor of 20. Our results demonstrate a highly practical, scalable approach for deep-learning gravitational wave searches and empirical background estimation for future observing runs.
Read the full abstract

All authors Chayan Chatterjee, Abigail Petulante, Haowei Fu, Yang Hu, Roy Lau, Karan Jani

Superconductivity

Impurity-dependent quantum geometry in 2D flat band superconductors

In flat band superconductors, quantum geometry, rather than kinetic energy, governs pairing and transport. However, how this geometry manifests in the response to a local, individually addressable impurity remains largely unexplored. Here, we study the Yu-Shiba-Rusinov (YSR) bound state induced by a magnetic adatom that hybridizes with every orbital of a two-dimensional flat band superconductor, and characterize its spatial profile using two quantities: the quadratic spread $Q_S$ and the localization length $\xi$. We show that $Q_S$ is not set by the quantum metric alone, as geometric corrections arising from inter-orbital interference and, more importantly, from the anisotropy of the adatom-lattice hybridization contribute on equal footing, making $Q_S$ strongly impurity dependent and tunable. On the other hand, $\xi$ emerges from the projection onto the flat band alone and is a universal property set by the overlap of compact localized states. All impurity dependence is instead isolated in the prefactor of the wavefunction whose magnitude correlates directly with $Q_S$. We, therefore, report a clean separation between universal and impurity-tunable physics from the induced YSR bound state in these systems, and confirm our results numerically via exact diagonalization.
Read the full abstract

Optimization and Control

Blaschke–Santaló diagram for the volume, the diameter, and the Cheeger constant

We study the Blaschke–Santaló diagram associated with the volume, the diameter, and the Cheeger constant. In the class of bounded open subsets of $\mathbb{R}^m$, $m\ge 2$, we give a complete description of the diagram. In the class of convex bodies of $\mathbb R^m$, we prove that the diagram is closed and simply connected as it is given by the region between two continuous functions for which qualitative properties such as monotonicity, local behavior, and growth estimates are studied.
Read the full abstract

Classical Analysis and ODEs

Critical points of point charge potentials along lines

We prove Conjecture 1.9 of Gabrielov-Novikov-Shapiro. More precisely, for every $p>0$, every nonconstant restriction to a line of a potential generated by $n$ point charges has at most $2n-1$ critical points. The bound is sharp, even for positive charges. The proof uses a duality with an auxiliary planar potential generated by collinear charges and Morse theory.
Read the full abstract

Computation and Language

Two tests of phase-structure features for transition prediction

Following arXiv:2607.25507, this report examines whether phase-derived features improve endpoint prediction over a combined baseline in two settings: a sealed contradiction comparison and a retrospective analysis of answer changes across matched pressure prompts. Study 1 froze a contradiction-category pipeline before sealed scoring. On 1,136 eligible primary cases, adding PC-2 produced a paired AUROC difference of +0.00087. The 99% bias-corrected accelerated interval included zero, and the prespecified +0.05 threshold was not met. A replication role with 1,063 cases showed a same-direction increment of +0.00019. The replication-direction condition passed, but both primary conditions failed, so advancement failed. Study 2 developed fifteen treatments on blocks b0-b4 using 1,415 eligible answer-change comparisons and twenty-repeat, five-fold grouped cross-validation. An execution on 9 September 2026 recomputed development statistics from saved prediction units and applied a three-condition gate; no treatment advanced. Layer 25 had the only positive mean-repeat PC-2 increment, approximately +0.00027, with a favorable sign in three of five seed blocks. These statistics measure development consistency, not whole-seed-block holdout performance. Agreement with earlier selection records does not establish an earlier notebook execution or that the rule was fixed before inspecting development results. The planned b5 selection and b6 evaluation were not completed through this gate. Neither study demonstrated the incremental benefit required by its applied advancement rule. The findings limit support for the evaluated feature constructions; they do not test the rotary score identity, its local pre-softmax bound, or the effectiveness of execution-boundary governance.
Read the full abstract

