Systems and Control
Reward-Rate Congestion Games and Replicator–Dinkelbach Dynamics
Hassan Abdelraouf, Vaibhav Srivastava and Vijay Gupta
Reward rate is a key performance criterion in cyber-physical and robotic systems where time, workload, and coordination costs are limiting resources. We introduce reward-rate congestion games, where
agents seek to maximize reward per unit execution time. The direct reward-rate game is generally not an exact potential game. We develop a Dinkelbach-based framework in which, for every fixed Dinkelbach parameter, the transformed game is an exact potential game. This yields a potential-level Dinkelbach iteration that terminates finitely at the optimal potential reward rate when the inner potential maximization problem is solved globally. We also provide a sufficient condition under which an
equilibrium of the transformed game is an
equilibrium of the original reward-rate game. To optimize aggregate performance, we introduce marginal externality corrections that make the corrected potential coincide with the Dinkelbach-transformed social reward-rate objective, thereby enabling optimization of the social reward rate. Finally, we develop a continuous-time replicator–Dinkelbach dynamics for reward-rate population games coupling fast replicator dynamics with a slow reward-rate update. We establish convergence of the fixed-parameter replicator dynamics, global asymptotic and local exponential stability of the reduced Dinkelbach dynamics, and local exponential stability of the coupled system for sufficiently slow Dinkelbach updates. The framework is illustrated on a continuous task-allocation problem.
Read the full abstract ▾
Computer Vision
ODPure: Backdoor Purification for Object Detection via Ensemble Corruption Consensus
Li Zeng, Mingcheng Duan, Longfei Fan et al. (+5 more)
With the development of applications like autonomous driving, object detection has gained significant
attention, while also highlighting critical vulnerabilities like
backdoor attacks that severely compromise model integrity. Specifically, such attacks involve altering the categories of objects (i.e., object misclassification), removing bounding boxes (i.e., object disappearance), or generating bounding box proposals for non-existent objects (i.e., object generation) when a predefined trigger is present in the input. Although
backdoor defenses for image classification are well-established, the research for object detection remains comparatively underexplored. Existing defenses address these threats by scanning outputs or models for potential
backdoors but require discarding either malicious data or models. This remedy fails to enable a continuous and accurate perceptual stream for the object detection pipeline. To address such limitations, we propose ODPure, a novel input-stage black-box defense for object detection, which is based on input purification that ensures stable perception flows. Tailored to the dense prediction nature of object detectors, our Corruption-Reconstruction-Selection (CRS) paradigm operates by neutralizing triggers through a diverse portfolio of corruptions to generate a massive pool of redundant proposals, then recovering fine-grained structural cues via generative priors, and finally employing voting to reach a consensus on the resulting detections. Comprehensive experiments demonstrate that our method provides robust defense against diverse
backdoor attacks and trigger types while preserving baseline accuracy. Our code is available at https://github.com/Alex66366/ODPure.
Read the full abstract ▾
All authors Li Zeng, Mingcheng Duan, Longfei Fan, Hangtao Zhang, Xianlong Wang, Yanchun Li, Xia Wen, Leo Yu Zhang
Optimization and Control
Second-Order Stationarity with Common Random Losses: Matching Tolerance Bounds
Wendao Wu, Haihan Zhang, Chenheng Zhang et al. (+5 more)
We establish tight
polynomial tolerance bounds for
stochastic second-order stationarity when each fresh oracle response is a derivative of one common random scalar loss. For a population objective $F$ with Lipschitz gradient and Hessian, the target is $\|\nabla F(x)\|\le \epsilon$ and $\lambda_{\min}(\nabla^2 F(x))\ge -\gamma$, with independent tolerances $\epsilon,\gamma>0$. Under bounded gradient variance and almost-surely bounded Hessian error, the minimax number of fresh gradient or Hessian-vector-product calls is $\widetilde{\Theta}\!\left(\epsilon^{-3}+\gamma^{-5}\right)$. The characterization fixes positive gap, smoothness, and noise parameters, suppresses logarithmic factors, and allows dimension to grow within an explicit
polynomial envelope. The upper bound removes the mixed term $\epsilon^{-2}\gamma^{-2}$ from the earlier fresh-HVP guarantee. Direct random-line Hessian estimates and a dyadic gradient tracker separate gradient drift from randomly signed
curvature motion. The lower bound realizes the endpoint costs through globally defined smooth random losses: a smooth partition localizes scalar noise without a chain-length penalty, while exact cancellation limits the information in the entire response. Consequently, the same tolerance exponents hold even for joint value, gradient, and full-Hessian responses with bounded value variance. For a population Hessian with Hölder exponent $\nu\in(0,1]$, fresh gradient/HVP complexity becomes $\widetilde{\Theta}\!\left(\epsilon^{-3}+\gamma^{-(3+2/\nu)}\right)$ under the corresponding dimension envelope.
