Peer Revenue Now accepting positions

Papers

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

All of arXiv

Monday 9 November

General Relativity and Quantum Cosmology

Thermodynamics of Kerr-Newman-Bertotti-Robinson black holes from background subtraction

The non-asymptotic flatness of the charged rotating black holes immersed in the Bertotti-Robinson magnetic field causes the ambiguity in the gauge choice for the time generator and the electric potential. By using the newly established form of the integrability condition together with the gauge choice for the asymptotically flat charged rotating black holes as the initial condition, we have found that the gauge can be uniquely fixed. With this fixed gauge, we further obtain the mass and angular momentum by the background subtraction method with the Minkowski spacetime as the reference spacetime. As a result, not only is the first law of black hole thermodynamics as well as the quantum statistical relation automatically satisfied, but also both the Christodoulou-Ruffini relation and the Smarr formula are obeyed.
Read the full abstract

All authors Chang-Xu Yan, Zili Mou, Yu Tian, Hongbao Zhang

Image and Video Processing

A four-step validation protocol for radiomic survival analysis in small cohorts: a body composition case study

Gastroenteropancreatic neuroendocrine tumors (GEP-NET) lack prognostic biomarkers, and computed tomography (CT)-based body composition radiomics at the third lumbar vertebra (L3) is a promising candidate. Small cohorts make high-dimensional radiomic models prone to overfitting and single-split evaluations variable and optimistic. This methodological study proposes a four-step validation protocol for radiomic survival analysis in small cohorts: (1) false discovery rate (FDR)-corrected feature selection, (2) a dual-path comparison of prior feature selection versus embedded regularization, (3) repeated nested cross-validation (CV) and (4) a negative control with Gaussian noise features. In 78 GEP-NET patients, 535 radiomic features from five L3 compartments were used to evaluate five survival models with Harrell's concordance index (C-index), time-dependent area under the curve (AUC) and integrated Brier score (IBS). Step 1 selected at least one feature in 18% of outer folds under noise (nominal level 20%) versus 84% with radiomic features. The dual-path comparison showed no consistent difference between strategies. Repeated nested CV reduced variability across seeds (41% median reduction in C-index standard deviation), but not enough to distinguish models. The nested CV C-index exceeded chance by 0.07-0.15 with radiomic features but not clearly under noise (0.40-0.42 with prior feature selection, 0.47-0.55 with embedded regularization). Frequently selected features tentatively linked greater subcutaneous fat, skeletal muscle compactness and texture to better prognosis. These findings suggest a modest prognostic signal in L3 radiomics, with considerable uncertainty that the protocol quantifies. The protocol may transfer to other small-cohort radiomic survival studies. Code: https://github.com/marielatalaveram/radiomics-survival-validation-protocol.
Read the full abstract

All authors Mariela Talavera-Muñoz, Cecilia Diana-Albelda, Miguel Antonio Sampedro-Nuñez, Victor Rodriguez-Laval, Jesus Bescos, Álvaro García-Martín

Computer Vision

National-Scale Crop Type Mapping in Morocco Using Google AlphaEarth Foundations Embeddings and the Fields of the World Field Boundary Map

Recent geospatial foundation models have introduced a new paradigm for large-scale Earth observation by providing pretrained embedding representations that can be directly exploited for downstream tasks. In this work, we investigate the use of Google AlphaEarth Foundations embeddings for national-scale crop type mapping in Morocco. The proposed workflow operates directly on publicly available 64-dimensional pretrained embeddings using lightweight downstream classifiers, eliminating the need to process raw satellite imagery during model training and inference. The approach targets three crop categories, namely winter cereals, tree crops, and other crops. For nationwide inference, the Fields of the World global field boundary map is used to constrain predictions to agricultural parcels, enabling efficient large-scale mapping. Experimental evaluation on an independent test set over ten repeated runs achieved an average classification accuracy of (87.17 +/- 0.42%). Nationwide inference over Morocco was completed in approximately five hours, with the dominant computational cost arising from cloud-based retrieval of the embeddings rather than model execution. These results demonstrate the potential of pretrained embedding representations for computationally efficient national-scale crop type mapping.
Read the full abstract

All authors Walid Elbarz, Mohamed Bourriz, Vincent Simonneaux, Hamd Ait Abdelali, François Bourzeix

Optimization and Control

Wolfe's Minimum-Norm-Point Method Generalizes the Simplex Method

Wolfe's method computes the point of minimum Euclidean norm in the convex hull of finitely many points. The method is parameterized by an insertion rule and it has been an open problem whether Wolfe's original rule guarantees a polynomial number of iterations (De Loera, Haddock, and Rademacher; STOC'18). We resolve this question by proving an exponential lower bound. In fact, we establish a tight connection to the simplex method for linear programming that allows us to transfer many lower bounds for simplex pivot rules to corresponding insertion rules of Wolfe's method. These include not only Wolfe's linopt rule via Dantzig's pivot rule, but also many other classical, randomized and history-based rules. As a corollary, we obtain the first exponential example for the Lawson-Hanson algorithm for nonnegative least squares, as well as PSPACE-hardness of deciding whether the linopt rule ever uses a given point.
Read the full abstract

