Mesoscale and Nanoscale Physics
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Sunday 6 September
We derive a two-scale continuum limit for the electrostatic energy of random charge density fields that are stationary and ergodic under lattice translations. Each microscopic cell is charge neutral in every realization, while its dipole moment may fluctuate and have a nonzero mean. Under assumptions on the stationary microscopic potential and the finite-volume fields, the local and nonlocal energies converge almost surely to deterministic limits. The local limit consists of the ensemble-averaged microscopic Coulomb energy and a cell-depolarization term, whereas the effective polarization determines the nonlocal limit. The cell-depolarization terms cancel in the total energy, which consists of the ensemble-averaged microscopic cell energy and the macroscopic electrostatic field energy. The deterministic specialization recovers the periodic two-scale limit. For an independent-cell random-displacement model, we verify the assumptions of the continuum-limit theorem and show that centered microscopic fluctuations can change the local energy without changing the effective polarization or the nonlocal energy.
Computer Vision
ChildGaze: A Benchmark Dataset for Collaborative Behavior Understanding in Children
Understanding collaborative behavior in children is important for analyzing social participation, peer interaction, shared attention, and engagement during play and learning activities. Reliable recognition of these cues can support research in child development, educational analysis, and human-centered computer vision. However, estimating where a child is looking does not necessarily reveal whether the child is actively participating in a shared activity. To support this higher-level analysis, we introduce ChildGaze, a child-centered behavioral annotation dataset built on the ChildPlay video collection [1]. ChildGaze introduces two behavioral labels, collaborative and non-collaborative, assigned independently to each child within a frame. The dataset provides face, left-hand, and right-hand bounding boxes for children and adults and organizes the annotations at the row, person, and frame levels. The current release contains 27 annotated video files, 10,641 frames, and 73,268 body-part annotation rows. Annotation reliability was evaluated on 1,187 frames using independent annotations from two annotators. The collaboration labels achieved 93.16% raw agreement and a Cohen's kappa of 0.8631, while bounding-box annotations achieved an overall mean IoU of 0.808. Baseline experiments with pretrained ViT and Swin Transformer models achieved up to 97.44% child-person-level accuracy and 96.80% frame-level accuracy, respectively. These results show that ChildGaze provides a reliable benchmark for studying collaborative behavior in naturalistic child-adult and peer interactions.
Read the full abstract
Solar and Stellar Astrophysics
IPHAS J190812.63+045728.1: A deeply eclipsing, X-ray bright cataclysmic variable with frequent outbursts
We present a multi-wavelength characterization of IPHAS J190812.63+045728.1. Independently discovered by us after being serendipitously observed by XMM-Newton, followed up with optical photometry with SPARC4/OPD and spectroscopy with GMOS/Gemini South, and exploring archival TESS and ZTF observations, we reveal the system to be an eclipsing accreting white dwarf (WD) system with an orbital period of 5.073 h. The deep optical and flat-bottomed X-ray eclipses completely occult the WD and the accretion disk, providing a model-independent, uncontaminated view of the donor star which we classify as an M2.3 dwarf with a mass of $\sim 0.50\,M_\odot$. The system is a modestly luminous (L$_{x;0.3-10 keV} \sim$ 1.7 $\times$ 10$^{32}$ erg s$^{-1}$) hard thermal (consistent with a shock-heated plasma cooling down from $k$T $\sim$ 26 keV) X-ray source. While these X-ray properties are typical of low luminosity intermediate polars (IPs), long-term archival photometry reveals frequent, large-amplitude outbursts ($\Delta m \sim 2$ mag in the ZTF $zr$ band) that are the typical characteristics of dwarf novae, and would be unusual for IPs. The weakness of the HeII $\lambda$4686 emission line does not allow us to decide between the IP and the dwarf nova scenarios. On the other hand, the optical and X-ray light curves do not exhibit periodic variations caused by the WD rotation. If the system is indeed an IP, the accretion geometry is likely unusual while the dwarf nova-like outbursts indicating disk instabilities would make IPHAS J190812.63+045728.1 a benchmark for testing disk instability models in the presence of magnetically truncated accretion disks. Otherwise, the system is a dwarf nova with X-ray properties at the extreme end of its class.