Read the full abstract ▾
All authors Wendao Wu, Haihan Zhang, Chenheng Zhang, Yanyi Li, Chunyuan Zheng, Cong Fang, Haoxuan Li, Zhouchen Lin
Computational Physics
Optimal limits on weak integrability breaking and protected thermal memory near qutrit exchange
Hanbing Liang and Fujun Liu
Although integrability does not universally require a continuous one-site
symmetry, we rigorously prove that every jointly analytic, regular Yang-Baxter deformation of the qutrit exchange interaction necessarily retains a nontrivial, analytically varying one-site charge. Breaking this local
symmetry imposes a fundamental physical constraint on approximate conservation, governed by the optimal uniform bound $\delta^3 \le C\varepsilon$ that explicitly relates the minimal one-site
symmetry defect $\delta$ to the local current-conservation residual $\varepsilon$. While breaking all one-site charges strictly forbids an exact integrable completion, an optimally compensated nearest-neighbor interaction saturates this cubic limit and anomalously extends the guaranteed infinite-temperature energy-current correlation window to order $|\lambda|^{-3}$ in the perturbation strength $\lambda$. Furthermore, we reveal a fundamental resonance obstruction for intrinsic conversion perturbations that strictly prevents any exact first-order repair of a broken one-site charge on any finite ring. Nevertheless, we demonstrate that the complete eight-dimensional charge
memory matrix remains thermodynamically protected and approaches the identity for timescales $t=o(|\lambda|^{-3/2})$, a robust feature of the full infinite-temperature dynamics when the thermodynamic limit is taken before weak coupling.
Read the full abstract ▾
Computer Vision
EmbodiedMemory-Bench: Benchmarking Embodied Memory for Long-Horizon Embodied Tasks
Lizhou Liang, Xinyu Zhong, Miao Pan et al. (+7 more)
Long-horizon
embodied interaction requires
agents to retain and continually update information about the environment as they observe, act, and encounter change. Yet current
agents struggle to maintain such
memory reliably. Our analysis traces this limitation to four key deficiencies: weak fine-grained visual
memory, unreliable dynamic world-state tracking, failing to record world state revealed by interaction outcomes, and limited generalization from prior experience. However, existing
benchmarks do not directly assess these
memory capabilities during long-horizon
embodied interaction. To address this gap, we introduce EmbodiedMemory-Bench (EMem-Bench), comprising 2,554 interactive episodes across four task families. EMem-Bench requires
agents to build and update
memory from interaction history, then use it to complete a later task by acting in the environment. We further present
Embodied-Memorizer (EMem), an external
memory system that organizes
embodied experience into spatial, event, and scene
memories. We also train EMem-8B, an 8B policy that manages and uses these
memories. We evaluate a diverse range of open-source and proprietary MLLMs and representative
multimodal memory systems. Results show that current models remain weak and uneven across the four challenges. Under matched backbones, EMem achieves the best overall performance among the evaluated
memory systems and improves both open-source and proprietary models, while EMem-8B further improves over its backbone. Project page: https://zju-omniai.github.io/EmbodiedMemoryBench/
Read the full abstract ▾
All authors Lizhou Liang, Xinyu Zhong, Miao Pan, Xiaohe Zhou, Xuanyu Liu, Qinfeng Li, Peng Li, Jintao Chen, Xuhong Zhang, Wenqi Zhang
Computer Vision
Diff-RF: Mutually Reinforced Image Registration and Fusion via Degradation-Aware Learning
Xunpeng Yi, Zaixi Du, Qinglong Yan et al. (+3 more)
Image registration and fusion aim to establish spatial correspondences from misaligned multi-modal source images, and integrate complementary information. However, in real-world imaging scenarios, source images are often affected by complex and diverse degradations, such as low illumination, noise, etc., which severely hinder the effectiveness of registration and fusion. To address this issue, we propose a mutually reinforced image registration and fusion
diffusion framework via degradation-aware learning, termed Diff-RF. It explores the intrinsic coupling between registration-fusion and information restoration in the degradation conditions, enabling high-quality fusion of unregistered images under complex degradation conditions. First, the intra-modal restoration module is designed to alleviate modality-specific degradations by leveraging information within each modality, thereby providing more reliable structural representations for registration and facilitating subsequent cross-modal fusion. Second, we develop a cross-modal
diffusion registration and fusion module that establishes bidirectional interaction between registration and fusion. By integrating fusion-derived visual cues and correspondence-based
geometric conditions into the
diffusion process, the proposed framework progressively refines spatial
alignment and exploits cross-modal complementary information to achieve collaborative enhancement. Rather than treating them as independent components, degradation-aware information restoration and the collaborative optimization of registration and fusion are tightly coupled, achieving overall performance improvements. Extensive experiments on multiple extended
datasets demonstrate that Diff-RF achieves superior registration accuracy and fusion quality under various degraded scenarios, exhibiting strong robustness and generalization ability.
Read the full abstract ▾
All authors Xunpeng Yi, Zaixi Du, Qinglong Yan, Yibing Zhang, Han Xu, Jiayi Ma
Probability
Stability under mixtures of transportation inequalities and restricted log-Sobolev inequalities
Radosław Adamczak, Dominik Kutek and Michał Strzelecki
We prove results about stability under mixtures for transport–
entropy inequalities and restricted functional inequalities of Poincaré and log-Sobolev type. We treat Talagrand inequalities of type $\mathbf{T}_2$ and $\mathbf{T}_1$ together with their generalizations, as well as weak transportation inequalities introduced by Gozlan et al. Our main results provide mild sufficient conditions under which a given inequality may be transferred from mixture components to the mixture itself.
