Peer Revenue Publish or profit

Papers

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

All of arXiv

Thursday 3 September

Quantum Physics

Robust multi-hypothesis quantum-state discrimination under unknown common unitary perturbations via least favorable priors

We study robust K-ary quantum-state discrimination when all candidate states are affected by the same unknown common unitary perturbation. The unknown perturbation does not represent the label to be identified, but acts as a nuisance factor that changes the performance of a fixed measurement. We formulate the problem as a minimax decision problem over the possible perturbations and propose using the Bayes-optimal collective measurement associated with a least favorable prior (LFP) on the nuisance-parameter space. For a finite discretization of this space, we show that the LFP can be computed by a semidefinite program and that the corresponding value coincides with the finite-grid minimax success probability. As numerical demonstrations, we consider a binary nonorthogonal qubit model and a nonorthogonal three-state qutrit model with an unknown common unitary perturbation. The LFP-based measurement substantially flattens the success-probability profile and improves the worst-case success probability compared with a reference-point optimal measurement and a uniform-prior Bayes measurement. The resulting LFP concentrates its weight on regions of the nuisance-parameter space that actively limit the robust discrimination performance, thereby providing both a constructive measurement design and a diagnostic description of the difficult nuisance-parameter regimes.
Read the full abstract

Optimization and Control

DREAM: Deep-Reparametrization of Adaptive Regularization Maps for Fast Zero-Shot Self-Supervised Learning

Adaptive regularization is an effective means of improving the flexibility of classical variational reconstruction methods while retaining their interpretability and mathematical structure. In this work, we propose an approach for learning deep-reparameterized adaptive regularization maps (DREAM) for Total Variation (TV) and Total Generalized Variation (TGV) through algorithm unrolling in a zero-shot, self-supervised setting, requiring no access to large paired datasets. We first validate DREAM on a two-dimensional image-denoising problem and then apply it to large-scale dynamic MRI reconstruction. The resulting adaptive models are theoretically grounded, as they retain the convergence properties of the primal-dual algorithm used as the building block of the unrolled architecture. Our results show that the regularization parameter maps can be learned efficiently, requiring substantially fewer weight updates than state-of-the-art methods to achieve comparable performance. We further demonstrate that the CNN parametrization of the maps acts as an implicit prior and provides effective dimensionality reduction, outperforming direct optimization of the adaptive parameters. Across both applications, the performance gap between supervised and zero-shot training is smaller for DREAM than for the end-to-end deep-learning methods considered for comparison. Remarkably, in dynamic cardiac MRI reconstruction, zero-shot training matches the performance of the supervised model. Moreover, the learned adaptive maps are highly interpretable, providing a compelling alternative to purely deep-learning-based methods when reference data are difficult to obtain.
Read the full abstract

All authors Thanh Trung Vu, Ander Biguri, Christoph Kolbitsch, Luca Calatroni, Kostas Papafitsoros, Andreas Kofler

Machine Learning

A Location-Invariant Estimator of Extremal Quantile Treatment Effects for Heavy-Tailed Distributions

Quantile treatment effects (QTEs) measure the effect of a treatment on the distribution of an outcome, and their estimation at extreme quantile levels is of central interest in applications where the target quantiles lie far beyond the range of the data. For heavy-tailed potential outcomes, existing extremal QTE estimators rely on extrapolation combined with a causal extreme value index (EVI) estimator, but the resulting estimator is not invariant under a common location shift of the potential outcome distributions, even though the population QTE is. We address this issue in two steps. First, we adapt the location-invariant Fraga estimator of the EVI to the causal setting using inverse propensity score weighting. Second, we replace the original extrapolation formula with a difference-based scheme, under which the location parameter cancels when quantile differences are taken. The resulting QTE estimator is therefore location invariant. We establish the consistency and asymptotic normality of the proposed extremal QTE estimators, and provide a consistent variance estimator, leading to asymptotically valid inference. A simulation study confirms the location invariance, the stability with respect to the threshold, and the coverage of the proposed methods.
Read the full abstract

All authors Xin Yu, Shuwei Huang, Jicheng Liu, Jielin Tang, Bolin Wang, Yunxiao Zhang, Tian Zhao

Cryptography and Security

A Black Box for Agentic Processes: Blockchain-Anchored Evidence for AI Agent Communication, Human Oversight, and GRC Audits

