Peer Revenue Now accepting positions

Papers

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

All of arXiv

Friday 6 November

Geometric Topology

Semifree actions and representation varieties

A group action on a topological space induces an action on the space's representation variety. We determine the fixed point sets of this latter action for semifree group actions on surfaces and circle bundles over surfaces. We also discuss implications for mapping tori of periodic maps.
Read the full abstract

All authors Patrick Dendis, David L. Duncan, Timothy Tarter, Lior Weber, Colin Willette

Earth and Planetary Astrophysics

Thermally unfavorable conditions of volatile existence in near-Earth asteroid 1998 KY26, the Hayabusa extended mission's target

The near-Earth asteroid 1998 KY26, the target of the Hayabusa2# extended mission, has been proposed as a dark comet that may host volatiles, particularly water, owing to an anomalous out-of-plane nongravitational acceleration that may result from outgassing. This asteroid, however, may also belong to a largely anhydrous taxonomic group. Whether outgassing-driven acceleration is possible strongly depends on how the asteroid stores volatiles in its interior and releases them upon heating at the surface. Thus far, the contentious topic of the existence of volatiles on 1998 KY26 that can cause outgassing has not been extensively examined from a thermophysical standpoint. Here, we show that all locations on the surface or in the interior of 1998 KY26 have orbit-averaged temperatures significantly above the sublimation point of volatiles in vacuum. Across 13 thermal characterizations, this observation is found to be consistent despite parameter variations in equivalent diameter, thermal conductivity, Bond albedo, specific heat capacity, and spin pole ecliptic latitude and longitude. Our temperature model results indicate that the water outgassing rate on 1998 KY26 would be too high to maintain buried ice, refining the understanding of the asteroid's composition.
Read the full abstract

Mesoscale and Nanoscale Physics

Coherent Electrometer Using Hybridised Double Quantum Dot - Resonator States

Probing the change in conductance or reactance of a charge sensitive system using an rf cavity is a standard way of doing fast single-charge electrometry. In this article, we present a detector concept which is instead based on coherent interaction. We show that the coherent interaction between a microwave resonator and double quantum dot gives rise to a fast, highly sensitive electrometer. The coherent electrometer achieves a state-of-the-art readout speed, resolving two charge states with SNR of 11.7 in 16 $\mu$s, with a readout signal two orders of magnitude lower in power than previously required to reach this speed. In this way, heating of the sample becomes negligible. Thanks to its sensitivity and low dissipation, the coherent electrometer opens up the avenue for quantum limited charge detection with nanosecond speed and minimal back-action.
Read the full abstract

All authors Samuel Andersson, Pierre Glidic, Andrea Cicovic, Harald Havir, Sebastian Lehmann, Kimberly A. Dick, Ville F. Maisi

High Energy Physics (Phenomenology)

Constraints on quirks from a Geant4-based reinterpretation of the CMS search for heavy stable charged particles

Quirks are new heavy fermions with electroweak charges that are also bound by a confining force whose scale $\Lambda$ is far below their mass. A produced pair therefore stays connected by a macroscopic string. We reinterpret the CMS search for heavy stable charged particles in 101 fb$^{-1}$ of 13 TeV proton-proton collisions for Drell-Yan produced quirks. The quirk dynamics is simulated with Geant4 in a parametric CMS-like detector, and the event selection is emulated and combined with the public likelihoods. The emulation reproduces the published limits on a heavy charged lepton within 11%. For an $SU(2)$-doublet quirk with one (two) infracolors, masses up to 870 (970) GeV are excluded for $\Lambda\lesssim70$ eV, where the quirks behave as independent heavy stable charged particles. For $\Lambda\gtrsim10$ keV, where the pair moves as one doubly ionizing object, the limit is 280 (320) GeV. In between, the quirks turn around inside the detector and the search has no sensitivity. For the $SU(2)$-singlet benchmark of the FASER search, with two infracolors, the limits reach 990 GeV below 100 eV, where FASER has no sensitivity, and 300 GeV above 10 keV, beyond the FASER exclusion of 175 GeV.
Read the full abstract

Materials Science

Atomically abrupt sp-sp$^2$ all-carbon heterojunction preserving the electronic identity of constituent allotropes

Covalently integrating distinct carbon allotropes with atomic precision remains, to date, elusive, despite the clear advantage of combining complementary functionalities while preserving their electronic identities. Here, we realize an all-carbon lateral heterojunction combining ($sp^2$) armchair graphene nanoribbons (GNRs) and $sp$-$sp^2$ hydrogenated graphdiyne (hGDY) on Au(111) through Fe-mediated hierarchical on-surface synthesis. Low-temperature scanning tunnelling microscopy and spectroscopy combined with density-functional theory reveal an atomically abrupt (<0.5 nm), asymmetric type-I interface. Redistribution of an edge-localized GNR state onto the first hGDY unit demonstrates local electronic continuity, whereas their characteristic electronic structures re-emerge immediately beyond the interface. Calculations further indicate that heterojunction formation weakens the interaction of both constituents with the underlying Au(111) substrate, widening their gaps and reducing adsorption-induced interfacial polarization. These results establish lateral covalent integration as a route to engineer atomically defined electronic interfaces and substrate coupling in all-carbon nanostructures.
Read the full abstract