Read the full abstract ▾
Quantum Physics
Certifying fermionic Gaussian states (and a little more) with optimal precision dependence
Ninnat Dangniam, Laphas Premcharoen, Metrasit Sripech et al. (+1 more)
Fermionic Gaussian states are the workhorse reference states for
qubit-based
quantum simulation of fermionic systems, yet existing certification protocols either proceed via fidelity estimation, leading to suboptimal sample complexity in the target precision $\epsilon$, only apply to Haar-typical states, or require adaptive measurements on a large fraction of
qubits. We give an adaptive protocol that certifies any $d$-mode pure fermionic Gaussian state using $O(d^2\epsilon^{-1})$ copies, single-
qubit measurements with only one
qubit measured adaptively per copy, and $O(d^3)$ classical processing time per copy. The $\epsilon^{-1}$ dependence is optimal for the certification problem, even among strategies using entangled measurements. The sample complexity is controlled by the spectral gap of a $d\leftrightarrow d-2$ down-up walk—a Markov chain studied in the theory of high-dimensional expanders, equivalent to a two-site Glauber dynamics in statistical physics—relating certification efficiency to relaxation time to
equilibrium. The worst-case bound is tight for this protocol and is attained by physically relevant states, including ground states of the fully dimerized Su-Schrieffer-Heeger (SSH) chain and BCS pair states. In contrast, numerics up to $d=14$ suggest that $O(d \epsilon^{-1})$ copies suffice to certify Haar-typical Gaussian states, a factor of $d$ improvement over the worst-case bound. Finally, because the bound depends only on the target's computational-basis distribution, the protocol extends to efficiently phase-dressed states—states obtained by injecting an efficiently computable diagonal phase to Gaussian states—with the same sample complexity. The class includes a continuous family of four-mode non-Gaussian magic states for matchgate computation.
Read the full abstract ▾
All authors Ninnat Dangniam, Laphas Premcharoen, Metrasit Sripech, Thiparat Chotibut
Medical Physics
Non-circular scan trajectories for reducing cone-beam artifacts in Gamma Knife CBCT images: a simulation study
Alexandra Alain-Beaudoin, Håkan Nordström, Luc Beaulieu et al. (+1 more)
Objective. Gamma Knife cone-beam computed
tomography (CBCT) images are deteriorated by cone-beam artifacts whose magnitude increases along the superior direction. In this study, a novel scan trajectory compatible with the Gamma Knife CBCT system is optimized to reduce cone-beam artifacts, with the aim to replace the current 200-degree single-arc scan. Approach. Data sampling analysis with tomographic incompleteness maps indicates the level of undersampling across the field of view based on the geometry of the system and of a scan trajectory. Moreover, CBCT simulations are performed from a virtual phantom with disks aligned along the axial direction and from a CT reconstruction of a stereotactic end-to-end validation (STEEV) phantom. CBCT projections are simulated for a given scan trajectory through a polychromatic forward projection model with added noise and scatter, then the CBCT image is reconstructed using an iterative
algorithm which minimizes weighted least squares. Main results. Both the incompleteness analysis and the CBCT simulations indicate adding lines to the current single-arc trajectory is more efficient to reduce cone-beam artifacts than adding more arcs, both in terms of number of additional projections and new artifacts. A line-arc-line trajectory with source axial steps of 3.5 mm removes virtually all cone-beam artifacts. The widths of the cone-beam artifacts created by the disks show a positive correlation between the artifact magnitude and the incompleteness value. Significance. A line-arc-line scan trajectory is promising to reduce cone-beam artifacts of the Gamma Knife CBCT images while being a compatible and reasonable trajectory for the current system design.
Read the full abstract ▾
All authors Alexandra Alain-Beaudoin, Håkan Nordström, Luc Beaulieu, Joakim da Silva
Computer Vision
Do Center Biases Propagate? Robustness of Pathology Foundation Models in Whole-Slide Image Classification
Ilán Carretero, Pablo Meseguer, Rocío del Amor et al. (+1 more)
Pathology foundation models (PFMs) have transformed computational pathology through powerful representation learning from histopathological images. PFMs provide rich, discriminative representations for whole slide image (WSI) analysis, enabling tasks such as slide-level classification under multiple instance learning (MIL). However, these representations may also encode non-biological signals associated with acquisition centers, potentially introducing spurious shortcuts into downstream predictions. In this work, we evaluate center-associated robustness in WSI classification using a controlled training setting with increasing class-center correlations quantified by Cramér's V. We
benchmark six PFMs across four
datasets and two MIL aggregators, while evaluating ComBat as a robustification strategy. We further introduce the Area Under the Cramér's V Curve (AUCC) to jointly capture absolute classification performance and its degradation as spurious correlation increases. Results show that center-related information encoded by PFMs propagates to WSI-level predictions, with robustness depending on both the PFM representation and MIL aggregation strategy. Additionally, ComBat harmonization does not provide consistent robustness gains across
datasets.
Read the full abstract ▾
All authors Ilán Carretero, Pablo Meseguer, Rocío del Amor, Valery Naranjo
Computer Vision
A Unified Framework and Dataset for Oriented Object Visual Grounding in Remote Sensing
Zeyu Ding, Yong Zhou, Jiaqi Zhao et al. (+5 more)
Visual grounding in remote sensing images aims to locate objects described by referring expressions. Most existing methods predict horizontal bounding boxes, which are often inaccurate for objects with arbitrary orientations. To address this limitation, we introduce O$^2$-VG, a family of models for oriented object visual grounding with three complementary designs. Specifically, O$^2$-VG-Trans is a cross-modality
transformer for oriented object visual grounding. It establishes a strong discriminative foundation for the model family. Building upon it, O$^2$-VG-Uni predicts universal oriented proposals for possible foreground objects without specific text
prompts. It also supports object
retrieval through cached proposal
embeddings. Using these universal oriented proposals as input
prompts, O$^2$-VG-VLM is an autoregressive
vision-
language model. It generates oriented box
token blocks in parallel through multi-
token prediction. In addition, we construct DIOR-R-RSVG, a
dataset for oriented object visual grounding in remote sensing images. It provides image, expression, and oriented box triplets for training and evaluation. Together, the O$^2$-VG family provides a flexible framework that spans discriminative
transformers and generative
vision-
language models. It achieves superior performance across multiple
benchmarks. Code is available at https://github.com/wokaikaixinxin/ai4rs and https://github.com/wokaikaixinxin/Eagle_o2_vg.