Algebraic Geometry

Counter-examples to Conjectures on NCCRs via a Reductive McKay Correspondence

We disprove three conjectures appearing in Van den Bergh's plenary ICM survey, posed by De Deyn and Van den Bergh, which concern the interplay between non-commutative crepant resolutions (NCCRs) and the stringy Euler characteristic. Our counter-examples are applications of a new McKay correspondence for reductive groups that we prove in this paper, extending the classical McKay correspondence for finite groups. We obtain explicit formulas for the stringy $E$-functions of reductive quotients in terms of the Weyl group action and stringy $E$-functions of toric varieties; we also obtain an equivalent formula expressed as an integral of the age function over Weyl-fixed loci of a compact real torus. For symmetric representations, our expressions are related to arithmetic Tutte polynomials, yielding combinatorial formulas for the stringy Euler characteristic.
Read the full abstract

Robotics

Library of Hands: Fix the Hands, Not the World Model

A world model predicts how a scene changes after an action, and a planner uses it to search for actions that reach a goal. If the camera moves or the controller changes, the same command moves the robot differently: when its commands are turned by 90 degrees, a frozen planner with the LeWM world model that succeeds in 77.5% of episodes succeeds in only 7.8%. Current fixes retrain the model, need training data, or forget a set-up when it returns. We propose Library of Hands (LoH), a training-free method that keeps the world model frozen and corrects only the map from commands to applied actions, which we call the hand. LoH reads the hand through the frozen model, with no extra sensor: it finds the action that best explains the change between two camera images (a reading). It stores every hand it identifies, so a returning set-up is recognised rather than re-learnt. In closed-loop tests, LoH more than doubles success on three simulated tasks, from 28.6% to 62.3%, and loses only 0.6 percentage points when nothing changes. It also works, unchanged, with two other world models and on two of three models pretrained on real robot data, and its gain grows with the quality of its readings.
Read the full abstract

All authors Hasaan Ahmad, Qizhen Ying, Yuxuan Sun, Tianjin Huang, Guangming Wang

Statistical Methodology

Spatially varying quantile downscaling of summer temperature from atmospheric circulation and long-term trends over peninsular Spain

Over recent decades, high temperatures have intensified worldwide, with impacts on human health, ecosystems, and the economy. Understanding their spatial evolution requires accounting for long-term temperature changes and the role of atmospheric circulation. Quantile regression provides a flexible framework for characterizing changes across the temperature distribution, with relevance to extremes. We analyze the spatio-temporal evolution of the 0.50 and 0.95 quantiles of daily maximum temperature during summer (June–August) over peninsular Spain from 1960 to 2023. We develop a Bayesian spatially varying quantile regression model for observations from 40 meteorological stations, using ERA5 geopotential fields at tropospheric levels to characterize atmospheric circulation, together with long-term temporal trends capturing the background climate change signal. The spatially varying coefficients allow the sensitivity of temperature to atmospheric circulation to differ across the region, linking synoptic dynamics with local temperature variability. The model separates changes associated with the dynamical atmospheric state from the long-term thermodynamic warming signal, while allowing their effects to vary across the temperature distribution. Results show that atmospheric circulation and long-term trends provide comparable contributions to the observed warming. Significant increases in median and upper-quantile temperatures are found across much of peninsular Spain, with pronounced heterogeneity across months and space, followed by quantiles
Read the full abstract

Robotics

VLA Adaptation with Real and Simulated Data: A Case Study on Robotic Chess

Vision-language-action (VLA) models are trained as general-purpose foundation models for robot learning, and require fine-tuning for each downstream task and robot. This raises practical questions such as how much data and training updates are needed, how checkpoints should be selected, and whether simulation can be leveraged to alleviate the cost of real data collection. We study these questions on a task that requires both precise manipulation and spatial understanding of the scene. In this task, an SO-101 arm moves a chess king between two squares that are given only by their names, such as e4 and a5. We fine-tune $\pi_{0.5}$ with various dataset sizes and update budgets, using teleoperated demonstrations as well as simulated demonstrations from a randomized scripted expert that imitates the variability of human demonstrations. In simulation, the success rate scales with both data and updates. Importantly, we find that a short learning-rate cool-down adds up to 31 points and is necessary for comparing checkpoints, and that validation loss and action error are unreliable for model selection. On the real arm, we fine-tune on up to 256 teleoperated demonstrations and evaluate ten policies in 25 trials each. Partial results indicate that pretraining on the simulated demonstrations helps substantially, most clearly when real data is scarce. We also present a negative result, in which training on the policy's own rollouts with filtered behavior cloning, RECAP-style advantage conditioning, corrections from an automatic coach, or residual reinforcement learning did not yield a measurable improvement over plain behavior cloning in our setting. We release the simulator, the expert and the real demonstrations.
Read the full abstract