Read the full abstract
Systems and Control
A Theory of Information Architecture for Networked Decisions: Freshness, Locality, and Coordination
A broader view of a network can help agents coordinate, but collecting that view takes time. When is a fresh local observation more useful than an older view of the whole network, and when should an agent look only partway across it? We study agents that repeatedly minimize a shared quadratic cost in an environment that evolves independently of their actions. An information architecture specifies what each agent observes and how old those observations are. Each agent uses a synchronized neighborhood snapshot, so extending its scope also ages the nearby readings used in the decision. We assume that the optimal action with complete information is affine in the state, and that the conditional mean of the current state equals the old state scaled by a common exponential decay factor. Using the classical Marschak–Radner decomposition, we derive the delay threshold beyond which fresh local information outperforms a complete but older global view. This comparison allows non-Gaussian states and nonlinear decision rules. We then determine how far an agent should look by balancing the benefit of additional observations against their greater age. In a Gaussian spatial model on a line, we derive the optimal observation radius: looking beyond the local node pays only when information travels more than twice the spatial correlation length in one coherence time. The preferred information architecture also depends on the decision being made. On a fixed 64-node ring, increasing only the reach of the decision weights from 0.7 to 6 hops shifts the optimal observation radius from one to five hops.
Read the full abstract
Materials Science
Barrierless Water Dissociation on Rare-Earth Sesquioxide Surfaces from First Principles
Water dissociation on metal oxide surfaces is a key elementary step in heterogeneous catalysis, photocatalysis, and radiation chemistry, yet its mechanistic details on rare-earth (RE) sesquioxides remain poorly understood. Here, we investigate water dissociation on the (110) surfaces of three cubic bixbyite oxides, Sc$_2$O$_3$, Y$_2$O$_3$, and Lu$_2$O$_3$, using molecular dynamics combining ab initio calculations with on-the-fly machine-learning force field acceleration. By sampling 25 independent trajectories per material, we obtain an unbiased picture of the reaction landscape inaccessible to conventional static calculations. Two distinct dissociation pathways are identified: a conventional proximal mechanism with a small but finite barrier of $\sim$0.1 eV, and a previously unreported distal mechanism that is effectively barrierless and energetically preferred at both the adsorption and dissociation stages. The low barriers are consistent with the periodic array of inherently undercoordinated RE$^{3+}$ sites in the bixbyite lattice, suggesting that ordered intrinsic coordination defects play a role analogous to stochastic oxygen vacancies in conventional oxides.
Read the full abstract
Other Condensed Matter
Single Dirac Fermions on a Lattice: A Unified View of Chiral and Parity Anomalies in Condensed Matter
Quantum anomalies are usually formulated in continuum field theory, where regularization may violate a classical symmetry. In a crystal, the compact Brillouin zone and the high-energy bands make that ultraviolet completion physical. This Perspective develops a unified lattice view of chiral and parity anomalies through the problem of realizing a single massless Dirac fermion. The central motif is a finite low-energy region in which the relevant symmetry is restored, embedded in a high-energy sector that breaks the symmetry and completes the lattice theory. In $1+1$ and $3+1$ dimensions, the quantized chiral response is protected by local chiral symmetry near the Fermi surface. In $2+1$ dimensions, the half-integer Hall response belongs to a single massless Dirac band with mirror symmetry, not to a generic massive Dirac insulator that breaks the symmetry explicitly. In both cases the lattice perspective reveals how the low-energy symmetry emerges from the high-energy sector and protects the anomalous and quantized response. Recent experimental progress in quantum materials is discussed in light of this unified view.
Read the full abstract
Complex Variables
A High-Rank Gap Theorem for Type-I Bounded Symmetric Domains
Let $\Omega_{r,s}$ and $\Omega_{r',s'}$ be Type-I bounded symmetric domains with $s>r$ and $s'>r'$. Let $F:U\to M_{r',s'}(\mathbb C)$ be holomorphic, where $\overline{\Omega}_{r,s}\subset U$, and suppose that $F(S_{r,s})\subset S_{r',s'}$. If \[ k(s-r)\leq s'-r'<(k+1)(s-r) \quad\text{and}\quad r'>kr, \] then, up to target coordinates, the map contains a fixed identity block of size $r'-kr$. The proof first turns the boundary equation into a matrix identity for the first derivative. A direct dimension count gives $g(s-r)\leq s'-r'$, where $g$ is the rank of a positive semidefinite matrix arising from this identity. The extremal property of the Shilov boundary and the scalar Hopf boundary lemma then show that certain rows of the map are constant. This gives the fixed identity block. We also give examples showing that its size is optimal and explain why the proof needs $F$ on a neighborhood of $\overline{\Omega}_{r,s}$.