Robotics

Scene Graph-based Driving Scenario Extraction for Automotive Egocentric Datasets

Extracting scenarios from unlabelled real-world sensor data streams is a critical but challenging task in the development process of automated driving systems (ADS). Automatically sifting through large datasets to spatially and temporally locate critical scenarios can enable scenario-based coverage analysis of ADS datasets. In this paper, we present a method for extracting scenarios from egocentric datasets using scene graphs and Linear Temporal Logic (LTL). We first process egocentric sensor data and HD maps to generate a sequence of scene graphs representing a driving scenario. Next, we use LTL to formally specify driving scenarios of interest, then extract all instances of the scenarios from the dataset using an off-the-shelf model checker, which evaluates the LTL formula against the sequence of scene graphs. Our approach can be used on both simulated and real world datasets. We evaluate the method on the training and validation datasets from Argoverse 2 consisting of 850 15-second real-world driving logs, and several videos of dashcam footage. We demonstrate the effectiveness of our approach for extracting and querying scenarios by evaluating against a rule-based benchmark based on track annotations and HD maps.
Read the full abstract

All authors Stefan Ramdhan, Kyanna Dagenais, Vera Pantelic, Victor Bandur, Mark Lawford

Computational Finance

Latent-Space No-Arbitrage Geometry of Generative Models for Implied Volatility Surfaces

Generative models for implied volatility surfaces must produce outputs that satisfy static no-arbitrage constraints. We study these constraints in latent space. For a fixed generator, we assign each latent code a scalar margin determined by the no-arbitrage conditions of the generated surface. The codes with nonnegative margin form the admissible latent set. We establish conditions under which strictly admissible codes remain admissible under small perturbations and the boundary of the admissible set is characterized by zero margin. For regular boundary components, we formulate a level-set equation whose local dynamics are directed toward the zero-margin set. The analysis treats the generator as a map from latent variables to surfaces and is therefore not restricted to a particular architecture. It applies to variational autoencoders, generative adversarial networks, and other generative models with a deterministic realization map. Numerical tests recover known boundaries in analytic examples. Experiments with a variational autoencoder trained on Heston surfaces show that similar reconstruction errors can correspond to different admissible regions and that the latent prior may be concentrated inside such a region. The computed boundary can also be used to modify latent codes that generate violating surfaces.
Read the full abstract

Soft Condensed Matter

Using microrheology to study dynamical heterogeneities

Dynamical heterogeneities are one of the hallmarks of supercooled liquids, and their properties and relevance have been studied with theory, simulations and experiments. In this work, we propose to monitor the dynamics of tracer particles (passive microrheology) to analyze the dynamical heterogeneities in a system of hard colloids close to the glass transition density, using Langevin dynamics simulations and mode coupling theory. Different observables, typical in the study of the dynamical heterogeneities are adapted to be calculated from the trajectory of a single tracer particle. The tracer dynamics shows a transition from a regime where it is most decoupled from the bath for small tracer size to a strong coupling regime for large tracers. Both theory and simulations show that the non-Gaussian parameter of the tracer is maximal for tracer sizes at the crossover between both regimes, and is highly dependent on the bath density. The dynamic susceptibility is also studied, but this parameter shows a minor dependence on both the tracer size or the bath density. Finally, the existence of regions with different mobility is also studied with microrheology. Although the tracer trajectory indeed shows stages with increased mobility, the estimated size of the regions decreases with the bath density, contrary to the results from cluster analysis in the bulk.
Read the full abstract

High Energy Physics (Theory)

Accretion, mergers, and metastability of fuzzy spheres in a three-matrix model

A three-matrix model that contains a plethora of fuzzy sphere solutions was defined twenty years ago. It has been shown that in a certain regime of the model, combinations of fuzzy spheres become classical solutions of the model. In the original paper and subsequent works, the stability of specific states has been discussed, showing that only a single largest fuzzy-sphere configuration is stable. Here, we focus on a complementary question: starting from some initial configuration, is the thermodynamically preferred state approached rapidly, or are there other long-lived, metastable configurations which may support interesting physics? We find that the answer in the large-matrix limit depends on the dispersion of the initial configuration. We identify the elementary transition as an accretion, in which one sphere absorbs a single unit from another, and show that it is driven by fluctuations of the sphere centers. By measuring these fluctuations in long simulations, we find quantitative agreement with the one-loop effective action, with no free parameters. For multi-sphere states, we find that the stiffness of an inner sphere grows logarithmically with the number of surrounding layers, so that concentric onion-like configurations – describing an emergent three-dimensional quantum space – become increasingly long-lived.
Read the full abstract