All authors Alice Cartoceti, Masoumeh Alihosseini, Gianni Conti, Eliecer Pelaez-Sifonte, Alessio Orbelli Biroli, Francesco Sedona, Paolo D'Agosta, Francesco Tumino, Andrea Li Bassi, Jorge Lobo-Checa, Carlo S. Casari, Simona Achilli

High Energy Physics (Experiment)

Sensitivity to Sterile Neutrinos with Elastic Neutrino-Electron and Highly Elastic Neutrino-Nucleus Scattering in the DUNE Near Detector

The Deep Underground Neutrino Experiment (DUNE) is a next-generation neutrino experiment designed for precision measurements of long-baseline three-flavor neutrino oscillations. The DUNE Near Detector (ND) is also sensitive to short-baseline oscillations arising from mixing with an eV-scale sterile neutrino. In this paper, we present a sensitivity study for sterile-neutrino oscillations in the $3+1$ framework using three complementary event samples in the simulated DUNE ND: highly elastic $\nu_\mu$ charged-current (CC), highly elastic $\nu_e$ CC, and neutrino-electron elastic-scattering ($\nu+e$) events. These highly exclusive samples are chosen to reduce neutrino-interaction and flux uncertainties: the low-hadronic-energy CC selections enhance the correlation between reconstructed lepton energy and neutrino energy, while the precisely calculable $\nu+e$ cross section provides an independent constraint on the neutrino flux. Residual interaction-model dependence remains in the CC samples. The statistical analysis considers $|U_{e4}|^2$, $|U_{\mu4}|^2$, $|U_{\tau4}|^2$, and $\Delta m^2$, with sensitivity results presented as fixed-$\Delta m^2$ slices for $|U_{\tau4}|^2=0$. The strongest sensitivity occurs for $4 \lesssim \Delta m^2 \lesssim 10~\mathrm{eV}^2$, where oscillations produce resolvable distortions of the reconstructed-energy spectra. For one year of on-axis forward-horn-current running, the 90% confidence-level sensitivity reaches approximately $|U_{\mu4}|^2 \simeq 0.01$ and $|U_{e4}|^2 \simeq 0.015$, depending on the value of the other mixing parameter.
Read the full abstract

Symplectic Geometry

Full fillings at the accumulation point and triangular limits of almost toric diagrams

Let $M$ be a closed symplectic four-manifold admitting an almost toric fibration over a disk. The ellipsoid embedding function of $M$ has an infinite staircase if it has infinitely many nonsmooth points. Cristofaro-Gardiner, Holm, Mandini, and Pires showed that these points must then accumulate at a solution of an explicit quadratic equation, and that at this accumulation point the ellipsoid embedding function agrees with the volume obstruction. We call this a full filling at the accumulation point, and study it in terms of almost toric fibrations. We show that if $M$ admits a sequence of almost toric base diagrams, each with a Delzant corner, converging nontrivially to a triangle, then $M$ has a full filling at its accumulation point. For the one-point blowups $H_b$ of $\mathbb{CP}^2$, we also prove the converse. As a consequence, $H_b$ has an infinite staircase if and only if it admits such a sequence converging to a triangle whose hypotenuse has irrational slope.
Read the full abstract

Cryptography and Security

Testing CCZ-equivalence to permutations via projections of Walsh zero spaces, with applications to APN functions

Whether almost perfect nonlinear (APN) permutations exist in even dimension $n\ge 8$ is a long-standing open question, the big APN problem. The only example known in even dimension, on six bits, was found from a quadratic APN function that is not itself a permutation, by Carlet–Charpin–Zinoviev (CCZ) equivalence. It is therefore natural to test APN functions for CCZ-equivalence to a permutation. This amounts to finding two complementary $n$-dimensional subspaces among the zeros of the Walsh transform, which lie in $\mathbb{F}_2^{2n}$. A known necessary condition only uses their projections to $\mathbb{F}_2^n$ and cannot decide every case. We give an exact test for functions whose Walsh supports are affine, including all quadratic functions. It reconstructs the subspaces from their projections, through searches for subspaces of $\mathbb{F}_2^n$ and for constrained bilinear maps. On the benchmarked functions in dimension twelve, it lists them $39$ to $55$ times faster than the generic search of the commonly used sboxU library. We then study the functions $x^3+a^{-1}\mathrm{Tr}(a^3x^9)$, which Göloğlu and Pavlů proved not to be CCZ-equivalent to permutations when $4$ divides $n$. With their result, our computations show that this holds for every even $n\le 24$. In dimension $18$ the necessary condition is satisfied, and the new test decides the case. We conjecture this for every even dimension. We also compile $25$ classes of quadratic APN functions in dimension ten from published instances, with equivalence certificates. No function in these classes or in their switching closure is CCZ-equivalent to a permutation. All code and data are public.
Read the full abstract