Autonomous AI agents increasingly communicate with other agents, invoke tools, exchange intermediate results, and request human approvals. These workflows create a new auditability problem: organizations must reconstruct what happened, when it happened, which agent or human was involved, which control or policy applied, and whether records were modified afterwards. Motivated by the 2026 OpenAI/Hugging Face incident, this position and architecture paper proposes a product- and vendor-neutral black-box architecture for agentic processes. The architecture creates blockchain-anchored cryptographic commitments for selected agent communications, human-in-the-loop approvals, tool calls, and process artifacts without placing sensitive content on-chain. We define an evidence model that distinguishes temporal anchoring and artifact integrity from event ordering, capture authenticity, authorized anchoring, and causal traceability. The latter properties require additional architectural controls. We then discuss practical use for Governance, Risk, and Compliance (GRC), including compliance testing, risk-based evidence selection, monitoring evidence streams, incident reconstruction, and regulatory reporting readiness under the EU AI Act, NIS2, and the Cyber Resilience Act (CRA). This position and architecture paper does not present an empirical performance or security evaluation. The approach does not prevent agent misbehavior or prove semantic truth. Rather, it strengthens the evidentiary basis for later verification of critical process traces.
Read the full abstract

High Energy Physics (Theory)

Quasi-local form for $\alpha$–$z$ Rényi QNEC from fixed-ray escorts

The fixed-ray escort integral representation expresses $\alpha$–$z$ Rényi divergence as an average over ordinary relative entropy of a family of escort states. Working in the standard UV-regulated density-matrix description of QFT subregions, we use this representation to derive an escort-averaged entanglement first law, an escort-averaged representation of the $\alpha$–$z$ information kernel, and an escort-averaged Bekenstein-type bound for ball-shaped regions in conformal field theories. For the conjectural $\alpha$–$z$ quantum null energy condition (QNEC), we obtain a quasi-local form in which the null energy is evaluated in an escort-averaged state and is corrected by an escort-transport term encoding the failure of escort formation to commute with restriction to a null-deformed region. The $z=\alpha$ specialization gives a similar quasi-local form of the Rényi QNEC for sandwiched Rényi divergence. We explicitly compute the Rényi QNEC, including the explicit escort transport term, for coherent-state excitations in a free scalar field theory. For the same coherent family, we obtain a positive $\alpha$–$z$ null Hessian, verifying the conjectured diagonal $\alpha$–$z$ QNEC for this family.
Read the full abstract

Materials Science

Hierarchical automation of scanning probe microscopy through agentic orchestration and algorithmic control

Rapid advances in agentic artificial intelligence enable scientific systems to interpret open-ended objectives, combine heterogeneous information, invoke specialized tools, and revise experimental strategies as evidence accumulates. However, physical experimentation also contains many tasks for which agentic reasoning provides little advantage and can reduce reliability. Quantitative analysis, optimization, spatial targeting, validation, and instrument execution are often better posed as deterministic or algorithmic operations with explicit objectives and verifiable outputs. Here, we introduce a hierarchical architecture for autonomous experimentation that separates these roles. Agentic components interpret scientific intent, construct task-dependent experimental representations, evaluate accumulated evidence, and select high-level actions, whereas deterministic algorithms perform numerical analysis, coordinate selection, validation, and physical execution. We implement this architecture in piezoresponse force microscopy. Starting from a broad scientific question concerning the relation between local domain structure and polarization switching, the system constructs spatial descriptors from multichannel imaging, selects and analyzes local hysteresis measurements, adapts the spectroscopy waveform, and terminates the experiment when additional measurements cease to provide new evidence. The autonomous trajectory also identifies a confounding relationship between polarization state and domain-wall proximity and recognizes that the requested contrast is not independently represented within the available field of view. These results demonstrate a route toward scientific autonomy in which agents determine what evidence is required while algorithms determine how that evidence is acquired reproducibly and within validated physical constraints.
Read the full abstract

All authors Boris N. Slautin, Sheryl L. Sanchez, Aidan Swanger, Yu Liu, Gerd Duscher, Vladimir V. Shvartsman, Mahshid Ahmadi, Sergei V. Kalinin

Artificial Intelligence

InSituMeasure: Probing Situated Measurement Grounding in Industrial Scenes with Multimodal Large Language Models

For trained operators, gauge reading requires little specialized knowledge, low cognitive effort, and high repeatability. Yet Multimodal Large Language Models (MLLMs) remain unreliable in continuous-valued measurement despite strong results on general multimodal benchmarks. Existing benchmarks expose this weakness but isolate measurement from realistic, knowledge-grounded settings, with limited situated context, specialized instruments, real-world noise, and matched diagnostic annotations, reducing realism and constraining root-cause analysis. We introduce InSituMeasure to evaluate situated measurement grounding. It contains 2,922 real industrial monitoring scenes across eight functional categories of professional engineering instruments, with dense gauge-attribute annotations and noise tags for failure diagnosis. We define metrics for numerical accuracy under predefined tolerances and unit consistency, rejection of fake or unanswerable tasks, and alignment between model failures and annotated error factors. Across 24 state-of-the-art MLLMs, the best model reaches only 25.7% joint value-unit accuracy and 51.8% confidence-diagnosis F1, revealing a substantial gap between general multimodal competence and reliable situated measurement. Further analysis identifies failures from text-induced shortcuts, overconfident responses, and authentic industrial noise, including mixed disturbances, viewpoint deviation, occlusion, and environmental interference.
Read the full abstract