Artificial Intelligence

Interpretability Is Not Model-Specific

Interpretability tools are usually built for one model at a time: a refusal direction, a truthfulness probe, or a steering vector is found with labeled data inside the model it explains. We ask whether such tools can be carried to other models through a map between residual streams fit only on unlabeled prompts. We transfer a refusal direction, tested causally by addition and ablation, and two read-out tools, a correctness direction and a sentiment probe, across 110 pairs of models from five families and 1B to 72B parameters and between base and instruction-tuned models, with controls for specificity and five seeds. How much transfers depends on what the map was fit on. A map fit on ordinary questions recovers 54% of the target's own refusal steering and 16% to 77% of its correctness gain; adding unlabeled prompts from the region where the tool operates raises these to 102% and 88% to 98%, while control directions recover at most 1%. We prove that a linear map behaves this way whenever two models correspond linearly on the data it covers, and a label-free coverage score derived from this result tracks recovery across more than 16,000 runs and explains earlier reports that maps between models fail across domains. Carried read-out tools outperform anything the target's unlabeled representations yield. Interpretability results, once found, can be reused across models at the cost of a few unlabeled prompts.
Read the full abstract

All authors Marcus Armstrong, Navid Ayoobi, Pradham Mummaleti, Alexander Chulzhanov, Arjun Mukherjee

Computer Vision

What Breaks Scribble-Supervised Segmentation? An Audit-Informed Benchmark for Scribble Annotation Robustness

Scribble-supervised segmentation is typically evaluated with a fixed annotation set, leaving robustness to distinct semantic and geometric imperfections poorly understood. Motivated by an audit of 12,031 manual PASCAL Visual Object Classes (VOC) 2012 scribbles and prior work, we introduce six controlled perturbation operators. Class corruption, spatial offset, and directional perturbation directly operationalize audit signals. Stroke-length reduction is motivated by prior scribble studies, while random dropout and fragmentation serve as diagnostic controls. Each operator reports requested and achieved changes. We evaluate manual and automatic scribbles across two image domains, two principal backbone settings, and multiple training pipelines, with an additional MiT-B1 PFA anchor evaluation. In the five-seed Tree Energy Loss (TEL)-ResNet101 VOC study, mean intersection over union (mIoU) drops from 70.08 to 1.46 under class corruption, compared with 35.57 for spatial offset, 52.20 for directional perturbation, and 63.24 for stroke-length reduction at their strongest settings. At the evaluated endpoints, semantic-anchor corruption produces the largest degradation, while geometric and support responses depend on the pipeline and dataset. Our benchmark provides a reproducible protocol for diagnosing annotation robustness beyond clean-set accuracy. The operator code is available at https://github.com/Wmhiram99/scribble-operator-benchmark.
Read the full abstract

All authors Minheng Wu, Di Wen, Ruiping Liu, Junwei Zheng, Yufan Chen, Kailun Yang, Kunyu Peng

High Energy Physics (Phenomenology)

Amplitudes separation and strong-electromagnetic phases determination in the $\boldsymbol{J/\psi}$ decays into baryon-antibaryon pairs

We extract the relative phase between strong and electromagnetic amplitudes of the decays of a vector charmonium into spin-$1/2$ baryon-antibaryon pairs. Starting from a flavour-SU(3) effective Lagrangian with vector and tensor currents, we construct a parametrization in terms of partial waves coefficients. The electric and magnetic phases, the phase of the integrated complex overlap and the effective rate angle are derived from the same source amplitudes. We retain two possible solutions and both of them describe the scalar $\Sigma^+\bar\Sigma^-$ scan phase within its uncertainty, while presenting tension with several electric-magnetic production phases.
Read the full abstract

All authors Chiara Alice, Antonio Amoroso, Matteo Barbagiovanni, Monica Bertani, Diego Bettoni, Fabrizio Bianchi, Alberto Bortone, Damiano Cabiati, Alessandro Calcaterra, Gianluigi Cibinetto, Fabio Cossio, Marco Destefanis, Francesca De Mori, Emma Di Fiore, Luciano Fava, Sara Garbolino, Isabella Garzia, Michela Greco, Marco Maggiora, Simonetta Marcello, Federico Matias Melendi, Giulio Mezzadri, Piero Patteri, Simone Pacetti, Angelo Rivetti, Manuel Dionisio Da Rocha Rolo, Francesco Rosini, Marco Scodeggio, Stefano Sosio, Stefano Spataro, Umberto Zarantonello