Optimization and Control

Transport Less, Solve Faster: Exploiting sparsity in algorithm design for Unbalanced Optimal Transport

In this paper, we study the computational aspects of the Unbalanced Optimal Transport (UOT) problem between two given positive measures $\mu$ and $\nu$. The UOT is a variant to the classic Optimal Transportation problem in which the linear constraints modeling mass preservation are relaxed and replaced by a convex penalty $\Dphi$ in the objective function. First, we show that any UOT problem admits a discrete solution as long as $\mu$ and $\nu$ are discrete, allowing us to express the optimality conditions characterizing the solution to the UOT problem in an easy-to-handle form. By exploiting the optimality conditions we provide an upper and lower bound on the marginals of the solution as well as a bound on the mass that any solution assigns to any couple of points. The bounds on the mass assigned to each couple of points reveal that, under suitable conditions, the support of any solution to the UOT is contained within a compact set that does not depend on the measures $\mu$ and $\nu$, but only on the penalty function $\Dphi$ and the cost function $c$. We then introduce a truncated version of the UOT problem that fully leverages the sparsity of the optimal solution by introducing a parameter $R$ that saturates $c$. We show that any solution to the truncated problem naturally yields an approximation of the optimal solution for the standard UOT problem and bound the primal sub-optimality gap between the real solution and the solution induced by the truncated problem via a decreasing power-like function of $R$. Lastly, we study the convergence speed of the Pairwise Frank-Wolfe algorithm when applied to the truncated UOT problem and show that such routine is able to identify the support of the solution within a finite number of iterations.
Read the full abstract

All authors Gennaro Auricchio, Francesco Rinaldi, Saverio Salzo, Giovanni Spisso

Machine Learning

Match the Geometry Before the Surrogate: Extreme Low-Budget Optimization in High Dimensions

Low-budget molecular optimization is often framed as Bayesian optimization in a learned latent space, but standard benchmarks typically allow hundreds or thousands of evaluations, whereas a realistic wet-lab round may allow only a few dozen. We revisit this regime using a SELFIES molecular variational autoencoder (VAE) with a Gaussian prior and GuacaMol objectives at N = 30 evaluations. The main finding is not a new surrogate model, but an interaction in candidate generation. The encoder means of a reference set of molecules concentrate near a narrow spherical shell, while many high-dimensional optimizers propose candidates from boxes or box-shaped trust regions. We study a simple constraint, the on-sphere scaffold, that restricts candidate generation to this shell. We then introduce two minimal optimizers to probe what remains useful once the support is fixed. GeoWalk performs geodesic local search on the shell without a surrogate. After 30 evaluations, it shows no detected difference from a Gaussian-process (GP) surrogate on the shell, which is ahead earlier in the budget. GeoCMA adds covariance learning in a random subspace. It does not improve on GeoWalk in the molecular benchmark. On LassoBench it has a lower mean loss than GeoWalk, but full-dimensional CMA-ES remains stronger on most tasks. A factorial experiment with a single GP crosses the candidate support (shell or box) with the acquisition maximizer (a finite candidate set or a continuous optimizer). In this benchmark, shell candidates outperformed box candidates when acquisition was maximized over a candidate set, but not with continuous maximization by L-BFGS-B. These results come from a controlled study of one Gaussian-prior VAE benchmark and do not recommend VAEs for molecular optimization. The practical lesson is to control the candidate support and the acquisition maximizer before attributing a gain to the surrogate.
Read the full abstract

Computation and Language

I play of the cello: Evaluating Constructional Bleeding in Multilingual LLMs with CALQUE

Languages use different morphosyntactic strategies to express the same meaning. Speakers of multiple languages are able to apply strategies from one language to another, either by mistake or intentionally, to produce a literal translation, or calque. Considering that large language models (LLMs) have been shown to generalize crosslingually, despite being predominantly trained on English, we investigate whether they overgeneralize constructional strategies from English to other languages. To test this, we create a dataset of 5583 sentence pairs across seven languages, matching literal translations of strategies from English into other languages, and literal translations of the other languages' strategies into English. We then quantify constructional bleeding by measuring how often strategies from English are assigned a higher probability than native strategies. We find that models often fail to assign higher probabilities to native strategies and tend to prefer English-derived strategies even when evaluating sentences in other languages. Our findings highlight the limitations of the current English-dominant training paradigm.
Read the full abstract

All authors Toshiki Nakai, Yu-Yin Hsu, Marie-Catherine de Marneffe, Joakim Nivre, Jakob Prange, Nathan Schneider, Wesley Scivetti, Tiago Timponi Torrent, Leonie Weissweiler

Artificial Intelligence

Do large language models rank moral values? A transitivity test of moral competence