All authors Chao Shen, Xinyuan Li, Yunfan Zhou, Jianguo Yao, Haibing Guan, Zhihai Wang, Xijun Li

Artificial Intelligence

LLM4CKD: Large Language Models for Early Stage Chronic Kidney Disease Screening

Early screening of chronic kidney disease (CKD) is critical for timely intervention, yet most machine learning (ML) and deep learning (DL) approaches require labeled data and model training, limiting their use in real-world screening settings. This study evaluates the effectiveness of large language models (LLMs) for CKD screening under zero-shot and few-shot in-context learning settings and compares them with traditional ML and DL methods. We propose a framework that uses clinically selected tabular features and structured prompt templates to enable LLM-based inference without task-specific training. LLM performance is evaluated across multiple prompt styles, feature configurations, and data settings, and compared with standard ML, DL, and tabular foundation model (TFM) baselines, and existing CKD screening tools. The results show that LLMs can achieve competitive performance using only a small number of examples, often matching or outperforming traditional approaches in low-data settings. However, their performance remains model-dependent and less stable as input complexity increases. In contrast, ML, DL, and TFM models show more consistent improvement with larger training data. Overall, the findings highlight a trade-off between data efficiency and stability, suggesting that LLMs may serve as a flexible complementary approach for CKD screening when labeled data are limited. To facilitate further research and reproducibility, the code has been made publicly available at https://github.com/akabircs/LLM4CKD
Read the full abstract

Astrophysics of Galaxies

Two sets of potential-density basis pairs for the study of radial perturbations in collisionless spherical stellar systems

The Kalnajs matrix method is a widely used framework for studying the global linear stability and possible Landau damping of collisionless stellar systems. However, for radial perturbations ($l=0$) in three-dimensional spherical models with infinite boundaries, standard biorthogonal sets such as the Clutton-Brock basis often converge slowly. This stems from a physical constraint: mass-conserving radial modes force the perturbed potential to decay faster than a point mass at infinity, whereas individual Clutton-Brock elements carry a fictitious net mass and decay only as $\mathcal{O}(1/r)$, producing unphysical asymptotic tails. We construct two new families of potential-density basis pairs that are free of these tails by construction. The first modifies the Clutton-Brock set: a specific linear combination of adjacent elements analytically cancels the leading $\mathcal{O}(1/r)$ term, giving potentials that decay as $\mathcal{O}(1/r^3)$ and densities as $\mathcal{O}(1/r^5)$. This breaks strict biorthogonality but yields a compact tridiagonal Gram matrix that enters the response equation at negligible additional cost. The second family is built from Jacobi polynomials that embed the required $\mathcal{O}(1/r^2)$ potential decay directly into their construction while retaining strict diagonal biorthogonality. Numerical tests demonstrate convergence that is uniform in radius for both expansions: the truncation error decreases exponentially when the asymptotic tail of the expanded potential matches the parity of the basis elements, and algebraically otherwise. Both bases reduce the dimension of the response matrix required for a given accuracy and are suited to studying radial perturbations in open stellar systems.
Read the full abstract

Computational Engineering

Differentiable Hybrid Modelling for Learning and Optimising Chemical Transport Processes from Experimental Data

Transport models underpin the design and optimisation of chemical processes, yet their constitutive laws are typically hand-picked, their experimental initial conditions are assumed exact, and their computational cost restricts optimisation to a handful of process variables. Here we introduce a differentiable hybrid modelling framework that embeds neural networks within a GPU-accelerated population balance equation solver to discover constitutive laws and assimilate initial conditions directly from experimental crystallisation data. Using thirteen real experiments with multi-dimensional size and shape measurements, we show that treating measured initial conditions as exact yields a physically inconsistent rate law despite improved predictive accuracy, whereas assimilating them recovers a physically consistent, non-linear law and reduces trajectory error sevenfold relative to conventional modelling. End-to-end differentiability further enables gradient-based optimisation of high-dimensional operating profiles, identifying a 300-variable temperature profile that reduces crystal aspect ratio by 25% at constant mass, facilitating superior product attributes. Our framework augments, rather than replaces, mechanistic modelling, offering a general route to learning transport models from data and optimising processes with them.
Read the full abstract

All authors Arthur Jessop, Mohammed Alsubeihi, Ben Moseley, Ashwin Kumar Rajagopalan