Quantum Physics

Quantum Residual Exploration for Selected Configuration Interaction

We propose a modular hybrid quantum-classical method to improve selected configuration interaction at a fixed final determinant budget. Quantum measurements supply complementary determinants, accepted only when they outrank classical candidates under a common energetic criterion. This targets multistep excitations that perturbative discovery may miss in strongly correlated systems. A demonstration on the Quantinuum Reimei trapped-ion quantum computer shows that measured proposals for H$_{12}$ at a stretched geometry lower the variational energy relative to classical-only selection at the same final determinant budget.
Read the full abstract

All authors Kentaro Yamamoto, Ikko Hamamura, Juan W. Pedersen, Duncan Gowland, Gabriel Laude, Shintaro Maekawa, David Muñoz Ramo, Elica Kyoseva, William Dawson

Emerging Technologies

GPS-Triggered Adaptive Trajectory Control and SDR-Based RF Jamming for UAV Geofencing

This paper presents a GPS-triggered geofencing and enforcement scheme that integrates real-time satellite positioning with Software-Defined Radio (SDR) based RF jamming to detect and neutralize unauthorized unmanned aerial vehicle (UAV) boundary violations. Conventional geofencing depends on onboard GPS and cooperative flight firmware. The proposed approach instead moves enforcement to a ground base station, so the response does not depend on the UAV. An ESP32- WROOM-32 microcontroller paired with a u-blox NEO-6M GPS receiver streams telemetry over Wi-Fi to a Python/GNU Radio base station. There, a lightweight Haversine boundary-violation algorithm computes the great-circle distance between the UAV and the geofence. This distance drives a piecewise-linear adaptive speed and altitude control law. The law progressively decelerates and descends the UAV inside a 1 km warning zone and brings it to a complete halt at the boundary. If the boundary is breached, a HackRF One SDR escalates to band-limited noise jamming of the 2.4 GHz control link. The hardware stack was integrated on a GoolRC S116 quadrotor with an end-to-end enforcement latency of about 1 s. Field trials achieved 95% disconnection accuracy within a 50 m radius, and u-center NMEA analysis confirmed sub-2-meter GPS positioning precision. Simulation over an 8-sector, 50 km geofence model confirms speed and altitude-invariant boundary enforcement for cruising speeds of 5 to 50 m/s and cruising altitudes up to 100 ft. A comparison against a PID baseline shows that the proposed law guarantees overshoot-free shutdown at the boundary without gain tuning. It also runs in constant time, which makes it well suited to low-cost, resource-constrained embedded platforms.
Read the full abstract

Classical Analysis and ODEs

Global zero distribution and pluripolar exceptional cones for complex linear differential equations

We study the zeros of solutions of complex linear differential equations of arbitrary order $n\ge2$ with exponential-sum coefficients. Using the initial-value space $\mathbb{C}^n$ to parametrize the solutions, we show that solutions with initial vectors outside a complete pluripolar cone of Lebesgue measure zero share an explicit common leading term for the global zero count and a common joint limit in radius and angle. The rescaled radial distance from the disk boundary has an exponential limiting distribution, while the rescaled angular deviations from each main ray are Gaussian. Inspired by logarithmic tracts in the Eremenko-Lyubich class $\mathcal{B}$, we pull exterior sectors back conformally to countably many curvilinear strips extending in finitely many directions and obtain zero-counting asymptotics in each strip, although the coefficients may lie outside this class. Moreover, the asymptotic formula for the global zero count and the joint limit in radius and angle persist under arbitrary entire perturbations of strictly smaller weighted exponential type.
Read the full abstract

Strongly Correlated Electrons

Ground-state phase diagram of the spin-1/2 Heisenberg zigzag ladder with anisotropic $\Gamma$ term revisited

We study the spin-1/2 Heisenberg antiferromagnet with an anisotropic $\Gamma$ term on a zigzag ladder, which can be exactly mapped to the spin-1/2 XYZ zigzag ladder by a spin rotation. This model has recently attracted considerable interest as an effective model for the quasi-one-dimensional magnetic semiconductor YbCuS$_2$. For the model, the ground-state phase diagram was previously proposed by Saito and Hotta [Phys. Rev. Lett. 132, 166701 (2024); Phys. Rev. B 110, 024409 (2024)]. We systematically reexamine the phase diagram through complementary numerical and analytical approaches, including exact diagonalization, the density-matrix renormalization group method, and low-energy effective theories. We thereby demonstrate that when the anisotropic $\Gamma$ term is present, the model exhibits the Néel, dimer, ferromagnetic, and double-Néel phases, all of which have a finite excitation gap. The gapless Tomonaga-Luttinger-liquid (TLL) phase appears only in the isotropic case where the $\Gamma$ term is absent. We also confirm that the transitions between the Néel and dimer phases and between the Néel and ferromagnetic phases are continuous transitions governed by Gaussian criticality with central charge $c=1$. The TLL parameter characterizing the critical states at the Gaussian transitions is estimated numerically.
Read the full abstract