LLMs are increasingly used to support decisions that require moral consideration. While LLMs perform well on morality benchmarks from psychology and philosophy, producing appropriate outputs on these benchmarks (moral performance) does not establish that an LLM weighs competing moral values at stake in a decision (moral competence). We test a minimal, necessary condition for such weighing: that an ordinal ranking can represent LLM moral choices. Such a ranking requires transitivity. To assess transitivity, we presented LLMs with pairwise trade-offs among moral value violations in vignette choice sets and extracted resulting choice probabilities from first-token distributions. Tested against three stochastic transitive choice models, most choice sets admitted weak or unstable orderings. Aggregating an LLM's responses, a common practice in LLM moral evaluation, produced transitive structure that matched no individual choice set. Evaluations of AI morality should establish that choices are structured before asking whether they are aligned.
Read the full abstract

All authors Katelyn A. Wasson, Kiwon Song, James M. Jennings III, Brittney Currie, Konstantina Sokratous, Clintin P. Stober

Robotics

Look Only Where You Plan for Real-Time End-to-End Urban Driving

Dense bird's-eye-view (BEV) construction accounts for 60% of the parameters in a strong recent driving planner. This paper presents LOWP, a planner that looks only where it plans: two sets of sixteen learnable queries read the cameras through the known calibration, one by sampling image features at projected 3D anchors and one by attending to features labeled with their viewing rays, so no dense BEV is built and no object or map is supervised. Evaluations on open-loop nuScenes, pseudo-simulated NAVSIM and closed-loop Bench2Drive show competitive planning with a compact representation. On nuScenes, average L2 changes from 0.302 to 0.296 m with 53% of the dense counterpart's parameters and 1.7 times its speed on the same GPU; the anchor-based reader performs better on turns than the global reader. On NAVSIM, scoring the planner's own trajectories alongside a vocabulary gives 85.9 PDMS without retraining. The closed-loop control configuration scores 65.5 on 218 routes, with a brake-induced standstill failure whose release rule has an inconclusive aggregate benefit. We distinguish measured latency from hardware-transferred estimates, which do not establish a common-hardware ranking.
Read the full abstract

All authors Haohua Que, Tianle Zhu, Qianyi Wu, Zhipeng Bao, Handong Yao

Computer Vision

DAC-MACE: Context-Conditioned Expert Routing for Compact Scene Coordinate Regression in Camera Relocalization

A camera becomes an absolute pose sensor when its images can be relocalized against a prebuilt map of the scene. On mobile robots and augmented-reality devices, that map must be small and quick to build. Scene coordinate regression (SCR) stores the scene in a compact network that predicts a 3D coordinate for every pixel, and mixture-of-experts (MoE) variants extend its capacity to larger outdoor scenes. In existing MoE-based SCR, however, the router assigns each pixel to an expert from the same local descriptor that the experts regress from, so repeated facades and windows at different locations receive similar routing inputs. We present DAC-MACE, which conditions soft expert routing on recalibrated and contextualized features. A residual context adapter (RCA) reweights the channels and locations of frozen encoder features, a global context encoder (GCE) mixes them with multi-scale self-attention, and an image-level pooled descriptor is added to the router input. A dual-threshold RANSAC–Levenberg–Marquardt estimator recovers the pose. On Cambridge Landmarks, DAC-MACE attains a mean per-scene median translation error of 13.0 cm over three seeds, the lowest among the RGB-only SCR methods compared (GLACE 13.6 cm, MACE 14.4 cm), and reduces the error on the largest scene from 24 to 19 cm relative to MACE. In a cumulative study, the MoE head, GCE, RCA, and the dual-threshold estimator lower the mean error from 17.0 cm to 14.4, 14.0, 13.6, and 13.0 cm. On 7-Scenes, adding GCE to a single ACE head raises Recall@(5 cm, 5$^\circ$) from 86.56% to 87.60%.
Read the full abstract

All authors Dingying Fan, Mingkai Liu, Haohua Que, Haojia Gao, Xianliang Huang, Qian Zhang, Xueqian Wang, Fei Qiao

Computer Vision

Domain Generalization for Style-Debiased Object Detection via Label-Consistent Training across Multiple Styles

Domain generalization (DG) for object detection aims to improve model performance on unseen target domains. Existing detectors, however, suffer substantial degradation on visually distinct unseen domains, as models confined to limited training styles tend to overfit superficial statistical cues rather than object-level semantics. Motivated by human perception, which attends primarily to semantically relevant features while ignoring irrelevant style cues, we propose Origin-Anchored Multi-Style Consistency (OAMSC), a source-only framework that couples multi-style data synthesis with cross-style consistency learning to suppress style-sensitive shortcut learning. For data preparation, we leverage off-the-shelf diffusion models to expand the source datasets into 20 distinct styles, with all generated images manually verified to ensure that object instances and annotations remain consistent across style variants. During training, we impose Origin-anchored consistency constraints at scene and object levels to reinforce the learning of domain-stable features while suppressing the influence of style-sensitive cues. Extensive experiments and ablation studies on multiple challenging DG benchmarks demonstrate the consistent effectiveness and strong generalization performance of OAMSC across diverse domain shifts. Notably, our approach applies to both one-stage and two-stage detectors, while its data-centric design is complementary in principle to other DG methods. The training code and two style-transfer datasets constructed for the experiments, comprising approximately 180K images in total, are publicly available on GitHub.
Read the full abstract

All authors Chaoran Zhang, Shirshak Bhusal, Zhilin Gao, Qiang Xue, Haoting Shen

Artificial Intelligence

Is Recursive Self-Improvement Safe?