All authors M. Hrmo, S. Kováčik, K. Magdolenová, A. Manta, K. Nedeľková, P. Rusnák, H. C. Steinacker, J. Tekel

Physics and Society

Modeling of Mobility and Energy Policies in an Agent-Based Framework: Case Studies for Chicago Region in 2050

Metropolitan regions are simultaneously pursuing several interventions to improve mobility, accessibility, and energy efficiency, necessitating integrated tools to understand how these policies interact to affect travel behavior, energy use, and infrastructure needs. This paper evaluates the combined impacts of electrification, freight demand management, road pricing, parking reform, and transit expansion on the Chicago metropolitan transportation system in 2050, using a business-as-usual (BAU) scenario as the baseline. We employ POLARIS, a large-scale agent-based modeling framework calibrated to 2019 conditions, to simulate nine policy scenarios for the seven-county northeastern Illinois region. The framework co-simulates activity-based passenger demand, endogenous freight generation, multimodal traffic assignment, and transit operations, with charging infrastructure and freight operations optimized for each case. Our findings reveal that under the high electrification scenario, total fuel mass declines by 68% while total charging energy increases by approximately 4-8x from BAU, resulting in a peak power demand near 4 GW concentrated in the urban core. Furthermore, freight management policies reduce freight VMT by increasing trip frequency but shortening distances, smart road pricing most effectively reduces auto VMT, and transit expansion boosts ridership by 18% relative to BAU. By presenting the first integrated, agent-based scenario framework for Chicago that jointly evaluates these interventions, this study provides actionable insights for regional transportation planning, grid infrastructure investment, and emissions reduction, highlighting the value of targeted charger upgrades and coordinated policy bundles.
Read the full abstract

All authors Md Rakibul Alam, Omer Verbas, Taner Cokyasar, Hadi Bhidya, Jesse Altman, Nora Beck, Joshua Auld, Pedro Veiga de Camargo, Jamie Cook, Felipe de Souza, Gopindra Nair, Hyunseop Uhm, Jan Zill

Complex Variables

On a construction of hermitian metrics on holomorphic vector bundles

With any holomorphic vector bundle $E$ over a compact base one can associate a line bundle, denoted $O_{PE}(1)$. According to a conjecture of Griffiths, if $O_{PE}(1)$ admits a positively curved hermitian metric $k$, then $E$ also admits a positively curved hermitian metric, $h$. In this paper we show that, while the conjecture may be correct, it is not possible to obtain $h$ out of $k$ by a fiberwise construction that is functorial.
Read the full abstract

Computation and Language

Latent Mechanisms of Language Control in Multilingual Language Models

Multilingual large language models can exhibit unintended code-switching – unnecessarily alternating between languages during generation. We present a comparative study of three methods that identify language-controlling latents in cross-layer transcoders: activation value-based selection (ValSel), activation frequency-based selection (FreqSel), and LLM-generated latent annotation-based selection (AnnSel). To evaluate the efficacy of these methods in identifying language-controlling latents, we introduce two multilingual benchmarks that exhibit code-switching for fine-grained analysis of language steering across seven languages. Through targeted intervention experiments on Gemma-2-2B and Qwen3-4B, we find that all three methods effectively manipulate generation language, with FreqSel achieving the strongest overall performance, while AnnSel offering interpretable latent selection through explicit language annotations. A knock-out analysis suggests the methods select non-overlapping but each-functional latent subsets, indicating redundancy rather than a single canonical language direction. Code and data can be found at https://github.com/rm-3284/Latent-Mechanism-Multilingual.
Read the full abstract

Analysis of PDEs

Quantitative Stability for Fractional Yamabe minimizers

We study quantitative stability for the fractional Yamabe functional introduced by González and Qing. Its minimizers correspond to conformal metrics with constant fractional curvature. We prove that the deficit of the fractional Yamabe functional controls, with a suitable power, the distance to the set of minimizers. Our approach is based on a Lyapunov–Schmidt reduction for the nonlocal functional, together with the extension characterization of the fractional conformal Laplacian, which provides the regularity theory needed for the local analysis. As a consequence, we also obtain an analogous quantitative stability estimate for the associated extension functional.
Read the full abstract

Digital Libraries

Do Large Language Models Favour Any Research Topics?