Machine Learning

Controlled Binary Assignment Flows

This paper introduces a geometric approach to the representation and learning of time-variant parametrizations of deep assignment flows, for binary data labelings on graphs. Each edge carries an optimally controlled parameter flow in a 2-dimensional totally geodesic submanifold, based on objectives derived from a non-local linear scale space representation of the input data. The aggregation of triplets of such flows in 5-dimensional totally geodesic submanifolds is learned from data. The fusion of these flows determines the trajectory in the high-dimensional symmetric parameter space and by geometric integration the layers of binary deep assignment flows. Numerical results demonstrate the expressivity and the limitation of our highly parallel geometric local synthesis approach to high-dimensional deep assignment flow parametrizations.
Read the full abstract

Numerical Analysis

Stochastic Multilevel Sampling for Inverse Problems via Coarse-Space Identifiability

Sampling only part of the data reduces the cost of each iteration in an inverse problem, but can make reconstruction updates unreliable. We study stochastic multilevel sampling, which computes updates on a coarse-to-fine hierarchy of correction spaces. The guiding idea is that a small sample may support a reliable coarse-grid update even when it provides insufficient information for a fine-grid update. For quadratic objectives, we identify conditions under which corrections computed from sampled data decrease the full-data objective. In small-batch experiments with computed tomography and periodic deblurring, the stochastic multilevel method improves reconstruction accuracy over matched single-level methods. In an increasingly ill-conditioned CT refinement with a fixed number of projection angles, the single-level error increases as the image dimension grows, whereas a deep hierarchy maintains or improves accuracy and outperforms full-data single-level optimization at the same number of forward and adjoint applications. The gains persist across the tested local solvers and strengthen with deeper hierarchies. A separate known-probe ptychography study provides complementary evidence for the benefit of hierarchical corrections for nonlinear inverse problems.
Read the full abstract

Systems and Control

Offset-Free Data-Driven Predictive Control: Implicit and Explicit Disturbance Formulations

Offset-free tracking is a key objective of any control algorithm. For predictive controllers with linear models, it is well understood that rejecting asymptotically-constant disturbances necessitates augmenting the prediction model with $p$ integrating states, where $p$ is the number of measured outputs. This can be achieved in model predictive control (MPC) by using a velocity form or a combination of disturbance model and observer. In data-driven predictive control (DPC), implicit incorporation of integrating states using the velocity form has been successfully adopted to obtain offset-free tracking; however, formulations based on an explicit disturbance model and observer have not yet been explored. In this paper, we first study how the integrating states are incorporated into the data-based representation used in DPC, translating the requirements of $p$ integrating states into the required modifications of the DPC formulation. We then propose a novel offset-free DPC formulation that incorporates an explicit constant output-disturbance model and an output disturbance observer. Numerical examples show that this formulation has similar disturbance-rejection capabilities to the velocity-form DPC, but using a smaller optimisation problem.
Read the full abstract

Machine Learning

Trade Promotion Optimisation for Consumer Goods

Trade promotions consume a large share of consumer-goods revenue, yet most planners still set discounts with a single, often unreliable, price elasticity per item. We present an end-to-end Trade Promotion Optimisation (TPO) system built as one chain: analogue SKUs, point elasticity, elasticity curve, price recommendation, promotional lift and net lift. Elasticity is measured at post-promo time from directional price-dip events against a rolling regular price, and at promo time by a causal estimator that residualises trend, seasonality and promotion, handles stock-out censoring with a Tobit likelihood and updates itself from the first clean promotion week. Items with weak evidence borrow the point elasticity and the curvature from analogue SKUs chosen by a cross-validated ridge model over category, brand, sub-category, price line and description. The optimiser separates self-price lift from substitution and post-promo dips and adjusts for inventory and trade funding. In a four-week promotion experiment on a 39-SKU consumer brand (mean elasticity -3.06), the plan lifts units by 20.6% and revenue by 6.9% at flat margin. Analogue pooling halves elasticity error relative to hierarchy averaging.
Read the full abstract

Chemical Physics

Damped polarizability and first hyperpolarizability of large systems with hybrid functionals: self-consistent and Lanczos solvers