Machine Learning

Unlocking Lossless Speedups in LLMs via Discrete Diffusion

Large Language Models (LLMs) owe much of their success to next-token prediction (NTP), but their autoregressive (AR) structure requires slow, sequential token generation. To overcome this bottleneck, we introduce diffusion-augmented LLMs, a new class of models that defines an AR model distribution while using diffusion to draw multiple tokens in parallel from that distribution. We decouple the parameters of these models into two sets: AR weights, trained using the standard NTP objective, and lightweight diffusion weights, trained to generate multiple tokens simultaneously. The diffusion weights are learned through a simple Diffusion Distillation phase that adds negligible overhead to existing LLM training pipelines. We also introduce $\Psi$-Spec, a family of samplers that enables lossless acceleration and inference-time scaling at a fixed context length. Unlike speculative decoding, our method requires no separate draft model. Unlike diffusion LLMs (d-LLMs), it accelerates generation without sacrificing the quality of the underlying AR model. The resulting models, called Uno, can be trained from scratch or built by augmenting existing open-weight AR LLMs. Uno achieves higher throughput than leading speculative-decoding methods at every evaluated batch size and delivers up to $3\times$ speedups over the base AR model, including at the largest batch size supported by the device. Notably, our 8B Uno model outperforms the leading open d-LLM, the 26B DiffusionGemma, and the proprietary Mercury 2 across all evaluated benchmarks in agentic tool use, coding, and long-context reasoning. We release code and checkpoints at: https://s-sahoo.github.io/uno/
Read the full abstract

All authors Subham Sekhar Sahoo, Lingjie Chen, Khiem Pham, Jonathan Geuter, Chaitanya Dwivedi, Varad Pimpalkhute, Yash Akhauri, Alexander Moreno, Mikhail Yurochkin, Zhenting Wang, Mostafa Elhoushi, Nolan Dey, Shane Bergsma, Joel Hestness, John Thickstun, Eric Xing, Zhengzhong Liu

Computer Vision

The Blind Spot in 2D Infants' Pose Estimation:Robust Learning from Noisy Annotations

Noisy annotations pose a significant challenge for supervised deep learning, as neural networks rely on large-scale, high-quality labeled data whose corruption can severely impair model performance. Although robustness to label noise has been extensively studied for classification tasks, it remains relatively underexplored in Pose Estimation (PE). This limitation becomes critical in clinical contexts, including neonatology, where PE of preterm infants is used to support the assessment of spontaneous motility, a key indicator of neurodevelopmental trajectories. In such settings, infants' images labeling is further hindered by visual challenges (e.g., keypoint self-occlusions, caregiver interference), making the annotation process inherently susceptible to errors. To tackle noisy annotations in PE, we introduce REliable keypoint selection via Memory of traINing Dynamics (REMIND), a clustering-based keypoint-selection strategy that exploits keypoint-wise training dynamics to identify noisy labels without assuming any prior knowledge of the noise distribution, thus enabling noise-free model training. When evaluated on the proprietary NeoPose dataset, comprising 46 videos of 46 preterm infants recorded in real clinical settings, REMIND correctly identifies noisy annotations across multiple corruption scenarios, achieving up to 93% Area Under the Curve (AUC) with three different PE architectures used in the relevant literature. To our knowledge, this is the first study to explicitly address label noise in preterm infants' PE, paving the way for the design of trustworthy learning-based algorithms for infants'monitoring support when data quality cannot be guaranteed.
Read the full abstract

All authors Emanuele Cardinale, Marco Proietti, Alessandro Cacciatore, Maria Francesca Spadea, Lucia Migliorelli, Sara Moccia

Computational Physics

Stochastic Operator Inference for reduced-order modeling of capillary wave turbulence using experimental measurements

Modeling complex physical phenomena directly from experimental data poses fundamental challenges: measurement devices introduce noise and biases into the data, the governing equations are often unknown or intractable, and the dynamics observed experimentally may exhibit stochastic behavior. In this paper, we consider capillary wave turbulence—an example of nonlinear wave interactions at a microfluidic interface—measured by ultra-high-speed digital holographic microscopy. With the goal to learn an efficient model directly from these data, we adapt a stochastic extension of Operator Inference to learn low-dimensional stochastic differential equation representations of the microscale wave dynamics from experimental measurements. Each experiment is repeated 16 times, and 10 different conditions are created by changing the nondimensional acoustic capillary number through an excitation device. In addition, we propose a new strategy to select the reduced-order model dimension, based on both mean and covariance errors. We further add Tikhonov regularization to the stochastic Operator Inference framework and show that, beyond its conventional role as a numerical stabilization technique in deterministic settings, it has a physically meaningful interpretation as controlling the spectral content of the learned stochastic dynamics. We demonstrate that the resulting stochastic reduced-order models faithfully capture salient physical features of capillary wave turbulence across a range of experimental conditions.
Read the full abstract

All authors Hyeonghun Kim, Lei Zhang, James Friend, Boris Kramer

Machine Learning

RobustSeiz: a model-agnostic framework for auditing the recording-environment, noise and adversarial robustness of EEG seizure detectors