Read the full abstract
General Relativity and Quantum Cosmology
Distinguishing lensing and precessional modulation in binary black-hole inspiral waveforms
Binary black holes (BBHs) emit gravitational waves (GWs) as they inspiral towards merger. These GWs can be gravitationally lensed by large-scale structure along the line of sight, potentially creating multiple images of the same source with fixed time delays determined by the lensing geometry. As the BBHs inspiral, the GW frequency increases, leading to successive constructive and destructive interference between the multiple images. BBHs also have spins $\mathbf{S}_i$ that may be misaligned with their orbital angular momentum $\mathbf{L}$. As the BBHs inspiral, these misaligned spins cause $\mathbf{L}$ to precess about the total angular momentum $\mathbf{J}$, modulating the GW emission similar to pulsar emission resulting from a misaligned jet rotating in and out of the line of sight. We investigate the ability of a single L-shaped GW detector to distinguish between these two sources of modulation. We find that precessional modulation can mimic the lensing modulation between two images with comparable magnifications when the time delay between the images is short enough that fewer than three interference fringes occur during the time the GW signal spends in the sensitivity band of the detector. As strong lensing is rare for GW sources at moderate redshift while misaligned spins are common for BBHs produced in certain formation channels, ruling out precessional modulation is essential to identifying genuinely lensed systems.
Read the full abstract
Machine Learning
Robust Dynamic Expansion for Continual Learning under Backdoor Attacks via Purification and Selective Recovery
Continual learning (CL) enables models to acquire new knowledge from sequentially arriving tasks while retaining previously learned knowledge. However, in practical scenarios, task streams collected from untrusted sources may contain backdoor-poisoned samples, posing a critical challenge to the stability, plasticity, and security of continual learners. In this work, we investigate a challenging setting termed Continual Learning Under Backdoor Attack (CLUBA), where each incremental task may involve a small proportion of maliciously manipulated training samples. Unlike conventional continual learning or backdoor defense scenarios, CLUBA requires models to simultaneously mitigate catastrophic forgetting, preserve adaptation capability, and prevent the absorption of malicious supervision during sequential updates. To address this challenge, we propose a robust dynamic-expansion framework that integrates sample purification, selective recovery, and robust expert routing into a unified continual learning paradigm. Specifically, we introduce Bi-Prototype Purification (BPP) to identify suspicious samples by exploiting semantic discrepancies in feature space. Based on purified data, Gradient Discrepancy-based Robustness Optimization (GDBRO) selectively recovers informative poisoned samples through pseudo-label correction and gradient consistency evaluation, improving robustness while maintaining model plasticity. Furthermore, Robust Feature Consistency-based Expert Selection (RFCBES) constructs perturbation-aware class prototypes to enable reliable expert routing under corrupted or shifted inputs.
Read the full abstract
Geometric Topology
Aspherical manifolds with nonvanishing tautological classes
For every even integer $m\geq 2$, we construct a closed, orientable, smooth, aspherical $(2m+1)$-manifold whose fundamental group has nontrivial center and for which infinitely many tautological classes in the $\mathbb{F}_2$-cohomology of the classifying space of its group of homotopically trivial diffeomorphisms are nonzero and not nilpotent. Specializing the calculation to cohomological degree $0$ gives examples of such manifolds that do not bound compact smooth manifolds. These provide counterexamples to a conjecture of Hebestreit–Land–Lück–Randal-Williams.
Read the full abstract
Analysis of PDEs
Generalized Whiting transform and Flux Estimates for Wave Equations in Subextremal Kerr
All recent advances \cite{DRS}, \cite{S-Rita1}, \cite{S-Rita2}, \cite{Millet}, \cite{MaS}, \cite{MaS2} on the boundedness and decay for wave equations on subextremal Kerr spacetimes $\KK(a,m)$, with non-small angular rotation, rely in an essential way on Whiting's groundbreaking discovery of a powerful integro-differential transformation with the help of which he ruled out exponentially growing modes in \cite{W} (see also \cite{Yacov}, \cite{AMPW} and \cite{Rita} for further developments). We give a purely physical space, generalized version, of the transformation which takes general solutions of {spin-$\sk$} wave equations in $\KK=\KK(a,m)$ to solutions of a secondary wave equation on a new metric manifold $\KKt=\KKt(a,m)$ with two stationary, asymptotically flat, ends. Moreover, for a certain subdomain $\DDt\subset\DDt_-$, which can be identified as the exterior of a black hole region, the new metric $\gt$ of $\KKt(a, m)$ is Lorentzian and the time translation $\T $ is timelike. Using the geometric properties of the two ends one can then deduce, by classical asymptotic Fourier analysis techniques, the desired bound for the flux of $\psi$ at the future event horizon. This estimate played a crucial role in our companion paper \cite{He-K1}.
