Mesoscale and Nanoscale Physics
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Sunday 6 September
We investigate whether an Fe–dehydroindigo coordination polymer can support Haldane-chain physics in its intrinsic, substrate-free limit. Spin-orbit-coupled density-functional-theory calculations and magnetic energy mapping yield a nearly isotropic antiferromagnetic coupling, together with much weaker single-ion anisotropy. Using these parameters, exact diagonalization and density-matrix renormalization group calculations give a nondegenerate ground state separated from the lowest triplet-derived excitations by a gap of about $13$–$14$~meV, on the scale expected for the spin-1 Haldane gap. The anisotropy produces only a small splitting of the low-energy modes. The dynamical spin structure factor places the lowest spectral weight near the antiferromagnetic wave vector $q=\pi$. These results identify the isolated cis-dehydroindigo chain as a material-specific coordination-polymer platform whose microscopic interactions place it in a gapped regime consistent with Haldane-chain physics.
Algebraic Geometry
Automorphism Groups of Smooth Cubic Fourfolds through Lattice Theory
Laza and the second author have classified all possible symplectic automorphism groups of smooth cubic fourfolds. There are $34$ such groups, and by Koike, there are $48$ families of cubic fourfolds with speficied symplectic automorphism group. For $42$ among those families, we complete the classification of all possible automorphism groups. The approach of this paper is mainly lattice theoretic, starting from known results about the period map of cubic fourfolds by Voisin, Hassett, Looijenga and Laza. We use a lattice-enumeration algorithm implemented in OSCAR.
Read the full abstract
Earth and Planetary Astrophysics
An Outer Giant Excites and Constrains the Orbit of $\pi$ Mensae d
The nearby $\pi$ Mensae system contains the eccentric giant planet b on a 5.7 yr orbit, the 6.27 day transiting sub-Neptune c, and a candidate planet d on a 123 day orbit between them. Published estimates place the eccentricity of d between approximately 0.2 and 0.4. We reanalyze 772 published velocities with three-Keplerian fits and explore three-dimensional b-d dynamics with N-body integrations. The high-eccentricity solution persists when the phase or apsidal prior is widened; exact importance reweighting to a population-informed Beta eccentricity prior gives $e_{d}=0.378^{+0.050}_{-0.049}$. In 20 kyr integrations, 84.5% of configurations near the prograde and retrograde coplanar endpoints survive, compared with 21.3% across $60^\circ \leq I_{bd} < 120^\circ$; stellar collisions dominate the failures. Secular forcing by b moves 53 of 62 one-Myr survivors initialized with $e_{d} \leq 0.1$ into $0.3 \leq e_{d} \leq 0.5$. Representative b+d configurations remain viable for 100 Myr, whereas controls with c and general relativity retain fewer systems. The outer giant can therefore excite the observed eccentricity of d while constraining its viable orbit toward prograde or retrograde near-coplanar configurations.
Read the full abstract
Numerical Analysis
A discontinuous Galerkin approximation of the Dean–Kawasaki equation
We introduce and analyse an arbitrary order spatial discontinuous Galerkin (dG) method for the Dean–Kawasaki equation, a highly singular SPDE modelling density fluctuations of $N$ diffusing particles in the regime of large particle number $N \gg 1$. Our starting point is a general procedure for discretising multiplicative, divergence-form noise on finite element spaces whilst preserving its cross-variation structure at the discrete level; the construction is explicit, elementwise, applies to continuous and discontinuous spaces alike, and extends to general mobilities. Using it, we prove weak error estimates of order $O(h^p)$ between fluctuations of the semi-discrete scheme and those of the underlying particle system, together with a correction that is exponentially small in the scaling regime $Nh^d \gg 1$ and arises because the scheme does not preserve positivity. The resulting method is locally and globally conservative and applies on unstructured simplicial meshes. Quantitative numerical experiments support the analysis, and further experiments illustrate the method beyond the scope of the theory: indicator function observables, external and interaction potentials, convection-dominated regimes and reflecting boundary conditions.
Read the full abstract
Computational Physics
A HIP-Compatible Accelerator Backend for Fourier-Bessel Particle-in-Cell Simulations on CPU/DCU Heterogeneous Clusters
FBPIC (Fourier-Bessel particle-in-cell) is a high-performance simulation code for relativistic plasma and accelerator physics. Its original accelerator backend relies on Numba CUDA, which limits its direct deployment on accelerators using the HIP (Heterogeneous-Compute Interface for Portability) programming environment, such as DCU (Deep Computing Unit) accelerators. In this work, we develop an accelerator backend compatible with HIP that enables FBPIC to run efficiently on DCU platforms while preserving its Python user interface and high level simulation workflow. For the evaluated LWFA (laser-wakefield acceleration) workloads, the proposed backend achieves 1.32-1.54x speedups over the original FBPIC implementation on an NVIDIA V100 GPU and enables efficient execution on the DCU platform. We also summarize the key lessons learned from porting FBPIC to the DCU platform. Multi-DCU experiments achieve a 1.88x strong-scaling speedup on four accelerators and a 2.72x increase in aggregate throughput at approximately 68% weak-scaling efficiency, with communication analysis identifying inter-node communication and synchronization as the main scalability limitations. Beyond FBPIC, the proposed approach provides a practical reference for porting and optimizing other scientific computing applications developed with Python on heterogeneous accelerator platforms.