Background and Objective: Automated electroencephalography (EEG) seizure detectors are validated on clean research recordings, whereas clinical recordings are altered by electrode problems, equipment settings and artifacts. As seizure treatment is time-critical, we tested whether detectors stay reliable and timely under such conditions. Methods: We developed RobustSeiz, an open-source framework that audits any trained detector without modifying it. Its novelty is the evaluation pipeline, not a new detector: a model-agnostic protocol applying controlled, seeded shifts (recording environment, noise and artifacts, worst-case adversarial), scoring that adds onset timing and model uncertainty to event-based accuracy, and paired statistics against clean recordings. We audited four deep-learning detectors on three public datasets unseen in training. Results: Robustness varied widely and was independent of clean accuracy. Adversarial changes of 5-10% of signal amplitude reduced the F1-score by up to 0.57, and electrode-contact noise silently lowered sensitivity from 0.57 to 0.06. Shifts delayed the seizures still detected by up to 62 s. Read with accuracy, dropout certainty fell under filtering shifts but rose as noise silenced detection; only a quarter of failures lowered it. Conclusions: Clean accuracy alone does not establish reliability or timeliness. RobustSeiz makes pre-deployment robustness audits reproducible.
Read the full abstract

Disordered Systems and Neural Networks

Quenched complexity of marginal states in the Sherrington–Kirkpatrick spin glass

The number of metastable states of a random corrugated landscape controls what dynamics, inference and learning can reach. In the Sherrington-Kirkpatrick model these states are marginal, and counting them requires breaking the BRST symmetry. So far only the annealed complexity was known, predicting states below equilibrium. We compute the quenched complexity function with a two-group Ansatz and one-step replica symmetry breaking. It branches continuously off the annealed curve at its instability point and vanishes near the Parisi free energy, while the soft mode turns toward orthogonality to the magnetization.
Read the full abstract

Artificial Intelligence

The Dually Flat Geometry of Planning as Inference

We present an alternative characterization of the occupancy measure of reinforcement learning, obtained by embedding the planning criterion into the dynamics through a resetting planning process. Its stationary measure, which we term visitation measure, is the object on which the information geometry of decision making is most naturally expressed. The achievable visitation measures form a dually flat statistical manifold whose two affine charts are the visitation probabilities and the log-policies, dual under the conditional entropy. This structure makes planning-as-inference generalize from linear rewards to nonlinear functionals of the visitation, each iterate solved by one natural-gradient step, and gives the temporal-difference error the interpretation of a marginal-utility estimate. We develop the geometry and its consequences for reinforcement learning and theoretical neuroscience.
Read the full abstract

All authors Nikola Milosevic, Asaki Kataoka, Nicolas Hinrichs, Kenji Doya, Nico Scherf

Information Theory

On the Impact of Site-Specific Training for a Real-World 5G NR System

Site-specific training can improve wireless receiver performance without increasing computational complexity. However, real-world results have so far focused on fully trainable neural receivers and single-layer transmissions. We study site-specific finetuning of three receiver architectures: fully trainable neural, model-driven neural, and model-based. We train and evaluate these receivers using new measurements from a standard-compliant 5G NR testbed at ETH Zurich with dual-layer uplink transmission, including measurement campaigns conducted more than six months apart. Our results show that site-specific finetuning (i) substantially improves fully trainable and model-driven neural receivers, while resulting in only marginal gains for the less tunable model-based receiver; (ii) enables a single neural receiver jointly finetuned for single- and dual-layer transmission to closely match receivers finetuned separately for each configuration; and (iii) remains effective across measurement campaigns separated by more than six months. We also investigate site-specific linear minimum mean-square error channel estimation using covariance matrices estimated from either synthetic channels or site-specific measurements. When combined with iterative detection and decoding, site-specific channel estimation achieves the lowest error rate observed in our datasets. Our finetuning code and measurement datasets are publicly available at https://github.com/IIP-Group/site_specific_training
Read the full abstract

All authors Reinhard Wiesmayr, Nuri Berke Baytekin, Chris Dick, Christoph Studer

Numerical Analysis

Closure complexity of longest-edge bisection for triangular meshes

On triangular meshes, we analyze local mesh refinement based on longest-edge bisection equipped with the serial longest-edge propagation-path closure. Ties are resolved by terminal priority: if the incoming shared edge is a longest edge of the neighboring triangle, the pair is declared terminal and that edge is bisected. For every adaptive mesh sequence $\mathcal{T}_0, \mathcal{T}_1, \ldots, \mathcal{T}_L$ with a sequence of marked sets $\mathcal{M}_0, \mathcal{M}_1, \ldots, \mathcal{M}_{L-1}$, we prove the cumulative closure estimate $$\#\mathcal{T}_L-\#\mathcal{T}_0 \lesssim\sum_{\ell=0}^{L-1}\#\mathcal{M}_\ell.$$ The proof derives single-mark locality from the uniform multiplicative gap between possible descendant diameters implied by finite similarity classes, and converts this locality into the cumulative estimate through a Binev–Dahmen–DeVore-type charging argument.
Read the full abstract