Read the full abstract
Computer Vision
Radiation, Rotation and Scale Invariant Feature Descriptor for Multimodal Image Matching
Multimodal image matching is a fundamental task for multi-source information fusion. However, geometric distortions and nonlinear radiometric differences (NRD) severely limit performance, especially under radiometric, rotation, and scale variations. To address this issue, we propose a radiation, rotation, and scale invariant (RRSI) feature descriptor. First, a dual-head regional sampling (DHRS) module simultaneously performs Cartesian and Log-Polar sampling on keypoint neighborhoods, retaining spatial structural properties while enhancing robustness to rotation and scale variations. We then jointly encode geometric and radiometric relations between multimodal images in a unified deep feature space, enabling feature encoding, interaction, and fusion across intra-modal, dual-head sampled, and inter-modal regions. Furthermore, we introduce a bidirectional cross-modal generative reconstruction constraint during training. By decoding implicit features into structural patches of the counterpart modality, this mechanism anchors modality-invariant geometric topologies without additional inference overhead. Experiments on optical-infrared and optical-SAR datasets demonstrate highly competitive matching performance and strong robustness to rotation and scale variations. RRSI supports the full rotation range from 0 to 360 degrees and scale factors up to four. Its generalization ability is further validated on multimodal images from computer vision, remote sensing, and medical imaging. The implementation will be made publicly available at https://github.com/yeyuanxin110/RRSI .
Read the full abstract
Robotics
OcclusionCBF: Backup Control Barrier Functions for Safe Navigation Among Hidden Dynamic Obstacles
Robots navigating under occlusion may enter states from which no admissible input can avoid a dynamic obstacle once it becomes visible. We present OcclusionCBF, a safety filter that extends backup control barrier functions to reachable-occupancy predictions for potentially hidden dynamic obstacles in occluded regions. The method certifies a prescribed backup rollout against collision-inflated occupancy and a verified terminal set, yielding affine constraints for minimally invasive quadratic-program filtering. We establish recursive feasibility of the resulting safety filter, and collision avoidance for every hidden-obstacle motion covered by the occupancy prediction. Randomized benchmarks, MetaUrban simulations, and hardware experiments demonstrate improved task success over reactive and occlusion-aware predictive baselines with millisecond-scale computation.
Read the full abstract
Machine Learning
Linear Algebra Foundations of Efficient Attention: A Phase Reversal in Rank Collapse Under SVD Compression
Linear algebra provides the framework of concepts (matrix rank, singular value decomposition (SVD), and eigendecomposition) that modern artificial intelligence employs to encode, compress, and propagate information through neural networks. This paper unifies fourteen separate peer-reviewed works analyzing the usage of these techniques in the context of transformer-based foundation model research, focusing on three areas of the topic: derivations and properties of self-attention matrices' output rank, compression methods that purposefully utilize this phenomenon, and the low-rank key-value (KV) cache projection and its semiseparable-matrix duality to linear attention and state-space structured models. We were motivated to conduct this work after observing an open problem in this literature: the interplay of the mentioned compression methods with natural rank collapse of the network. With this paper, we report an original finding that using SVD compression of attention projections actually has the opposite effect on the rank collapse of the network: while it strongly suppresses it at initialization, it accelerates on pretrained models (for GPT-2 124M, GPT-2 Medium 355M, and Pythia-160M) with minimal risk of object aliasing artifacts appearing (verified on all compression ratios) and is consistent across four rank estimation methods. A controlled causal decomposition of the effect in both settings showed that the reason for this behavior can be explained by the choice of the subspace SVD makes when compressing the matrix better than the reduction of the operator norm it achieves, explaining roughly 76% of the effect at initialization and 83% on the pretrained weights, providing a refinement to the calibration-aware compression viewpoint and an explanation of why it outperformed naive SVD truncation.