The phrase recursive self-improvement (RSI) refers to the idea of an AI system designing and creating another, better AI system, which in turn designs and creates an even better one, and so on. While the idea has a long history, it has always been an abstract, futuristic idea – until recently. Today's frontier AI has attained a set of abilities, including in programming, mathematics, research, tool use, and working coherently over extended periods of time, that make RSI seem plausible. AI companies are starting to pursue it, but many of the involved people are also expressing concerns, asking whether humanity should really set this in motion and under what conditions. I will discuss the situation, argue that we do not even have clear frameworks for thinking about the likely outcomes of RSI, and do a quick game-theoretic analysis of how to choose whether to initiate RSI when another party is making the same choice.
Read the full abstract

Computer Vision

STAIRS: Staging the Inverse Problem via Admissible-Support Scheduling for Unrolled MRI Reconstruction

Accelerated magnetic resonance imaging (MRI) forms images from incomplete Fourier measurements, shifting acquisition burden to computational reconstruction. Deep unrolled networks interleave learned regularization and data consistency, yet typically estimate the full unmeasured spectrum at every cascade. We propose STAIRS (Staging the Inverse Problem via Admissible-Support Scheduling), which projects intermediate reconstructions onto nested Fourier supports that expand to the full spectrum and conditions each update on its support transition. Conditional increments of a regularized log-determinant score of training covariance define a normalized spectral coordinate, accounting for dependencies between newly and previously admitted frequencies. Dyadic targets allocate larger coordinate increments to earlier cascades and are mapped to fixed supports before training, enabling coarse-to-fine reconstruction with unchanged sampling masks and data consistency. On fastMRI and Stanford2D, STAIRS improves matched Transformer (HUMUSNet) and convolutional (NAFNet) backbones in single- and multi-coil reconstruction, including radiologist-annotated pathology regions. It also outperforms predictor-based progression with lower parameter overhead and inference latency. Stage-wise analysis shows repeated error correction by projection and lower intermediate reconstruction errors than the backbones alone throughout support expansion. Ablations favor dyadic allocation over uniform expansion for square and line-band supports and over manual front-loading for square windows. Support projection alone improves reconstruction, and transition conditioning strengthens the gain. Cross-dataset evaluation shows improvements for both backbones in both directions. Support allocation thus complements regularizer design in unrolled MRI reconstruction.
Read the full abstract

All authors Xiangjian Hou, Chang Ni, Chao Qin, Zhiwei Tan, Xin Wang, Xinrui Ye, Rong-Rong Chen, Chun Yuan, Xiaodong Ma

High Energy Physics (Phenomenology)

Evidence for quark+quark pairing in proton formation

The question of pairing between quarks inside baryons is addressed by calculating the fragmentation functions that describe production of nucleons (protons and neutrons) in high-energy collisions and comparing attendant predictions with phenomenological global fits to available data and directly with data on hadron multiplicities. The fragmentation functions are obtained from nucleon parton distribution functions using the Drell-Levy-Yan relation and subsequently solving a coupled set of pion, kaon, and nucleon jet cascade equations. Strong empirical evidence is found for the presence of diquark correlations. In comparison with global fits, $d$ (down) quark to proton fragmentation ($d\to p$) exhibits the clearest signals. The strongest evidence is found in comparisons with the momentum-dependent proton multiplicity distribution: FFs built from a proton wave function that contains diquark correlations deliver a $z$-profile that is fully consistent with available data, whereas that obtained from a no-pairing wave function is disfavoured by data at the $9\sigma$ level.
Read the full abstract

All authors Hui-Yu Xing, Dan-Dan Cheng, Craig D. Roberts, Jose Rodriguez-Quintero

Computation and Language

Beyond the Prompt: Leveraging Pre-Decoding States for Jailbreak Detection in dLLMs