Complex Variables

Reconstruction of a slice regular function from some of its real components

A basic property of holomorphic functions $f:D\to \mathbb{C}$ defined on domains $D$ of $\mathbb{C}$ is that $f$ is uniquely determined by its real part up to an additive constant. The same is true for slice regular functions defined on circular slice domains of the division algebra of quaternions or octonions. The aim of this paper is to extend the latter result to more general classes of algebras, including, among many others, the Clifford algebras $\mathbb{R}_{p,q}$ and the split octonions $\mathbb{S}\mathbb{O}$. New phenomena appear, as well as unexpected connections with graph theory and the theory of slice-Nash functions.
Read the full abstract

Soft Condensed Matter

Designing corrugated surfaces to guide colloidal self-assembly

The self-assembly of colloidal particles enables the creation of structured materials with programmable functionalities; however, controlling interaction specificity and aggregate morphology in a reversible and scalable manner remains a major challenge. Here, we investigate the selective depletion-induced self-assembly of 3D-printed flat polygonal colloids, where nanoscale surface topography is engineered through precise modeling in two-photon polymerization. By designing anisotropic lateral surfaces, we direct specific interactions that govern aggregate morphology, yielding dimers, chains, zigzag, and honeycomb structures depending on the surface configuration. The specificity of interaction is tuned by varying the length scale of the topographic surfaces, the depletant concentration and the ionic strength of the solution, revealing a transition from selective to non-selective aggregation regimes. The relative placement of lateral interacting surfaces on the colloids enables assembly into aggregates spanning a broad range of sizes, while tuning the interaction strength selectively stabilizes distinct structural motifs. We demonstrate this interplay between geometric arrangement and interaction energy experimentally and corroborate through both theory and simulations for specifically hexagonal shaped colloids. This study establishes a versatile framework for programming colloidal interactions via micro-architectural design, offering new routes for fabricating reconfigurable and functional soft materials.
Read the full abstract

All authors Dinesh Kumar Sahu, Jude Ann Vishnu, Lisa Shafroth, Martin Lenz, Olivia du Roure, Julien Heuvingh

Combinatorics

The type and cardinality of minimal presentations of numerical semigroups with embedding dimension four

For a numerical semigroup $S$ with embedding dimension four, we study the relationship between its type $t(S)$ and the cardinality of its minimal presentations $\eta(S)$. Using an approach based on the geometry of the Apéry set, we prove that $4t(S) + 5 \geq \eta(S) \geq t(S)-11$. This resolves a problem of Moscariello and Sammartano asking whether $t(S)$ is bounded by a function of $\eta(S)$, and improves the previously known bound $9t(S) + 4 \geq \eta(S)$ due to Bresinsky.
Read the full abstract

Computers and Society

Shifting from Injection to Interaction: Rethinking Web Security in the Age of LLMs and Beyond

Large language models (LLMs) are becoming integral to web applications and browser agents, transforming online interactions while introducing new attack vectors and reshaping longstanding web vulnerabilities. Classical threats such as cross-site scripting (XSS) can be amplified through LLM-mediated interactions, while LLM-specific vulnerabilities can propagate across web applications, introducing attacks such as prompt injection. Securing modern web systems therefore requires understanding interactions between traditional and LLM-specific threats across the system lifecycle. Unlike prior surveys treating web and LLM security separately, this survey provides a unified analysis of how LLMs amplify web vulnerabilities across client-side, server-side, and pipeline layers while evaluating defenses and their limitations. The analysis examines extending NIST and ISO/IEC AI security frameworks to the security needs of LLM-enabled web environments. Three unresolved challenges are identified: adversarial natural-language instructions, autonomous agent security, and post-deployment security through continuous monitoring and adaptation. An LLM-aware monitoring and control framework is proposed, integrating semantic input validation, prompt integrity protection, output isolation, agent governance, and runtime monitoring. This unified perspective characterizes the evolving threat landscape and outlines future directions for secure AI-enabled web systems.
Read the full abstract

All authors Nivedita Singh, Alsharif Abuadbba, Yansong Gao, Surya Nepal, Hyoungshick Kim

Programming Languages

Typed Flexible-Arity Slotted E-Graphs: A Soundness Construction and an Alloy Case Study

Slotted e-graphs represent open terms modulo consistent renaming, while algebraic operators benefit from canonical sequence, bag, or set children. We compose the two at a specification level: typed slot-mapped invocations inhabit operator-declared ports whose sibling quotient and recursive flattening licenses are certified separately. A generic finite quotient presentation proves exactness of its least-orbit normal form, while certified records specify effective-support kernel extraction and collision. For abstract obligation traces carrying local endpoint certificates, we prove finite-unfolding equational soundness. An Alloy case study compares seven related pipeline arms on a frozen corpus and a controlled transformation suite.
Read the full abstract

Atomic Physics

Magnetic Dipole Hyperfine Constants of 15 Energy Levels in $^{169}$Tm II Determined by Collinear Laser Fluorescence Spectroscopy