Read the full abstract
Solar and Stellar Astrophysics
Coordinated Coronagraphic Observations from Proba-3 and Aditya-L1: Investigating CME Energetics in the Inner Corona
Constraining the plasma properties and energetics of coronal mass ejections (CMEs) in the inner corona is essential for understanding their early evolution, yet remains challenging because of limited observations. We investigate the mass and density evolution, and energy partitioning of two CMEs observed on 2025 September 14 and 16, and assess the potential of coordinated Proba-3/ASPIICS and Aditya-L1/VELC observations to constrain CME energetics in the low corona. We present the first coordinated observations of CMEs obtained simultaneously by Proba-3/ASPIICS and Aditya-L1/VELC. Using ASPIICS white-light observations, we estimate the CME mass, volume, number density, and the evolution of kinetic, thermal, and magnetic energies. Magnetic energies are estimated from observed CME properties using observationally constrained, physically motivated assumptions. Simultaneous VELC Fe XIV 5303 Angstorm observations provide independent estimates of the emission measure, electron number density, thermal energy, and CME lateral extent. The two events exhibit markedly different energy partitioning in the low corona. For the 14 September CME, the kinetic and magnetic energies are comparable, while the thermal energy remains nearly two orders of magnitude smaller, indicating limited plasma heating. In contrast, the 16 September CME exhibits a substantial thermal-energy enhancement, with thermal energy eventually becoming comparable to kinetic energy. For both CMEs, the estimated magnetic energy remains comparable to or exceeds the kinetic energy over the observed height. Our results demonstrate the scientific potential of synergetic ASPIICS and VELC observations for constraining CME mass, density, and energetics in the inner corona, providing new observational constraints on the early evolution of CMEs.
Read the full abstract
Rings and Algebras
Generic Nullity of Generalized Commutators
We study the generic nullity of generalized commutator operators \[L_{\mathbf{A}}(X)=s_{k+1}(A_1,\cdots,A_k,X)\] on matrix algebras, where $s_{k+1}$ denotes the standard polynomial. Dixon and Pressman conjectured an explicit formula for the generic nullity of $L_{\mathbf{A}}$, and Brassil and Reichstein proved the conjecture when $k$ is even. In this paper, we settle the remaining case where $k$ is odd. Our proof first treats the boundary cases $k=2n-3$ in dimensions $n$ and $n+1$ using degree decompositions and graph-theoretic interpretations of alternating trace forms, and then establishes a dimension-extension argument from $n$ to $n+2$. Consequently, together with the result of Brassil and Reichstein, we obtain a complete proof of the Dixon-Pressman generic nullity conjecture over any field of characteristic zero.
Read the full abstract
Computer Vision
One Shared LoRA Weight for MRI Reconstruction across Acceleration Factors
Accelerated MRI reconstruction recovers images from undersampled k-space. However, different acceleration factors produce distinct artifact patterns. Existing methods often train separate models for each factor, leading to poor cross-factor generalization and high training and storage costs. We propose Shared LoRA, a parameter-efficient framework that freezes the pretrained SHFormer backbone and trains a single shared set of LoRA adapters together with a lightweight gating network. During training, undersampled inputs are generated by randomly sampling acceleration factors and their corresponding sampling masks, enabling the shared adapters to learn reconstruction knowledge across factors. Given the acceleration factor, GateNet generates layer-wise coefficients to dynamically modulate the residual strength of each adapter. Experiments show that Shared LoRA achieves the best or competitive PSNR and SSIM across acceleration factors, while its trainable parameters account for only about 5.3% of the total model parameters. Its performance at lower acceleration factors remains largely unaffected as the jointly trained factor set expands, and it generalizes stably to unseen neighboring factors.
Read the full abstract
Numerical Analysis
Rough variational principles and applications to adjoint systems
We consider a Type-II variational principle driven by geometric rough path with split boundary conditions naturally suited to adjoint systems. From this rough variational principle we derive rough Hamilton's equations, establish their pathwise conservation laws and associated Hamilton–Jacobi equation. We then specialise the framework to rough adjoint systems, obtaining pathwise conservation and quasi-conservation laws that underpin adjoint sensitivity analysis with respect to initial conditions and parameters. On the discrete side, we construct a rough Galerkin discretisation of the rough Type-II variational principle and show that it generates a symplectic flow with discrete analogues of the continuous conservation laws. We establish it's equivalence to a class of Rough Symplectic Partitioned Runge–Kutta (RSPRK) methods and analyse its convergence and naturality properties. Lastly, we perform numerical experiments to validate the predicted convergence rates and demonstrate that RSPRK methods preserve the adjoint conservation laws to machine precision, yielding more accurate and stable gradients in optimisation problems than non-symplectic alternatives.