Read the full abstract ▾
All authors Zeyu Ding, Yong Zhou, Jiaqi Zhao, Wen-Liang Du, Xixi Li, Hancheng Zhu, Rui Yao, Abdulmotaleb El Saddik
Quantum Physics
Distilling Datasets into Shallow Circuits for Quantum Machine Learning
Guang Lin and Qibin Zhao
In
quantum machine learning, training a
quantum model requires each sample to be prepared as a
quantum state by a loading
circuit that must be re-executed for every shot at every training step. The total burden therefore scales with both the number of samples and the cost of preparation. Existing approaches reduce the loading cost of individual inputs or distill data and compress their representations before applying a separate
quantum encoding. This separation can leave resulting samples costly to prepare or cause additional loss of the information during subsequent compilation into shallow
circuits. We propose
quantum dataset distillation (QDD), which distills the full
dataset directly into a small set of shallow
circuits. Each synthetic sample is parameterized as a staircase
circuit corresponding to a low-rank tensor network, making the sample and its loading
circuit the same object. The
circuit parameters are optimized classically with exact gradients using distribution matching under an explicit loading budget, without post-hoc state-preparation synthesis. On MNIST and Fashion-MNIST, only 10
circuits per class ($0.0017\times$ the full
dataset size) achieve accuracy comparable to full-data training and outperform selection baselines under matched sample and loading budgets. In finite-shot training, QDD reaches 95% of the full-data accuracy with more than $100\times$ fewer cumulative shots. Additionally, we validate QDD on real
quantum hardware, demonstrating its practical deployment potential.
Read the full abstract ▾
Cryptography and Security
MimicSat: A Reconfigurable Cyber-Physical Testbed For Small Satellite Systems and Cybersecurity Research
Nisha Vinayaga-Sureshkanth, A H M Nazmus Sakib, Mahsin Bin Akram et al. (+2 more)
MimicSat provides a common experimental environment for examining how changes in satellite subsystem behavior propagate to mission outcomes across software-based and hardware-based execution. Its design is motivated by controlled spacecraft cybersecurity studies involving attacks, faults, and defensive responses. In MimicSat, a mission encompasses the spacecraft and ground activities required to achieve defined objectives, and an experiment consists of one or more mission runs used to study selected conditions or interventions. To support such studies, MimicSat offers software-based and hardware-based execution environments that implement the same mission functions and data exchanges, while allowing specific functions to be realized differently. A shared mission definition preserves command and telemetry semantics across environments, and collected observations retain provenance about the originating participants and acquisition paths. As a result, mission behavior can be compared across execution configurations without redefining the surrounding mission. This paper presents the architectural principles of MimicSat, its software and hardware execution forms, and their integrated operation. MimicSat also supports satellite systems engineering, mission operations, resilience studies, and related experimental use cases.
Read the full abstract ▾
All authors Nisha Vinayaga-Sureshkanth, A H M Nazmus Sakib, Mahsin Bin Akram, David R. Silva, Murtuza Jadliwala
Analysis of PDEs
The Bifurcation Phenomenon for the Regularized Two-Phase Problem Associated with the $p$-Laplacian
Alaa Haj Ali, Catherine Lebiedzik and Peiyong Wang
In this paper, we verify a bifurcation phenomenon regarding the multiplicity of weak solutions, subject to the Dirichlet boundary condition, of a regularized two-phase free boundary problem associated with the $p$-
Laplacian. In fact, we prove the existence of a mountain pass solution when the boundary data is small. In this case, three weak solutions of the regularized problem exist: the $p$-harmonic function, a minimizer of the corresponding functional, and the mountain pass solution. Finally, we consider the special case of radially symmetric solutions. By solving the associated nonlinear ordinary differential equation in the unregularized case, we show explicitly how the Bernoulli condition at the free boundary gives rise to the bifurcation of solutions.
Read the full abstract ▾
Cryptography and Security
Pinpointing Super-Quadratic Quantum Enumeration Speedups: Exact and Certified Evaluation of the Guessing-Moment Exponent under Product-Distribution Advice
Carsten Schubert, Niklas Paskarbeit, Maximilian J. Kramer et al. (+2 more)
Grover's
algorithm gives an optimal quadratic query advantage for black-box search. In cryptanalysis, however, the search often comes with additional probabilistic advice over the candidates, frequently of product form, e.g. from side-channel leakage on independent key coordinates. Classically, guessing in likelihood order is optimal in expectation. In the
quantum setting, Montanaro showed how to achieve an optimal expected query complexity, beating plain Grover on every non-uniform advice distribution (up to a constant overhead factor). What has been missing so far is a finite-size method for evaluating the
quantum-classical guessing-moment separation induced by a given advice distribution. We provide such a method for product-distribution advice, thereby sharpening the previous
entropy-based estimate of Bashiri et al. We reduce the classical and
quantum guessing moments to functionals of the one-dimensional surprisal distribution, obtained for product advice by convolving the per-coordinate surprisal laws. When the surprisals lie on a common arithmetic grid (the commensurate case), the logarithmic moments and hence the speedup exponent can be evaluated as finite sums without discretization error; exponential tilting makes this computation numerically stable. For general product advice, we discretize the surprisals onto a common grid and derive an a-posteriori bound on the resulting binning error. We apply the framework to cold-boot leakage on seeds and block-cipher keys, to template-attack posteriors, and to synthetic i.i.d. Bernoulli posteriors calibrated to average residual ranks reported for Keccak side-channel attacks on ML-KEM and ML-DSA. The resulting exponents substantially exceed 2 in several skewed-advice settings, reaching 3.92 for cold-boot advice and 3.12 in the synthetic models, and include cases where the previous
entropy-based bound did not establish an exponent above 2.