We present measurements of the hyperfine structure, transition frequencies and relative intensities in 14 lines of 169Tm+ ions starting from the $4f^{13}({}^2\!F^o_{7/2})6s_{1/2} (7/2, 1/2)_{3,4}^o$ levels. The magnetic dipole hyperfine ($A$) constant of all 15 involved ionic levels are determined, 14 of which have sub-MHz accuracy. For five levels no previous measurements were reported, while all others are in excellent agreement with the literature and improve on the previous accuracy by typically 1 - 2 orders of magnitude. We compare transition strengths and include improved transition frequencies. These results are useful for the quantum computing community, modelling of Tm$^+$ abundances in stellar spectra, as well as for an ongoing measurement program on short-lived Tm isotopes at ISOLDE/CERN.
Read the full abstract

Fluid Dynamics

Dissipation scaling in wind turbine wakes exposed to free-stream turbulence

The nature of the dissipation of turbulent kinetic energy (TKE) is investigated experimentally in the wake of a diameter $D=0.58$m wind turbine exposed to several “flavours” of high-Reynolds-number free-stream turbulence (FST). For low- and moderate-intensity FST, an annular region of elevated normalised dissipation, $C_{\varepsilon}$, develops in the outer wake, coinciding with a ring of enhanced turbulence intermittency at both small ($\ell \leq \lambda$) and large ($\ell \geq D$) scales, where $\lambda$ is the Taylor microscale. In the blade-tip region, $C_{\varepsilon}$ scales with $\sqrt{Re_D}/Re_{\lambda}$, where $Re_D$ is a global Reynolds number and $Re_{\lambda}$ a local turbulent Reynolds number based on $\lambda$-a scaling indicative of dissipation being out of equilibrium with the inter-scale flux of TKE in the inertial range of the energy cascade. This non-equilibrium regime is interpreted in light of the observed intermittency : large-scale intermittent events (i.e., low-wavenumber perturbations/“kicks”), driven by persistent tip-vortex and tip-shear-layer dynamics under low-intensity FST, require a finite time to cascade down to the dissipative scales, thereby introducing an imbalance between the inter-scale energy flux and dissipation. At the wake centreline, by contrast, $C_{\varepsilon}$ remains approximately constant with streamwise distance, reflecting either classical Kolmogorov-type equilibrium turbulence or balanced non-equilibrium turbulence, with intermittency confined to the small scales. Under high-intensity FST, large-scale intermittency is suppressed, consistent with the erosion of tip-vortex structures, and no comparable scaling for $C_{\varepsilon}$ could be identified using a single turbulent Reynolds number for these cases, where two similarly intense streams of turbulence, but of different origins, are adjacent to one another.
Read the full abstract

Computer Vision

Catalogue Photography as a Cold Start: Toward Deployable Rotary Milling Tool Recognition

Verifying that manufactured batches of rotary milling tools, also known as carbide burrs, conform to production order sheets remains a largely manual and error-prone quality assurance task. Automating this process with computer vision faces a critical cold-start constraint since no labelled imagery from the deployment environment is available, leaving manufacturer catalogue photography as the sole source of supervision. We investigate how far catalogue supervision can support an industrial recognition pipeline under domain shift, explicitly measuring the gap between catalogue separability and performance on held-out field photographs. Our findings reveal three key insights. First, off-the-shelf frozen feature extractors do not reliably separate the two task attributes, head shape and tooth profile, motivating targeted representation learning. Second, metric learning produces near-perfect unsupervised cluster discovery on catalogue images (adjusted Rand index 0.94–0.97), yet on field photographs under half of the accuracy gained from training survives. Third, the largest transfer gains do not come from model scale or representation complexity, but from simple changes that reduce domain sensitivity: converting images to grayscale (+0.22) and constraining retrieval against the known order sheet (+0.11). We therefore treat catalogue photography as a useful cold start rather than a deployment-ready training domain, and provide empirical baselines and an evaluation protocol for catalogue-to-field transfer in precision tool manufacturing.
Read the full abstract

All authors Abilash Philip Madavath, Chandra Yuvesh Aubeeluck, Augustin Raju, Nicolas Pyschny, Felix Hackelöer, Florian Zwanzig

Strongly Correlated Electrons

Evolution of electron spin resonance through a metallic quantum critical phase diagram

In the heavy-fermion metal YbRh2Si2, quantum criticality at a suppressed antiferromagnetic order is governed by the interplay of local magnetic moments and itinerant conduction electrons. We demonstrate how this can be investigated by a new experimental approach that enables the observation of electron spin resonance (ESR) across a broad range of frequencies and fields at very low temperatures. This allowed us to cover a large part of the phase diagram from the paramagnetic Fermi-liquid phase to the phase with antiferromagnetic order and including the quantum-critical regime. Both the ESR g-factor and the linewidth present distinct behaviors in these three regimes, providing further insight into the physics across a quantum critical point. Notably, when cooling down at a field directly towards the quantum critical point, both g-factor and linewidth continuously decrease. Furthermore, we observe a very good matching of the g-factor behavior upon field-tuning and temperature-tuning towards the quantum-critical point. We analyze and discuss the results in the context of present theories on ESR in strongly correlated electron systems.
Read the full abstract