Read the full abstract
Analysis of PDEs
Large-time behavior of solutions to bipolar Euler-Poisson equations with space-dependent damping
This paper is concerned with the Cauchy problem for the one-dimensional bipolar Euler-Poisson equations with space-dependent damping, which arise in the hydrodynamic model of semiconductors. Notably, the two pressure functions are allowed to be different, and the doping profile is nonzero. By means of the time-weighted energy method with carefully chosen weights, we prove that if the initial perturbation around a constant steady-state is small enough, the Cauchy problem admits a unique global smooth solution, which converges to the constant steady-state at algebraic decay rates. We show that the decay rates coincide with those established for the constant-coefficients damping case, thereby overcoming the influence of spatially dependent damping on the decay rates of the solutions. A key feature is that no smallness of the space-dependent damping coefficients is required.
Read the full abstract
Materials Science
First Principles Thermodynamics of Zr B Segregation at Grain Boundaries in Recycled Nd2Fe14B
Grain-boundary chemistry is central to the coercivity and thermal stability of Nd2Fe14B permanent magnets, particularly in recycled magnets where recovery of the hard-magnetic phase does not by itself restore the intergranular microstructure. Our recent experimental study showed that nanoscale ZrB2 precipitates can emerge at grain boundaries and triple junctions during recycling despite an overall Zr concentration of only about 0.1 at.%, where they are associated with boundary stabilization and suppression of grain coarsening. Here we use spin-polarized density-functional theory with a Hubbard correction (DFT+U) to determine the atomistic thermodynamics underlying this preferential localization. A systematic set of composition-matched bulk/grain-boundary DFT+U supercells provides a common correlated-electron description across boundary chemistries. We compare B, Zr, Dy, and ZrB2-like local configurations in bulk and grain-boundary environments of Nd2Fe14B. Excess B is strongly stabilized at the boundary, while Zr also shows an independent thermodynamic preference for the interfacial region. When Zr and B are combined in a ZrB2-like configuration, the boundary preference is retained, indicating that the interface remains favorable as Zr-B coordination develops. These results establish a thermodynamic pathway for the co-localization of Zr and B prior to ZrB2 formation. Magnetic-state analysis further shows that all compared structures remain within the same high-moment Fe-sublattice regime, and boundary-localized defects generally perturb the normalized magnetization less than their matrix counterparts. The calculations therefore provide a first-principles explanation for why Zr-B chemistry concentrates at intergranular regions and how such boundary-localized states can support the microstructural stability required for high-coercivity recycled Nd-Fe-B magnets.
Read the full abstract
Materials Science
Chemical and interface confinement effects in promoting plastic co-deformation in high-strength nano-scale eutectics
Eutectics offer a route to overcome plastic incompatibility in disparate-phase heterostructures by coupling microstructural refinement with phase-specific chemical and crystallographic hierarchy. Here, we investigate laser-rapid-solidified Al-(Si,Ge) eutectic composites designed along the univariant ternary eutectic path, solidifying with an Al-rich face-centered cubic (fcc) matrix and (Si,Ge)-rich diamond-cubic (dc) fibers as constituent eutectic phases with faceted interfaces. The microstructure was hierarchical in nature with finer structures within eutectic phases: nanoscale (Si,Ge) clusters in the Al phase and growth twins within the (Si,Ge) fibers with some Al retention. Although increasing Ge content coarsens the eutectic spacing, the yield strength is slightly enhanced compared to the relatively finer Al-Si, and tensile ductility of Al-(Si,Ge) is higher than that of Al-Si. In situ SEM micromechanical testing combined with post-mortem STEM and TEM showed that in the Al-rich phase, Ge segregated to the deformation-induced sub-grain boundaries that confined glide dislocations. Simultaneously, the (Si,Ge)-rich fibers remain crack-resistant at large plastic strain and exhibit deformation-induced planar faults, consistent with localized partial-dislocation activity at highly stressed interfaces and twin boundaries. These coupled mechanisms due to interface confinement and hard phase chemistry enable co-deformation of the metallic and covalent phases, providing a pathway for designing high-tensile-strength and ductile hierarchical eutectics beyond conventional length-scale-controlled strengthening.
Read the full abstract
Number Theory
Mean Values and Normal Order of a Successive-Iterate Ratio of Dedekind's Arithmetic Function
We obtain asymptotic formulas for the three sums \[ \sum_{n\le x}\frac{\psi(\psi(n))}{\psi(n)},\qquad \sum_{n\le x}\frac{\psi(n)}{\psi(\psi(n))},\qquad \sum_{n\le x}\log\frac{\psi(\psi(n))}{\psi(n)}, \] where \(\psi\) denotes Dedekind's arithmetic function. We also determine the normal order of the successive-iterate ratio; more precisely, \[ \frac{\psi(\psi(n))}{\psi(n)} \sim \frac{6e^\gamma}{\pi^2}\log_3 n \] for almost all positive integers \(n\).