Read the full abstract ▾
All authors Carsten Schubert, Niklas Paskarbeit, Maximilian J. Kramer, Jean-Pierre Seifert, Marian Margraf
Robotics
Large-Scale Geometric Map-Based Localization of UAVs in GNSS-Denied Urban Environments
Garth Terlizzi and Kaveh Fathian
Unmanned aerial vehicles (
UAVs) operating in GNSS-denied urban environments require alternative methods for position estimation. Existing approaches based on satellite image
retrieval or learned descriptors are sensitive to appearance variation and degrade rapidly as the search area grows. We present a
vision-based localization system that matches building patterns observed from a downward-facing
UAV camera against a reference building footprint database. Our approach detects buildings in aerial imagery, accumulates observations across frames into a unified map, and matches local building arrangements against reference footprints using a novel geometry-driven descriptor that augments local triangle structure with per-building shape features. By encoding spatial relationships between nearby buildings rather than visual appearance, the system is robust to appearance variations and remains discriminative over large search areas. Evaluations on seven flights across four municipalities in a large metropolitan area demonstrate 100% Recall@1 at search areas of approximately 113 km$^2$ and 254 km$^2$, and 71.4% Recall@1 when expanded to approximately 452 km$^2$, encompassing up to 277,000 buildings. In contrast, baseline methods degrade rapidly and achieve 0% Recall@1 at 254 km$^2$ and 452 km$^2$.
Read the full abstract ▾
Mesoscale and Nanoscale Physics
Active control of THz plasmon propagation in a one-dimensional electronic waveguide
Thomas Vasselon, Uzer Ahmad, Clément Geffroy et al. (+14 more)
Quantum nanoelectronics is pushing towards ever higher operating frequencies in order to realise
quantum technologies capable of processing information at unprecedented speeds. A particularly promising direction is the development of flying electron
qubits, which offer the prospect of
quantum operations on picosecond timescales. Achieving in-flight
quantum control in this regime would establish a fundamentally new paradigm for studying
quantum entanglement and enable a novel form of
quantum information processing based on propagating electronic wavepackets. Here we report a first step towards this goal by injecting ultrashort electron wavepackets into an engineered
quantum nanoelectronic device. We demonstrate active control over the propagation speed of an electron wavepacket with a temporal duration as short as 4 ps in a quasi-one-dimensional electron waveguide with a length ranging from 10 $\mathrm{\mu}$m to 40 $\mathrm{\mu}$m. This advance provides a key building block for ultrafast
quantum operations using flying electrons. Beyond its technological implications, our approach offers a platform for exploring the intrinsic dynamical processes that govern
quantum transport and coherence in nanoscale electronic systems.
Read the full abstract ▾
All authors Thomas Vasselon, Uzer Ahmad, Clément Geffroy, Lucas Mazzella, Matteo Aluffi, Seddik Ouacel, Jashwanth Shaju, Kevin Bredillet, Nathan Roussel, Jean-François Roux, Pierre-Baptiste Vigneron, Arne Ludwig, Andreas D. Wieck, Matias Urdampilleta, Hermann Sellier, Giorgos Georgiou, Christopher Bäuerle
High Energy Physics (Phenomenology)
Exclusive photoproduction of vector toponium within the color dipole picture
Victor P. Goncalves, Bruno D. Moreira and Haimon Trebien
The exclusive photoproduction of the toponium vector
meson ($\psi_t$) is investigated within the color dipole model. The overlap function between the light-cone wave function of the vector toponium state and the
photon wave function is estimated using the Boosted Gaussian approach. Considering the dominance of small dipoles, the dipole-
proton scattering amplitude is estimated in terms of parton distribution functions, and predictions for the photoproduction cross-section in the energy range of the proposed Large
Hadron electron
Collider (LHeC) are presented. Moreover, the cross-sections and rapidity distributions for the exclusive $\psi_t$ photoproduction in ultraperipheral
proton-
proton and
proton -
nucleus collisions at the Large
Hadron Collider (LHC) and Future Circular
Collider (FCC) energies are calculated. We find that exclusive $\psi_t$ photoproduction is, in principle, feasible in ultraperipheral $pp$ collisions at the FCC-hh.