All authors Marc Scheffler, Jörg Sichelschmidt, Conrad Clauss, Mojtaba Javaheri Rahim, Boris I. Kochelaev, Cornelius Krellner, Christoph Geibel, Frank Steglich, Martin Dressel

Functional Analysis

The Endpoint Fractional Riesz Estimate on the Hamming Cube

Let $1<p<2$ and $D_j$ be the discrete partial derivative on the Hamming cube $\Omega_n = \{ -1, 1\}^n$. Let $\Delta=\sum_{j=1}^nD_j$ be the discrete Laplacian. We prove the endpoint inequality \[ \left\|\left(\sum_{j=1}^n|D_jf|^2\right)^{1/2}\right\|_{p} \lesssim_p\|\Delta^{1/p}f\|_{p}, \quad \forall f:\Omega_n \to \mathbb C. \] This result answers the conjecture proposed by Naor, Eskenazis and Ivanisvili (see [BenEfraimLustPiquard, IvanisviliVolberg] or [Remark 45, EskenazisIvanisvili]). The proof relies heavily on the noncommutative semigroup BMO theory [JungeMei].
Read the full abstract

Computation and Language

Alignment-Free Text-Audiobox for Voice Dubbing and Full-Duplex Dialogue Synthesis

We present Alignment-Free Text-Audiobox (Text-AB), a unified framework for high-quality voice dubbing and full-duplex dialogue synthesis. Building on a Diffusion Transformer trained with a flow-matching objective, Text-AB departs from the Audiobox system along three dimensions. First, it operates in a latent diffusion framework using DAC-VAE features that encode 48 kHz waveforms into a 25 Hz latent sequence, giving over 10x higher compression than previous EnCodec representations while improving resynthesis quality. Second, Text-AB is alignment-free: it consumes raw text via an off-the-shelf text encoder and learns text-speech alignment through cross-attention, removing the need for forced alignment and explicit duration prediction. Third, we scale model and data substantially, pretraining a 3B-parameter model on 480k hours of monolingual speech, followed by supervised fine-tuning on three downstream tasks: cross-lingual voice dubbing, full-duplex dialogue synthesis, and emotional full-duplex dialogue synthesis. At inference, Text-AB supports one-shot generation for up to ~1 min of speech and arbitrarily long-form generation via a multi-diffusion scheme, plus a multi-stage reranking strategy that enhances quality based on automated metrics. On a real-world dubbing benchmark, Text-AB delivers a step-change improvement over the latest internal dubbing system, with large gains in prosody similarity, voice similarity, naturalness, and shareability. For full-duplex dialogue synthesis, it approaches human recordings on short-form conversations and substantially outperforms the latest internal model on long-form human-likeness and expressivity, while natively modeling turn-taking, back-channeling, and emotional dynamics. For emotional dialogue synthesis, emotion conditioning significantly improves emotion alignment and emotional interaction quality over the unconditioned baseline.
Read the full abstract

All authors Sanyuan Chen, Min-Jae Hwang, Sho Inoue, Anna Sun, Bokai Yu, David Kant, Dongmin Hyun, Dorian Desblancs, Gregory Antonovsky, Oleg Repin, Peng-Jen Chen, Xutai Ma, Zehai Tu, Juan Pino, Wei-Ning Hsu

Soft Condensed Matter

Boundary- and Screening-Induced Bubbly Phases in Autophoretic Active Matter

Spatial confinement and chemical screening fundamentally reshape the non-equilibrium phase behavior of autophoretic active particles. Here, we present a systematic study mapping the collective dynamics of self-propelled particles governed by chemo-attractive translational forces ($\mu_t < 0$) and chemo-repulsive rotational torques ($\mu_r > 0$) across varying screening parameters $\kappa$, torque magnitudes $\mu_r$, and boundary condition coefficients $\Lambda^c$. Beyond standard chemotactic macro-phase separation and dynamic clustering, we report the emergence of novel boundary- and screening-induced bubbly phases, classified into boiling and bursting bubbles. Using a metric triad of steady-state cluster fraction $\langle S \rangle$, temporal fluctuation magnitude $\sigma_S$, and coordination number $\langle Q \rangle$, we draw phase diagrams to demarcate phases for no-flux boundaries ($\Lambda^c = 1$) and chemically permeable interfaces ($\Lambda^c = 0$) . Increasing chemical screening ($\kappa$) systematically suppresses long-range attraction, driving sequential phase transitions from macro-scale collapse toward bubbly states, dynamic micro-clusters, and homogeneous gas phases, while simultaneously inducing aggregate shape anisotropy. These findings provide predictive design rules for controlling active assembly and transport in microfluidic environments.
Read the full abstract