We present a damped coupled-perturbed Kohn-Sham method to compute the dynamical polarizability and first hyperpolarizability of finite systems, implemented in the CRYSTAL code for semi-local, global and range-separated hybrid functionals. The approach is based on the adiabatic time-dependent density functional theory for the complex-valued (hyper)polarizabilities, describing optical phenomena such as absorption, and the nonlinear second-harmonic generation, rectification and Pockels effect. These are obtained either at individual frequencies with a self-consistent solver based on Pulay extrapolation, or at an arbitrary grid of frequency and damping factors with a Lanczos solver, from which the first-order amplitudes required in the hyperpolarizability can be reconstructed. With hybrid functionals, the self-consistent solver tends to become impractical at some point beyond the absorption onset, where there are many excitations for which the external frequency lies between their energy and those of their contributing single-particle transitions, and its convergence eventually stalls for large systems. In contrast, the convergence of the Lanczos method is essentially independent of size, hence the cost of the dynamical response calculation scales with system size as the hybrid ground-state calculation. The implementation is validated in a small molecule and, as an application, we study the first and second-order optical response of a series of hexagonal boron nitride flakes of up to 324 atoms. The use of range-separated hybrid functionals allows to capture the strongly-bound first excitation of the monolayer, and evidences the failure of the self-consistent method to converge excitations with strong electron-hole attraction.
Read the full abstract

All authors A. García-Blázquez, A. Erba, A. Boccuni, J. K. Desmarais

Optimization and Control

Multi-objective Markov Decision Process: Set-valued Bellman's principle and fixed point perspective

We study infinite-horizon discounted Markov decision processes with multiple cost objectives and finite state and action spaces. We consider the set of value functions generated by randomized Markov policies, where each function describes the expected discounted costs from all initial states under a single policy. We establish a set-valued Bellman equation and characterize this set as the greatest fixed point of the associated Bellman operator. We also show that choosing an attainable cost vector separately for each initial state can produce a function that no single Markov policy realizes. Nevertheless, the attainable costs at individual states satisfy their own Bellman recursion. These results provide a dynamic programming and fixed point framework for multi-objective Markov decision processes.
Read the full abstract

All authors Zachary Feinstein, Gabriela Kováčová, Ariel Neufeld, Birgit Rudloff

Machine Learning

Certify, Report, or Refuse: Harm-Constrained Conformal Prediction for Clinical Actions

Routing a melanoma case to routine follow-up is a serious action error that diagnostic coverage alone does not rule out. We present Harm-Constrained Risk-Optimal Conformal Prediction (HC-ROCP), a training-free framework that uses frozen classifier probabilities and adds a harm penalty when choosing prediction sets and actions. We test the resulting policies on disjoint calibration splits with multiplicity adjustment. HC-ROCP certifies supported claims, reports descriptive fallback policies, and refuses unsupported subgroup claims. We evaluate these outputs retrospectively on COVID-19 radiographs, ISIC 2019 dermoscopy, and EyePACS-to-APTOS retinopathy transfer. Across ten calibration seeds, HC-ROCP reduces COVID-19 severe-action errors from the exact risk-optimal baseline's 4.68% to 0.54%. At a 5% budget, global claims are certified for COVID-19 and ISIC, but the rare squamous-cell-carcinoma claim is refused. At 1%, all settings receive descriptive reports. For each budget, we also report the minimum severe-case count needed to pass the global harm gate with zero errors, providing a starting point for further data collection.
Read the full abstract

High Energy Physics (Phenomenology)

Pentagon Functions for Two-Loop Leading-Colour $pp\to t\bar{t}H$ Amplitudes with a Light-Quark Loop

We present a basis of special functions and an efficient numerical implementation for the planar two-loop master integrals that enter the leading-colour amplitudes for $pp\to t\bar{t}H$ with a closed light-quark loop. We provide analytic results for the special functions at transcendental weight one and two in terms of logarithms and dilogarithms, while higher weights are evaluated as one-fold integrals. We focus on the nested roots that arise, studying their analytic structure on the physical region and constructing numerically stable representations of letters and $\mathrm{d}\log$ forms that involve them. We find that a number of $\mathrm{d}\log$ forms have an antiderivative that cannot be represented by a single logarithm over the physical region and tackle them numerically with adaptive Chebyshev approximation. We implement the functions in C++ as an extension of the PentagonFunctions++ library. We study the performance of our implementation over physical phase-space points and find that the evaluation is accurate and fast enough for phenomenological application. This provides an important building block for NNLO studies of $t\bar{t}H$ production at the LHC.
Read the full abstract

High Energy Astrophysical Phenomena

GRB 241030A: A fallback-accreting magnetar description of the early X-ray emission