Diffusion large language models (dLLMs) have recently emerged as a strong alternative to autoregressive models, generating responses by iteratively refining masked tokens in parallel rather than decoding sequentially. While enabling more efficient and flexible inference, dLLMs remain vulnerable to jailbreak attacks, including recent attacks that exploit properties unique to dLLMs such as bidirectional conditioning and parallel masked-token refinement. In this work, we show that these same properties also expose a safety-relevant detection surface absent from standard autoregressive decoding: even when a jailbreak is difficult to identify from prompt representations alone, it can still be detected from hidden states over masked response positions before any tokens are generated. By training linear classifiers on prompt and pre-decoding response representations, we empirically show that neither view strictly dominates, with each detecting attacks that the other misses. We introduce $\texttt{ReFuse}$ (Representation Fusion), an inference-time detector that fuses the sigmoid harmfulness scores of the two classifiers. Without modifying model weights or the decoding procedure, $\texttt{ReFuse}$ reduces average ASR across multiple jailbreaks from 63.29% for the undefended model and 11.27% for a prompt-only classifier to 3.31%, while keeping average benign refusal on standard utility benchmarks below 1% on LLaDA-8B. Together, these results establish $\texttt{ReFuse}$ as a lightweight jailbreak detector for dLLMs that improves safety beyond prompt-only detection.
Read the full abstract

All authors Adam Hazimeh, Amel Abdelraheem, Ke Wang, Mariam Salman, Ljiljana Dolamic, Gérôme Bovet, Pascal Frossard

Human-Computer Interaction

VeV: Video-Enhanced Vibration Data Augmentation For Structural Difference

Surface-mounted vibration sensing provides non-invasive structural monitoring with high temporal fidelity across diverse ubiquitous computing scenarios such as smart farms, yet structure-borne wave models remain fragile to physical relocation and structural shifts. However, addressing this spatial data discontinuity remains limited by costly manual recalibration, insufficient physical grounding in data-driven synthesis, and substantial sim-to-real gaps in numerical digital twins. To resolve this spatial bottleneck without labor-intensive data recollection, we propose VeV, a physics-consistent spatial augmentation framework that synthesizes vibration responses at unmeasured target coordinates from sparse reference observations by coupling a monocular Vision-Derived Propagation Possibility Map for structural wave-guiding paths with a Visual Feature Map for transient source kinetics and implicit material dynamics. Evaluated on a 1,237 sensor-hour sow pig and piglet farm across disparate structures, VeV reduces waveform reconstruction MSE by over 57.5% compared to data-driven baselines and achieves up to 81.69% zero-shot classification accuracy, while reaching peak performance within just 30-50 calibration chunks to effectively eliminate cold-start retraining overhead and establish a scalable foundation extending to human-centric floor vibration sensing and ambient intelligence.
Read the full abstract

All authors Joonhee Lee, Jiale Zhang, Sungmin Lee, Ruiqing Wang, Jeffrey Shulkin, Yiwen Dong, Pei Zhang, JeongGil Ko

Plasma Physics

GUERNICA: A continuum kinetic code for neutral dynamics in fusion plasmas

We introduce GUERNICA, a GPU-accelerated continuum kinetic code for atomic hydrogen transport in fusion edge plasmas. It solves the Boltzmann equation with charge exchange, ionization, recombination and elastic neutral-neutral collisions. GUERNICA discretizes two-dimensional configuration space with the discontinuous Galerkin method on unstructured meshes and velocity space with the discrete velocity method in three dimensions, enabling direct simulation of the neutral distribution function without statistical noise. Unlike Monte Carlo methods, which introduce statistical noise and become computationally expensive in strongly collisional regimes, GUERNICA provides a deterministic alternative. An FFT-based evaluation of the full charge exchange operator reduces its computational complexity from $O(N_v^2)$ to $O(N_v \log N_v)$, enabling computationally practical simulations in three-dimensional velocity space. We verify GUERNICA against analytic tests and validate it against one-dimensional Monte Carlo results from DEGAS2. The numerical tests demonstrate the expected convergence orders and discrete conservation properties, while the DEGAS2 comparisons show agreement in neutral density and temperature to within 1.6%. Performance studies confirm the expected scaling of the FFT-based charge exchange operator and demonstrate efficient execution on a single GPU. Finally, we demonstrate GUERNICA in two dimensions using a static plasma background from a MAST-U SOLPS-ITER simulation, with $8.6\times 10^8$ phase-space degrees of freedom on a single GPU, illustrating the method's ability to resolve neutral dynamics in experimentally relevant divertor conditions.
Read the full abstract

Algebraic Geometry

Automorphisms of Kummer and Artin-Schreier curves defined by Chebyshev polynomials

Let phi_m be the monic Chebyshev polynomial characterized by phi_m(t + t^-1) = t^m + t^-m. We study the geometric automorphism groups of the smooth projective models of y^n = phi_m(x) and y^p - y = phi_m(x). We classify the hyperelliptic family in genus at least two, in characteristic zero and odd characteristic. For every m >= 5, with p not dividing m in positive characteristic, we determine the projective stabilizer of the roots of phi_m together with infinity. For tame Kummer covers, we describe the normalizer of the cyclic deck group and compute it explicitly for m >= 6. It is the full group when m > 2n. If m >= 3 and n > 2m, the full group has order 2n in characteristic zero and in every characteristic prime to 2nm. The proof uses the dihedral Galois closure of the map to the y-line and Dickson's subgroup classification. In odd characteristic, we also classify all Artin-Schreier curves in the family with genus at least two. For m >= 3 prime to p >= 5, the group has order 2p. In characteristic three, exactly five degrees are exceptional. We give the groups and their actions, and treat the quadratic cases separately.
Read the full abstract