Read the full abstract
Number Theory
Normal Behavior and Periodic Points of a Pseudo-Aliquot Map Associated with the Dedekind Psi Function
We study the iteration of (T(n)=\psi(n)-n), where (\psi) is the Dedekind psi function. We show that, for almost all integers, successive iterates satisfy strong asymptotic growth relations governed by (T(n)/n). As an application, we prove that for every fixed (\ell\ge1), the set of integers satisfying (T^{(\ell)}(n)=n) has asymptotic density zero, and hence so does the set of periodic points of any fixed period.
Read the full abstract
Robotics
Geometric Distributional Control: Learning Progress with Partial Structural Knowledge
Real-time control often sits between two limiting regimes. Predictive optimization and model-based control are powerful when dynamics, parameters, objectives, and online planning models are specified; reinforcement learning can relax this requirement, but must infer long-horizon value signals from sequential data and interaction, making training slow, high-variance, and hard to scale in large action spaces. This middle regime is common in systems including autonomous driving, warehouse robotics, traffic control, and delivery drones: partial geometry, physics, rules, or constraints are known, yet the local direction of task progress remains uncertain. Geometric Distributional Control (GDC) is designed for this partial-knowledge setting. It factorizes control into feasibility and progress: known geometry, rules, constraints, and response maps define an executable scaffold, while progress-weighted feasible data learns the missing directional signal on that scaffold. The learned score acts as a Bellman-like local value-gradient, selecting actions that make progress without requiring global Bellman recursion, a fully specified planner, or a black-box policy that absorbs both feasibility and preference. This knowledge can be lightweight and partial, such as simple dynamics, safety filters, local maps, constraint projectors, or lower-level response maps; it need not encode full dynamics or a long-horizon objective. Offline, GDC fits a progress-tilted distribution from short known-feasible snippets with weak signed progress certificates. Online, its score is projected through the scaffold and applied in receding-horizon feedback. We prove that this score descends a data-induced soft progress value and validate GDC on structured multilevel optimization and SUMO route-progress driving, where it improves over known-only solvers and learning baselines while preserving scaffold-enforced feasibility.
Read the full abstract
Cryptography and Security
Beyond QA Matching: Perturbation-Response Fingerprinting via Probability Distributions for Large Language Models
Large language models are often instruction-tuned, specialized, quantized, or otherwise transformed, making fine-grained provenance difficult. In this paper, we introduce BReF, a training-free fingerprint that compares how probability distributions over four answer-option labels A/B/C/D move under controlled textual perturbations. For each pair of models, BReF selects 25 jointly responsive probes and compares their perturbation log-ratio (PLR) response directions by global cosine similarity. On a unified benchmark with 34 checkpoints, 22 documented direct-parent relations, and 411 suspect-candidate pairs, BReF retrieves the documented parent in 22/22 cases (MRR=1.0000), with DP-DF AUC 1.0000. Same-family discrimination is harder (DP-SF AUC 0.8969), and paired tests show a significant exact-retrieval gain over a magnitude-only Top-25 control. Together with static, random-probe, permutation, calibration, and transformation-level controls, the results show that strong pooled separation does not guarantee correct parent ranking among closely related checkpoints, verifying the superiority of our work.
Read the full abstract
Data Structures and Algorithms
Computation and Applications of Euclidean and Normed Representations of Massive Data
This thesis investigates Euclidean-space and $\ell_p$-norm representations of different forms of data, with a focus on efficient algorithms for computing these representations in settings where the amount of data is very large. The applications of such representations are also discussed: they may be used to summarize the data in a more compact form for downstream tasks while preserving its most salient properties; and further, they may also be used to extract insights about the data that may not be apparent in its original form. This thesis presents novel algorithms for computing Euclidean representations in the case where the data comes with linear structure, as well as in the case where the data comes only with metric, or distance structure. In addition, we give algorithms for computing $\ell_p$-norm representations in linear structured cases. The analyses of these algorithms use tools from geometry, probability, and optimization, and these tools are used to illuminate other algorithms for similar problems.