The bright, variable early X-ray emission of GRB~241030A at $z=1.411$ motivates investigating continued central-engine activity before the smoother afterglow emerges. We model a possible underlying magnetar contribution to the unabsorbed 0.3–10~keV XRT flux over 80–430~s, evolving the spin under fallback-disk and electromagnetic torques. A descriptive estimate gives $P_0\simeq4.25$~ms, $B\simeq3.04\times10^{15}$~G and source-frame $t_{\rm fb}\simeq143$~s, with exploratory temporal-block uncertainty ranges. The effective efficiency reaches its geometry-dependent ceiling of 81.2. At observer time 200~s, this realization has $P_*\simeq4.82$~ms and magnetospheric, corotation and light-cylinder radii $r_{\rm m}\simeq235$~km, $r_{\rm co}\simeq48$~km and $r_{\rm LC}\simeq230$~km. It is on the ejector branch; light-cylinder entry near 255~s initiates negative propeller torque and steepens the decline. The smooth component declines throughout the fitted interval; additional variable emission contributes to the observed maximum near 180~s. A correlated-residual Bayesian analysis permits a fainter component and broader dynamical histories, with median $P_0=16.8$~ms and central 68 per cent interval 2.76–286~ms. However, posterior predictive checks reveal that this statistical description typically underpredicts the observed flux level and variability, limiting its physical interpretation. Fallback-modified magnetar activity remains a possible descriptive scenario, but these analyses do not establish a unique underlying component or robust constraints on the initial spin and disk-interaction history. We also discuss Lense-Thirring precession and alternative mechanisms as possible origins of the early X-ray variability.
Read the full abstract

Numerical Analysis

Shape Holomorphy of the Calderón Projector on Lipschitz Surfaces: Frequency-Explicit Bounds for the Helmholtz Operator

We establish the holomorphic dependence of the Calderón projector for the Helmholtz operator upon shape boundary deformations of a family of bounded Lipschitz domains in two and three spatial dimensions, with bounds explicit in the wavenumber and the size of the holomorphic extension. More precisely, after transporting the four boundary integral operators contained in the Calderón projector to a fixed reference boundary, we prove that the resulting parameter-to-operator maps admit bounded holomorphic extensions in their natural fractional Sobolev spaces. Unlike approaches that require boundary regularity of class $\mathscr C^{1,1}$ or higher, our argument traces the holomorphic dependence on the shape by transferring elliptic estimates to the boundary through the fundamental solution and the Dirichlet and conormal trace operators. This construction is the key step that allows us to treat Lipschitz-regular boundary deformations. Furthermore, we show that the domain of holomorphy can be chosen independently of the wavenumber $\kappa$, while the operator bounds in the corresponding fractional Sobolev spaces contain algebraic factors in $\kappa$ multiplied by $\exp(C\kappa\delta)$. In particular, for real domains, corresponding to $\delta=0$, our single layer estimate recovers, to our knowledge, the best available frequency dependence on general Lipschitz boundaries. For the double layer, adjoint double layer, and hypersingular operators, we obtain novel frequency-explicit bounds. The results proved in this work provide the analytic foundation for frequency-explicit uncertainty quantification, model order reduction, and high-dimensional approximation of boundary integral operators, their Galerkin discretizations, and the resulting boundary densities and quantities of interest on families of parametrically varying boundaries.
Read the full abstract

Strongly Correlated Electrons

Higher tensor network states in two spatial dimensions: fundamental theorem for open boundary conditions

We introduce a class of two-dimensional tensor network states obtained by categorifying matrix product states. This procedure leads us to consider bond spaces provided by a 2-vector space associated with the 2-cells of the underlying lattice. Tensor network representations are then defined as module categories over the free monoidal category on a monoidal signature encoding the lattice geometry. We prove a fundamental theorem stating that two representations generating the same states with suitable open boundary conditions are equivalent as module categories. Finally, we introduce a notion of 2-injectivity, a categorified analogue of the injectivity condition for matrix product states.
Read the full abstract

Computer Vision

When Does Sliced Inference Pay Off? Resolution, Tiling and Dense Fusion for Small Object Detection

Sliced inference such as SAHI is the standard remedy for small objects in high-resolution imagery, but its gains are usually measured against low-resolution baselines. We revisit sliced inference with scale as the explicit variable and ask when tiling actually pays off. We compare nine dense detectors, eight YOLO models of four generations and FCOS, on VisDrone and DOTA, under pre-specified hypotheses, calibration-only tuning and a single evaluation on held-out test data. Two findings emerge. First, much of the benefit attributed to slicing is a resolution effect: on VisDrone, a single pass trained at a higher resolution outperforms tiling with box fusion (merging decoded boxes) for every detector that could be trained at that resolution, and at equal compute a higher resolution helps more than a larger model. Second, tiles carry complementary information that box fusion wastes. We introduce SCAHI (Slicing with Coverage-Aware Hyper Inference), which accumulates the dense outputs of all passes with weights that give every image position the same total coverage, and reads objects off a peak grid fine enough to keep crowded neighbours apart. On identical passes, SCAHI outperforms box fusion for every YOLO detector on both datasets and matches it for FCOS. On DOTA, whose images far exceed the resolution a single pass can be trained at, tiles with SCAHI beat the best single pass for every detector, by 4.3 to 5.6 mAP50 at 2048 px, and replacing only SAHI's merge step by SCAHI adds 2.3 to 4.0 mAP50. An ablation attributes the gain to the border weighting and the resolution of the peak grid. Source code is available at https://github.com/beckerio/scahi.
Read the full abstract