Distributed Computing

FirstSwap: Pre-Consensus Trade Approvals for Decentralized Market Design

Decentralized exchanges typically give traders no execution guarantee until a block is finalized. In the meantime, external prices move, traders strategically delay submission, and block producers can influence inclusion and ordering. We introduce FirstSwap, a DEX built on a constant-function market maker pricing curve that gives traders an early pre-approval before consensus, guaranteeing inclusion and execution within their submitted price limit, with final ordering determined by limit prices rather than by a block producer. We demonstrate the benefits of trade pre-approvals in an economic model combining auction theory with CEX-DEX arbitrage. When the CEX price changes gradually, FirstSwap rewards earlier traders with better prices, discouraging strategic delay and yielding higher social welfare and gains to AMM liquidity providers than a batch auction. When the CEX price jumps abruptly, FirstSwap yields higher welfare and smaller AMM losses than fair ordering. The central challenge is delivering fast trade pre-approvals without consensus. We introduce a decentralized, scaled admissibility protocol under which each validator reserves only a fraction of the available liquidity at every price. A trader obtains approval in a single round of messages with a quorum of validators. Somewhat surprisingly, despite Byzantine validators who may approve trades arbitrarily, we show that all quorum-certified trades are jointly admissible and execute within their limits. Our protocol extends naturally to a setting where buy and sell orders can be pre-approved against each other before a block is finalized.
Read the full abstract

All authors William Diamond, Fatima Elsheimy, Dahlia Malkhi, Geoff Ramseyer, Srivatsan Sridhar

Strongly Correlated Electrons

On the nature of low-energy electronic states in twisted bilayer graphene: a first-principles study

The nature of low-energy electronic states in twisted bilayer graphene (TBG) near the magic angle has been extensively discussed in the literature on qualitative grounds, since these states play a key role in the construction of effective few-band models toward a many-body theory. Here, we report extensive and systematic first-principles calculations of TBG electronic states in the magic-angle range, using density functional theory with no further approximations. We construct moiré-scale localized Wannier states for the low-energy bands and develop an effective ten-band model for each $K$ valley. The Wannier orbitals contain both atomic and moiré-scale details and are centered near the high-symmetry stackings in the moiré cell. We calculate Coulomb and exchange integrals of these localized states, quantities that are the basic ingredients in many-body models. We also compare the values of these integrals to those obtained or assumed in previous works. By symmetrizing the ten-band Hamiltonians and basis orbitals, we provide insight into how the dominant hoppings between the flat-band orbitals evolve as a function of twist angle.
Read the full abstract

All authors Albert Zhu, Daniel Bennett, Daniel T. Larson, Mohammed M. Al Ezzi, Efstratios Manousakis, Efthimios Kaxiras

Quantum Physics

Can One Hear the Size of a Clock?

Weyl's law counts the vibration modes that fit into a cavity, and a quantum system with $d$ levels can oscillate at up to $d(d-1)/2$ frequencies without decay. For systems that jump at random between states, no such count was known: they oscillate only out of equilibrium, and the random timing of their jumps damps every oscillation. We find the missing count: if $Q$ is the ratio of frequency to decay rate, a system of $d$ states sustains fewer than $2(d-1)/(\pi Q)$ oscillations whose ratio exceeds $Q$. Nested cycles approach this bound, and measured oscillations reveal how large a hidden system must be.
Read the full abstract

Machine Learning

TORNADO: Adaptive Latent Stochastic Transport for Calibrated Probabilistic PDE Forecasting

Probabilistic neural surrogates for PDEs enable ensemble forecasting, but often exhibit miscalibration when used for uncertainty quantification, especially in compute-constrained, few-step inference regimes. Existing generative approaches typically rely on fixed global stochasticity, which cannot adapt predictive uncertainty to the local dynamical state. We propose TORNADO, a latent probabilistic forecasting framework that formulates next-step prediction as stochastic transport between VAE encoder posteriors. Unlike existing approaches with globally prescribed stochasticity, TORNADO learns a SDE with state-dependent diffusion, enabling the predictive uncertainty to adapt to the local latent dynamics. The model is trained using analytic conditional endpoint targets derived from a Gaussian interpolant, enabling simulation-free learning. An additional energy distance regularization term improves distributional alignment. Empirically, TORNADO improves uncertainty calibration over representative generative baselines, including stochastic interpolants, diffusion models, and flow matching, while maintaining comparable predictive accuracy. Our results show that adapting stochasticity to the latent state, rather than prescribing it globally, is an effective mechanism for improving calibration in practical neural PDE forecasting settings. Project page: https://tornado-sde.github.io.
Read the full abstract