Read the full abstract
Programming Languages
SparseConflicts: Handling Conflicting Data Layouts in Sparse Tensor Contractions
Optimizing sparse tensor computations is challenging due to the use of compressed storage formats, which leads to non-affine loop nests and a vast, complex schedule space. The performance of a given schedule is sensitive to the sparsity pattern of the input tensors, making it difficult to find a single optimal solution. When input tensors in the same tensor contraction have conflicting data layouts in relation to the iteration order, it requires costly-both in time and memory-layout transformation, such as transposition. A promising but under-explored alternative is to generate a schedule that avoids explicit transposition, but this has not been systematically supported in existing compilers. This paper presents a new code generation strategy that generalizes the intermediate representation of the TACO sparse tensor compiler to generate a single loop nest, circumventing explicit transposition of tensors when the tensors have conflicting data layout iteration orders. We extend TACOś iteration graph to express a search-based strategy for locating elements in tensors with conflicting layouts, and we introduce new intermediate representation nodes to lower these schedules to efficient code. This enables the systematic generation of loops that do not require explicit transposition, thus avoiding the overhead of materializing temporary tensors. We evaluate our approach on a set of sparse tensor contractions using both real-world and synthetic datasets. Our results demonstrate that for computations with misaligned data layouts, our fused approach achieves up to 2x speedup for some sparsity patterns over the traditional approach of explicitly creating a transposed temporary. We also provide guidelines for when this new scheduling strategy is likely to be beneficial.
Read the full abstract
Data Structures and Algorithms
Query-Oblivious Coresets for Softmax Attention: Improved Bounds and Efficient Constructions
A query-oblivious coreset for a softmax-attention head is a subset of the key–value pairs whose attention output is within $\varepsilon$ of the full one for every query in a ball. With $\rho$ the query radius times the centred key radius, we build such coresets of size $O(e^{\rho}(\sqrt{d_v}+\sqrt{d_k\log(1+\rho)})/\varepsilon)$ in randomised polynomial time, improving the existential $O(\sqrt d\,e^{\rho+\frac12\log\rho+o(\log\log\rho)}/\varepsilon)$ of Liberty, Andoni and Kleiner to within $\sqrt{\log(1+\rho)}$ of their $\Omega(\sqrt d\,e^{\rho}/\varepsilon)$ lower bound: a spherical lift of both balls into one exponential-kernel instance admits the Bozzai–Rothvoss chaining bound, and Chevet's inequality splits key from value dimension. A sampling cap $O(e^{2\rho}/\varepsilon^{2})$ completes the envelope, and in fixed dimension Tai's diameter-free bound removes the logarithm. A census of every head of Qwen2.5-7B-Instruct and Llama-3.1-8B-Instruct finds $\rho\ge24.100$, so every whole-ball bound with a factor $e^{\rho}/\varepsilon$ prescribes a coreset larger than the cache: the algorithmic contribution is asymptotic on these models.
Read the full abstract
Systems and Control
Rethinking Safety for Generalist Robots
Generalist robots promise to transform our society: the same system that prepares a meal or folds laundry might also repair a car, inspect infrastructure, or care for a loved one. Yet this versatility introduces risks far beyond the collision- and force-based safety notions that have long dominated robotics. Notions of safety must now consider context (e.g., turning off a building's electricity is only safe during scheduled maintenance), user intent (e.g., asking the robot to “clean the kitchen” includes unspoken expectations that the robot should not mix dangerous but powerful cleaning agents like bleach and ammonia), hard-to-model physical consequences (e.g., burning food during meal preparation), and more. We argue the need for a new era of robot safety—embodied AI safety—that broadens the hazards considered across the robot's lifecycle while recognizing that the safety of bits cannot be separated from the safety of atoms. We present a taxonomy of emerging risks and a full-stack research agenda to guide the safe deployment of generalist robots.
Read the full abstract
Plasma Physics
Separation of bi-dispersed microspheres in dusty plasma ratchet experiments
It is demonstrated experimentally that the effective separation of bi-dispersed microspheres (dust particles) in the underdamped and strongly-coupled regime is realized using a designed dusty plasma ratchet. Experimental findings reveal that these dust particles can undergo directional transport at varying speeds, even moving in opposite directions depending on the discharge conditions, enabling successful particle separation. Numerical simulations of the plasma environment surrounding the dust particles are performed using fluid simulations of the capacitively coupled discharge of Argon. The simulation results indicate that the bi-dispersed dust particles are suspended at different balance heights within the plasma sheath and experience distinct ratchet potentials that govern their directional transport, resulting in varied flow velocities. The discovery of height-dependent transport of dust particles here provides insights of transport fundamental of underdamped strongly-coupled particles in dusty plasma ratchets.