All authors Stefan Becker, Jens Bayer, David Münch, Wolfgang Hübner

Differential Geometry

On Continuum Theories in a Locally Conformal Setting

We study continuum and kinetic dynamics on locally conformally symplectic (LCS) and locally conformally cosymplectic (LCCos) manifolds. The main geometric tool is the Lichnerowicz-de Rham differential, which allows local geometric data that are compatible only up to conformal rescaling to glue into global structures. In the LCS setting, we analyze the locally conformal Vlasov equation from the viewpoint of continuity. Its source term does not represent particle creation or annihilation: it measures the mismatch between the ordinary phase-space volume and the conformally weighted volume determined by the Lee form. Absorbing this weight into the density yields a source-free continuity equation. We also compute the entropy production of the transported density, and illustrate the framework through a thermostatted ensemble, an LCS-driven Vlasov model with a momentum-dependent gyroscopic force, and their combination. We then propose a new definition of LCCos manifolds in which the defining one-form and two-form carry the same conformal weight. Unlike the usual convention, this allows local cosymplectic Hamiltonian systems to glue directly into global LCCos Hamiltonian dynamics, and it normalizes the evolution vector field by the conformal unit rather than by the ordinary unit. We develop the corresponding Hamiltonian, evolution and continuum dynamics, and illustrate them with a time-dependent mass example exhibiting transport in the distinguished time direction, momentum-space drift, and conformal source terms.
Read the full abstract

All authors Begüm Ateşli, Oğul Esen, Miroslav Grmela, Michal Pavelka

Earth and Planetary Astrophysics

Testing Models of Disk Formation and Gravito-Viscous Evolution Against Observations of Young Protoplanetary Disks

Young protoplanetary disks have received increasing attention in recent years, due to observations revealing the presence of substructures in disks younger than one million years, suggesting that planet formation may begin earlier than previously thought. Other recent observations have provided new insights into the evolution of gas and dust in these early stages, enabling comparisons with evolved Class II disks. Yet the physical processes governing the earliest stages of disk evolution remain poorly understood. In this paper, we present two-dimensional thin-disk hydrodynamical simulations of the formation and early evolution of protoplanetary disks, under the influence of self-gravity and turbulent viscosity alone. We explore various properties of disks forming from the collapse of clouds with a range of initial masses and we compare our results with surveys of protoplanetary disks across multiple evolutionary stages. Our simulations broadly reproduce key observational constraints, including the mm fluxes of Class 0 disks and the dust disk radii of later stages. However, our models do not explain the observed gas and dust disk masses of Class I and II disks: while they account for the most massive observed disks, they fail to reproduce the least massive ones. We discuss that simulations incorporating the influence of the star-forming environment might be needed to resolve this discrepancy.
Read the full abstract

All authors Fabiola Antonietta Gerosa, Paola Pinilla, Eduard I. Vorobyov, James McKevitt

Robotics

Correctness Boundaries of Typed Plans for Natural-Language Robot Fleet Control

We examine what plan validation and execution reports establish in natural-language robot fleet control. Commands are traced through typed plans, deterministic corrections, fleet coordination and outcome checks in a three-level Robot Operating System 2 implementation with twenty simulated robots. An offline benchmark of 44 cases checks plan structure and selected command requirements. Across five local configurations without a response schema, 70.0-87.3% of responses pass. All runs in a selected 18-run simulation series meet their pose-based task conditions. This series includes two successful repeats that replaced incomplete runs after a software change. Saved plans and logs document duplicate pickup assignments and goal adjustments within the original task zones. Separate formation traces show a structurally accepted plan with the wrong formation shape and a rejected route, and an arrival report that omitted a requested robot after replanning. Carry completion records pickup and subsequent drop-off zone presence for the same robot. The cases distinguish changes to executable actions from changes to the requirements covered by a completion report. They motivate checking the effective assignments, goals and reported outcome against the original command.
Read the full abstract

All authors Artem Tuzov, Dmitrii Vizitei, Evgenii Shlomov, Senya Belykh, Anastasia Malakhova, Alexander Maloletov