Read the full abstract ▾
Computer Vision
Xueqi Qiu, Xingyu Miao, Jingjing Deng et al. (+3 more)
Unified 3D
vision-
language systems must combine complementary geometry, scale, and appearance cues while supporting tasks from instance
segmentation to
language-guided
reasoning. Existing methods often process point clouds, voxel grids, and multi-view images independently; directly combining these heterogeneous representations may leave substantial feature discrepancy unresolved, while subsequent unconstrained adaptation may distort their internal geometry. We propose an align-then-fuse framework that first applies triple pairwise cosine
alignment to establish segment-level correspondence across the three representations and then retrieves task-conditioned features with a
prompt-guided query decoder. Before fusion, representation-specific query features are transformed by learnable mappings constrained to the special orthogonal group. These mappings preserve inner products and Euclidean distances within each representation, permitting controlled representation-specific re-parameterisation without arbitrarily distorting its internal geometry. The transformed features are subsequently combined through Adaptive Fusion under downstream task supervision. Experiments cover eight
datasets for instance
segmentation, visual grounding, question answering, and dense captioning. Compared with PQ3D, the model improves average precision by 3.2 points on ScanNet200 and grounding accuracy by 2.9, 10.6, 4.6, and 4.1 points on ScanRefer, Nr3D, Sr3D, and Multi3DRefer, respectively, while also improving performance on ScanQA, SQA3D, and Scan2Cap. Ablations further support the complementary roles of
alignment and orthogonal re-parameterisation and the effectiveness of Adaptive Fusion.
Read the full abstract ▾
All authors Xueqi Qiu, Xingyu Miao, Jingjing Deng, Haoran Duan, Yang Long, Ling Shao
Physics Education
Determination of the Charge of Electron Using Faraday's Law of Electrolysis
Vladimir Ya. Kezerashvili and Roman Ya. Kezerashvili
We propose a simple experiment for measuring the charge of the electron based on Faraday's law of electrolysis, assuming the known value of Avogadro's number. The experimental setup consists of electrolytic
cells connected in series, so that the same electric current passes through each electrolyte. By measuring the mass deposited at the electrodes and relating it to the total electric charge transferred through the electrolyte using Faraday's law, the charge of the electron can be determined. The experiment requires only standard laboratory equipment, and its performance and analysis are within the skills and knowledge of most students taking
algebra-based general physics courses. It is therefore particularly suitable for instructional laboratories in high schools and two- and four-year colleges. We demonstrate that students can obtain a reasonable value for the charge of the electron from simple measurements, providing a direct experimental connection between Faraday's law, Avogadro's number, and the electron charge.
Read the full abstract ▾
Differential Geometry
Derived Smooth and Banach Higher Groupoids: Representability and Descent
Qingyun Zeng
We study homotopy theories of higher groupoids in specified geometries, distinguishing
geometric representability from realization and descent. Ordinary smooth and
Banach-open sheaves admit full Brown categories of fibrant objects. For ordinary
geometric groupoids, a crossing-axes obstruction motivates an empty-compatible incomplete Brown structure. Its split-
Banach version requires a fixed chart-compatible plot site with local kernel-product closure. These results include all finite unique-horn bounds and their union. For represented derived models, finite matching and an outer-prism filtration give the structural Brown calculus. Nuiten's finite-
geometric realization
theorem yields the finite derived-smooth structure and its localization. Enriched realization base change on the original small Kan-enriched site gives the outer-unbounded hypercomplete structure, with positive-degree
geometric horn conditions and Reedy fibrancy in every degree. For
Banach domains, structured spectra, strict-open gluing and split-open hyperdescent establish
geometric finite limits and hypersheaf representability. They give the represented Brown structure, using the same hypercompleted weak-equivalence class at every outer bound. A square-zero test detects nonordinary self-intersections of the chosen
Banach charts, including infinite-dimensional charts. For connective associative affines, homotopy split horns, Pridham's finite effectivity
theorem and a pointwise diagonal-fibration argument give categories of fibrant objects for his Artin and Deligne–Mumford classes. The associative theory uses quasi-isomorphisms of
algebras and realization in presheaves.
Read the full abstract ▾
Networking
Load Balancing with Partial Queue Information - Threshold Optimality and Indexability
Sathwik Chadaga and Eytan Modiano
We consider the problem of load balancing in a system with one dispatcher and $N$ parallel servers. The dispatcher must select one server to dispatch new jobs at every time-step and each server buffers incoming jobs in a queue. However, the dispatcher does not know the servers' backlogs and must make dispatching decisions based on previous observations. The dispatcher's objective is to dispatch jobs to the shortest queue. This problem can be formulated as a restless multi-arm bandit (RMAB) problem where each arm's state is its corresponding belief vector. Our goal is to verify Whittle indexability for this problem and derive a low complexity Whittle index policy. Previous Whittle indexability results cannot be directly applied due to the multi-dimensional nature of the belief vector. To overcome this issue, we define the RMAB state as the tuple of the most recent backlog observation and the time since this observation. We consider two model variations, a standard finite queue model and a blocking queue model. We show that the single-arm decoupled problems of both these models have threshold optimal solutions under some assumptions. For the standard finite queue model, we prove indexability and derive the Whittle index policy in closed form. For the blocking queue model, we derive a sufficient condition for indexability under threshold optimality and use it to show indexability for some special cases.
Read the full abstract ▾
High Energy Physics (Phenomenology)
Solar capture and annihilation of fermionic and scalar dark matter motivated by the LUX-ZEPLIN high-recoil candidate
Antonio Capolupo, Gabriele Pisacane and Raoul Serao
Motivated by the LUX-ZEPLIN high-energy
nuclear-
recoil candidate, we ask whether
solar evolution can distinguish dark-matter models that are degenerate in their leading xenon
recoil response. We compare a split fermionic electroweak doublet with a specified inert scalar doublet, matching their masses, excitation energy, coherent transition interaction, and leading elastic
nuclear response. This matching fixes the same direct-detection signal and the same optically thin first-capture source in the Sun, while leaving post-capture dynamics model dependent. We follow finite-temperature capture, orbital
evolution, scattering, excited-state decay, escape, thermal entry, and annihilation. The different excited-state lifetimes generate distinct capture histories, but, for the specified cooling interaction, both retained populations thermalize and reach capture–annihilation
equilibrium, yielding only a modest difference in their late annihilation rates. We also derive an analytic protected-capture criterion and test the result against shared
halo,
solar-composition, and
nuclear-response variations. Although the total annihilation rates remain similar, model-dependent gauge-
boson branching fractions and polarizations produce distinct
neutrino spectra at production. A source-level comparison with published IceCube templates shows that the nominal partial $WW$ sources exceed the corresponding pure-$WW$ template limits, without constituting a detector-level exclusion or model discrimination.