All authors Jorge Mifsut-Benet, Armand Kassaï Koupaï, Ramon Daniel Regueiro-Espino, Nicolas Baskiotis, Patrick Gallinari

Analysis of PDEs

Stability in the Non-Neutral and Neutral Regimes of the Holonomy-Based Inverse Problem in the Class \(\mathcal{OC}\)

We study anchor reconstruction stability in return-map inverse problems for class $\mathcal O_C$. Local identifiability is governed by the first nonvanishing derivative of the holonomy map $H = T_\gamma - \text{Id}$. In the non-neutral case, we prove local Lipschitz stability and multiplier estimates. In the neutral regime, using Faà di Bruno equations and Malgrange preparation, we derive quantitative $1/k$-Hölder estimates for order-$k$ anchor perturbations. We also establish finite jet stability, real-analytic dependence, and global $C^0$-stability of order $O(\varepsilon_F^{1/k})$. Numerical benchmarks on circular-core return geometries validate these theoretical convergence rates and square-root laws.
Read the full abstract

Computer Vision

Beyond Frame-Level Accuracy: Analyzing Video-Level Stability in Transformer-Based Deepfake Forensics

Deepfake videos generated by advanced generative models pose serious threats to digital security, misinformation control, and media forensics. Although numerous detection methods have been proposed, their reported performances are often based on heterogeneous preprocessing pipelines, inconsistent data splits, and model-specific training protocols, hindering fair architectural comparison. This work presents a controlled and reproducible comparative study of convolutional neural networks (ResNet-18 and EfficientNet-B0) and a transformer-based architecture (ViT-B/16) within a unified, face-centric deepfake detection framework. All models are trained using identical preprocessing, augmentation, optimization settings, and data splits on Celeb-DF v2 and FaceForensics++, and results are validated across three independent random seeds with bootstrapped confidence intervals. Experimental results show that while CNN-based models perform competitively at the frame level, their performance degrades after video-level aggregation. In contrast, ViT-B/16 demonstrates more stable video-level performance, achieving 92.8% [95% CI: 91.7–94.0] video-level accuracy and an AUC of 0.950 [95% CI: 0.934–0.964] on Celeb-DF v2 (3-seed bootstrapped). Cross-dataset evaluation across the same three random seeds reveals a large, asymmetric generalization gap (AUC 0.79 FF++$\rightarrow$Celeb-DF vs. 0.64 Celeb-DF$\rightarrow$FF++), indicating that in-domain video-level stability does not transfer to unseen manipulation domains. These findings indicate that global self-attention improves robustness and video-level stability for deepfake detection within a given data domain, while highlighting the need for further work on cross-domain generalization.
Read the full abstract

All authors Syed Azmul Hasan, Mowmita Pervin Hera, Afroza Lona Mim, M M Jihad Hossain, Hazrat Ali, Mohammad Farhad Bulbul

Machine Learning

FedLoVA: Value-Only Aggregation for Federated LoRA Fine-Tuning of Large Language Models

Low-Rank Adaptation (LoRA) is a popular parameter-efficient fine-tuning (PEFT) method for fine-tuning of large language models (LLMs), especially in federated learning, due to its strong performance and communication efficiency. In practice, LoRA is applied to the query and value projections of transformer layers without considering the distinct roles of these components. In this work, we analyze the impact of these projections in federated fine-tuning of LLMs using LoRA and observe that the value projection converges significantly faster than the query projection. Motivated by this finding, we propose FedLoVA, a federated fine-tuning framework that trains LoRA adapters on both query and value projections locally while sharing only the value projection with the server. This approach reduces communication overhead while improving convergence efficiency. Experiments on natural language understanding, multilingual sentiment analysis, and generation tasks show that FedLoVA reduces the communication cost by up to $7.8\times$ compared to existing federated LLM fine-tuning baselines while achieving comparable or better performance. The code is available at: https://github.com/EnsiehKhazaei/FedLoVA.
Read the full abstract

Machine Learning

Can Time Series Foundation Models Serve as Surrogates for Power Systems Dynamic Simulation?

High-fidelity dynamic simulation is essential for scientific and engineering applications, but its computational cost limits large-scale scenario screening. We investigate the use of pretrained time-series foundation models (TSFMs) as computationally efficient surrogates to address this limitation. In-context learning and post-training strategies, including retrieval-augmented reference sample selection and physics-informed regularization, are considered in the investigation. Results are benchmarked against a traditional AI model representative of the state-of-the-art in the domain. Experiments using an open-source TSFM and data from a simulated 9-bus power system indicate promising results. Retrieval augmentation improves in-context learning performance, and fine-tuning leads to superior results compared to the traditional AI model, especially when including physics-informed regularization. The fine-tuned TSFMs achieve lower errors and improved computational costs compared to the traditional AI model, and a 563$\times$ speedup compared to the original physics-based simulation.
Read the full abstract