Large Language Models (LLMs) can estimate the quality of published journal articles, potentially supporting human assessment when evaluations are needed. Whilst there are reasons to believe that LLMs may have biases in this role, there is no statistically strong evidence yet. The current article addresses this gap with an exploration of the types of articles that attract high or low LLM scores in 73,489 articles from 15 health and life sciences journals. Based on comparing the words in the titles and abstracts of higher and lower scoring articles for two LLMs in various ways, the results suggest that topics favoured by GPT-OSS-120B include viruses, genes and cells and its disfavoured topics include surveys, patients and students. It is not clear whether these patterns reflect underlying quality differences or AI biases, however. The same method found systematic differences between the topics favoured by GPT-OSS-120B and Gemma 3 27B, such as Gemma 3 27B giving relatively higher scores for machine learning research, proving that at least one of the two LLMs has AI bias. Finally, comparing the scores for full-text articles compared to scores for titles and abstracts also finds differences for both LLMs, showing that they both can exhibit AI bias for at least one of these two input types, and probably both. Overall, the results show that it is important to consider LLM biases when deciding whether to use them for research evaluation tasks.
Read the full abstract

Statistical Mechanics

A Human-AI Theorem Connecting Spontaneous and Field-Induced Mechanisms of Collective Behavior in One Dimension

Can an artificial intelligence (AI) generate a scientific hypothesis outside a human collaborator's active hypothesis space (AHS), and can human-AI research be organized to make such breakthroughs more likely? We document such a case while proving a theorem that connects two basic organizing mechanisms of statistical physics: collective behavior arising in zero field from competing interactions and that induced or controlled by an external field. A zero-field $O(n)$-vector open chain with arbitrary inhomogeneous nearest- and next-nearest-neighbor interaction functions $U_i(S_i\cdot{S}_{i+1})$ and $V_i(S_i\cdot{S}_{i+2})$ is microscopically, via a temperature-independent mapping at the Hamiltonian level, equivalent to a simpler $O(n)$ open chain with nearest-neighbor interaction $V_i(\boldsymbol\sigma_i\cdot\boldsymbol\sigma_{i+1})$ and axial single-spin potential $U_i(\sigma_i^z)$ for every integer $n\ge1$ and every system size $L\ge1$. The homogeneous linear specialization maps the foundational frustrated $J_1$-$J_2$ model onto the canonical $J$-$h$ field model—with $n=1,2,3$ being the Ising, XY, and Heisenberg classical spin models, respectively; the theorem resolved a longstanding challenge for $n=3$ published in 1990. Its proof was done with an AI-synthesized recursive Householder moving frame and understood via a human-recognized hidden reciprocity. An analogous theorem holds when the continuous $O(n)$ spins are replaced by the $q$-state Potts spins, implying a closed-form exact solution of the $J_1$-$J_2$ standard Potts open chain for every $q\ge2$ and every $L\ge1$. The emergence of these theorems from a human-AI co-development framework suggests that sustained AI involvement throughout a systematic research program may incubate autonomous scientific breakthroughs and make aspects of the discovery process experimentally testable.
Read the full abstract

Quantum Physics

Which Otto Engine Is the Fastest?

A general thermal machine is characterized by distinct time or energy scales: temperature, coupling to the heat baths, internal free dynamics and interactions, and external driving. We address the question: what parameters determine the speed of a given engine, and how can they be used to reliably compare different types of engine? Specifically, we consider four realizations of the Otto engine, each operating with the same Otto efficiency, and aim to characterize and compare their power outputs. The main insight of this paper is that, irrespective of their very different dynamical implementations, the power of each engine can be factorized into a common characteristic work and an implementation-specific characteristic time. The characteristic work depends only on the internal frequencies of the machine and the temperatures, whereas the characteristic time depends on the coupling strength to the baths and on the implementation-specific driving. This observation allows us to compare the power of different engines through their characteristic times and thereby reduce the relevant parameter space to parameters that capture the dynamical aspects of engine performance. Despite different dynamical implementations, this framework reveals simple common features. In all cases, the maximum operating speed is limited by a common timescale given by the sum of the characteristic thermalization times of the hot and cold baths. Moreover, the four engines fall into two distinct asymptotic classes: implementations in which driving and thermalization occur simultaneously exhibit quadratic scaling, whereas those in which work extraction and thermalization alternate exhibit linear scaling. These results expose general dynamical features that determine the operational speed of quantum Otto engines.
Read the full abstract