Solar evolution therefore provides a complementary probe of particle-model differences hidden by a matched xenon signal.
Read the full abstract ▾
High Energy Astrophysical Phenomena
4U 1624-490 as a possible intermediate-mass X-ray binary: constraints from XRISM and NuSTAR spectroscopy
R. Iaria, A. Anitra, T. Di Salvo et al. (+7 more)
High-inclination dipping
neutron-
star X-ray binaries provide a direct view of
accretion-flow structure because the line of sight intercepts ionised
plasma above the disc. We analyse two XRISM/Resolve observations of 4U 1624$-$490 with simultaneous NuSTAR coverage, dividing the data into seven orbital-phase intervals. Phase-resolved fits include neutral partial covering, a photoionised absorber, thermal Comptonisation, and ionised disc reflection. Fe XXV and Fe XXVI absorption are detected over most of the orbit. The absorber column density, ionisation state, covering fraction, and projected velocity vary with orbital phase and the lines are systematically blueshifted, with line-of-sight velocities of about $85$-$440\ {\rm km\,s^{-1}}$. Absorption-only models leave broad Fe-K residuals, while ionised reflection improves the fits. The velocity modulation has a semi-amplitude $K=210\pm14\ {\rm km\,s^{-1}}$ and, if interpreted as orbital motion, implies a mass function $f(M)=0.83\pm0.17\,M_\odot$. Residual structure in the blue wing of Fe XXVI suggests additional kinematic complexity near the dip, but the modulation remains consistent when the two dip-adjacent measurements are excluded. If the absorber traces the
neutron-
star orbital motion, the inferred donor mass is $M_2=2.66\pm0.32\,M_\odot$, compatible with a Roche-lobe-filling B9 V-B9.5 V
star. Under this interpretation, 4U 1624$-$490 may be an intermediate-mass rather than a classical low-mass X-ray binary. (Abridged abstract)
Read the full abstract ▾
All authors R. Iaria, A. Anitra, T. Di Salvo, A. Sanna, F. Barra, L. Burderi, C. Maraventano, S. Caserta, W. Leone, C. Miceli
Software Engineering
From Agent Output to Authorized Transition
Christopher Koch
Agentic engineering systems can edit repositories, run tools and tests, build firmware, synthesize schematics, and prepare deployable or manufacturable artifacts. The assurance problem is therefore shifting from whether an
agent can produce an output to whether an engineering lifecycle is justified in acting on claims about that output. Current products and standards provide sandboxes, approvals, hooks, traces, policy enforcement, attestations, bills of materials, and assurance representations, but these capabilities remain fragmented. This paper presents the Agile-V Assurance Spine, a cross-domain transition contract for software, firmware, and PCB engineering. Evidence is admitted only when it establishes required properties through an authoritative source profile, is bound to the exact artifact and frozen policy baseline, remains current with respect to declared dependencies, and satisfies risk-appropriate independence and authority. Gate decisions are recorded as receipts; approvals and exceptions are exact-scope and time-bounded; and authorization is rechecked at the effect boundary before merge, deployment, flashing, release, or fabrication. A bounded review of contemporary research, commercial platforms, open-source infrastructure, and standards positions the model relative to evidence-gated lifecycle control, continuous assurance, runtime admission, provenance, and AI/ML inventories. The paper contributes a precise vocabulary, compositional architecture, domain profiles, mapping to open-source implementations, and an
adversarial evaluation agenda. It does not claim regulatory conformity or demonstrated production superiority.
Read the full abstract ▾
Analysis of PDEs
Sharp Quantitative Stability for the Affine $p$-Sobolev Inequality, Part II: The Case $1<p<2$
Song Fan, Gui-Dong Li and Jianjun Zhang
We prove a sharp quantitative stability result for the affine \(L^p\)-Sobolev inequality, for \(1<p<2\), introduced by Lutwak–Yang–Zhang (J. Differential Geom., 62 (2002),17–38). Moreover, the stability exponent is shown to be optimal, and equal to \(2\).
Materials Science
Disentangling spin polarization from driven circular ionic motion in EuTiO$_3$
Clifford J. Allington, Matthew J. Lutz, Enoch Ho et al. (+30 more)
Coherently driven circular ionic motion has been reported to produce large helicity-dependent optical responses attributed to transient magnetization. However, the microscopic nature and magnitude of this phenomenon, sometimes referred to as dynamical multiferroicity, remain strongly debated. Here, we directly study the connection between ionic circulation induced by a high-field circular terahertz (THz) drive and its effect on the
spin system of EuTiO$_3$ using X-rays. This material exhibits an optical response consistent with the putative
magnetic signal observed in related non-
magnetic materials. Furthermore, the presence of Eu$^{2+}$ ions enables the use of X-ray
magnetic circular dichroism (XMCD) to test for the creation of transient
spin polarization. Simultaneously, ultrafast X-ray diffraction (XRD) measures the circular ionic motion from which we determine the mechanical angular momentum. We find a classical ionic contribution of $3\times10^{-8}~\mu_B$ from XRD and upper limits of $0.03~\mu_B$ and $0.11~\mu_B$ from the sensitivity of XMCD at the europium $M_5$ and $L_2$ edges, probing the $4f$ and $5d$ shells of Eu$^{2+}$, respectively. These observations show that large-amplitude circular ionic motion in this system is not accompanied by a detectable
spin polarization despite the clear signature in the optical data, with the XMCD upper bounds and classical contribution differing by many orders of magnitude.
Read the full abstract ▾
All authors Clifford J. Allington, Matthew J. Lutz, Enoch Ho, Fabian Graf, Martina Basini, Hiroki Ueda, Michael Grimes, Alexander B. Elliott, Megan F. Biggs, Ravi Finn, Shih-Wen Huang, Merideth A. Henstridge, Matthias C. Hoffmann, Takahiro Sato, Roberto Alonso-Mori, Diling Zhu, Quynh L. Nguyen, Vincent Esposito, Jeffrey T. Babicz Jr., Elizabeth Skoropata, Biaolong Liu, Eugenio Paris, Arnau Romaguera Camps, Elia Razzoli, Roman Mankowsky, Flavio Capotondi, Nicolas Jaouen, Daniele Pergolesi, Milan Radovic, Matteo Savoini, Steven L. Johnson, Jeremy A. Johnson, Urs Staub
Mesoscale and Nanoscale Physics
Local time-reversal-invariant topological markers in two dimensions
Thomas Klein Kvorning, Miguel F. Martínez and Julia D. Hannukainen
Local topological markers provide a real-space characterization of topological phases and are suited to noncrystalline models. We derive a local topological marker, the Chern-Simons descendant marker, for characterizing the
topology of two-dimensional time-reversal-invariant near-Gaussian states from the one-particle density
matrix \(\rho\). We construct an auxiliary three-dimensional state characterized by the local Chern-Simons marker, and use dimensional reduction to obtain the Chern-Simons descendant marker for the original two-dimensional state. This derivation introduces a local Hermitian involution \(S\) that is odd under time reversal. When \([S,\rho]=0\), the local Chern-Simons descendant marker reduces to half the difference of the local Chern markers in the \(S=\pm1\) sectors. Constructing an \(S\) that commutes exactly with \(\rho\) generally requires fine-tuning, but this is not needed—the local Chern-Simon marker remains well defined whenever \(\Gamma=\{\rho,S\}-S\) has a spectral gap around zero, even when \([S,\rho]\neq0\). This leaves the practical problem of identifying a suitable \(S\) for a given state, which, importantly, is straightforward even when
spin is not a candidate. We show this by considering both strong Rashba and Dresselhaus
spin-orbit coupling, for which knowledge of the physical origin of the state reduces the general problem of choosing $S$ to a one-parameter family \(S(\theta)\). We evaluate the optimal angle $\theta$ maximizing the spectral gap of $\Gamma$ analytically assuming translation invariance, and find that the gap of $\Gamma$ remains open for a broad range of angles around the optimum. The same value of \(\theta\) remains valid in disordered systems, providing an easy way to obtain choice of \(S\) for evaluating the local Chern-Simons descendant marker of the state.
Read the full abstract ▾
Machine Learning
Log-Depth Recurrent Language Modeling
Yiqin Wang, Nuri Cingillioglu and Charles Pert
Language modeling using
Transformers has become commonplace despite their fixed computational depth and quadratic runtime with respect to input
tokens. Recurrent models on the other hand offer linear depth but no parallel execution. In this work, we extend balanced-tree recursive operators from sequence encoding to autoregressive prediction, enabling all prefix representations to be computed with logarithmic depth and linear runtime. Our experiments provide an initial characterization of this model class, demonstrating robust length extrapolation and performance approaching that of ALiBi-based
Transformers, highlighting its potential as an alternative architecture for
language modeling.
Read the full abstract ▾
Number Theory
Distinguishing Gauss sums
Moshe Adrian, Jack Diamond, Kenneth Kramer et al. (+1 more)
Let $\F_q$ be the field of order $q = p^f$ for
prime $p$. If $G(\chi)$ is the Gauss sum attached to a multiplicative character $\chi$ on $\F_q^\times$, then $G(\chi) = G(\chi^p)$. We investigate the converse question: when does the equality of Gauss sums $G(\chi_2) = G(\chi_1)$ imply that $\chi_2 = \chi_1^{p^j}$ for some integer $j$. If $\chi_2 = \chi_1^{p^j}$, we say that $\chi_1$ and $\chi_2$ are Frobenius-conjugate. We use the Stickelberger factorization of ideals in cyclotomic fields to give an easily testable criterion for equality of Gauss sums, based on $p$-adic digit expansion. As an application, we develop several conditions under which Gauss sum equalities between characters on $\F_q^\times$ are explained by Frobenius-conjugacy. For example, if $\chi_1$ has order $q-1$ or $\frac{1}{2}(q-1)$ and $G(\chi_2) = G(\chi_1)$, then $\chi_1$ and $\chi_2$ are Frobenius-conjugate. We also include several examples, based on the explicit evaluation of certain pure Gauss sums by R.J. Evans.
Read the full abstract ▾
All authors Moshe Adrian, Jack Diamond, Kenneth Kramer, Geo Kam-Fai Tam