Read the full abstract
Astrophysics of Galaxies
ADF22-WEB: Massive galaxy formation shaped by cosmic web filaments in the z = 3.1 proto-cluster core
We present a census of dust, molecular gas, and galaxy structure in 18 dusty star-forming galaxies (DSFGs) at $z=3.09$ embedded in Ly$\alpha$-traced cosmic web filaments in the core of the SSA22 proto-cluster, using multi-band ALMA, JWST imaging, and JVLA CO(1–0). Using up to six ALMA bands and Herschel/SPIRE data, we construct rest-frame far-infrared spectral energy distributions. The DSFGs span two orders of magnitude in far-infrared luminosity ($L_{\rm FIR}\sim10^{11}$–$10^{13}L_\odot$), with median values of $\log(L_{\rm FIR}/L_\odot)=11.56^{+0.32}_{-0.37}$ and $T_{\rm dust}=25^{+6}_{-3}{\rm K}$. For eight DSFGs with high-resolution ($\sim1$ kpc) JWST and ALMA imaging, we measure stellar and dust surface densities and find a positive correlation, suggesting that structural compaction of the stellar component proceeds together with that of the interstellar medium. We report 12 detections of CO(1–0), 18 of CO(3–2), four each of CO(8–7) and CO(9–8), and one of CO(12–11). The median brightness temperature ratio between CO(3–2) and CO(1–0) is $r_{31}=0.66^{+0.05}_{-0.04}$, consistent with field galaxies at similar redshifts. The high-$J$ CO lines show relatively low excitation, and the CO(8–7)/CO(3–2) ratio correlates with star-formation rate surface density, suggesting that molecular gas excitation is primarily regulated by star formation. Lower-mass galaxies show excess molecular gas fractions and depletion times relative to field scaling relations. By combining quiescent galaxy samples in the same field, we identify a sequence from gas-rich systems with low stellar surface density to gas-poor, compact systems with high stellar surface density. These results suggest that massive galaxy evolution in dense environments is governed by a baryon cycle linking gas supply, star formation, and structural transformation within the cosmic web environment.
Read the full abstract
Algebraic Geometry
On the existence of smooth varieties of any dimension carrying no Ulrich bundles for some very ample line bundle
Using some recent examples of Anghel in dimension $2$, we prove that for any $n \ge 3$ there exist many smooth $n$-dimensional varieties $X$ with a very ample line bundle $H$, such that there is no Ulrich bundle for $(X,H)$. Moreover, we find many new examples in dimension $2$.
Number Theory
Unsupported Cyclotomic Divisors in Three-Prime Integer Tilings
Cyclotomic divisibility imposes strong prime-power structure on integer tiles. We study unsupported cyclotomic divisors: mixed-order divisors for which none of the prime-power components of the order divides the mask, although every prime in the order divides the tile cardinality. Kiss, Łaba, Marshall and Somlai asked whether such a phenomenon can occur in the three-prime setting. We prove that unsupported cyclotomic divisors already occur for periods with three distinct prime factors. For primes \(p<q<r\), we characterize the square-period case: an unsupported factor \(\Phi_{pqr}\) occurs in a tiling of \(\ZZ_{(pqr)^2}\) if and only if \(r\in\langle p,q\rangle\), and every such tile lies in a single residue class modulo \(r\). Among cyclic tilings with the unsupported order dividing the specified modulus, the smallest modulus is \(180\); if the order has three distinct prime factors, it is \(900\). An Apéry-set construction gives examples for every triple at period \(p^2q^2r^3\).The proof of our results combines Fourier rigidity, a three-cylinder decomposition, and an integer mass obstruction.
Read the full abstract
Quantum Algebra
Thickening realization and positivity properties of canonical bases
Let $\mathbf{U}$ be a quantum group associated with a symmetric Cartan datum, let $\dot{\mathbf{U}}$ be its modified form, and let $\dot{\mathbf{B}}$ be the canonical basis of $\dot{\mathbf{U}}$. Lusztig conjectured that the structure constants of the multiplication, comultiplication, and bilinear form in $\dot{\mathbf{U}}$ with respect to $\dot{\mathbf{B}}$ belong to $\mathbb{N}[v,v^{-1}]$. We introduce the thickening realization, which relates $\dot{\mathbf{B}}$ to the canonical basis of the negative part of a larger quantum group $\tilde{\mathbf{U}}^-$. More precisely, it identifies the relevant structure constants in $\dot{\mathbf{U}}$ with the structure constants in $\tilde{\mathbf{U}}^-$. As consequences, we prove, for arbitrary symmetric Cartan datum, $\dot{\mathbf{B}}$ has the positivity properties for the comultiplication and bilinear form, as well as the positivity for the multiplication whenever one factor is spherical parabolic. In particular, Lusztig's conjecture holds for simply-laced finite type. We also prove the canonical bases of a broad class of tensor products of integrable modules have the positivity properties for the transition matrices and actions by $\dot{\mathbf{B}}$.
Read the full abstract
Quantum Physics
Contact order governs the onset of entanglement cascades
Entangling rates describe a direct entangling channel, but give no information when that channel is forbidden. For finite-dimensional analytic pure-state dynamics, we show that the onset of genuine $(n+1)$-partite entanglement is determined by the contact order $m_\star$ between the physical trajectory and the $S|E$ product manifold. It is the first Taylor order that cannot be reproduced by any product curve, or equivalently the first nonvanishing order of the Fubini–Study distance from the product manifold. We derive a time-ordered recursion that removes curvature-induced kinematic terms and computes $m_\star$ from the Taylor coefficients of $H(t)$. This provides a geometric description of an entanglement cascade, in which new subsystems can join multipartite entanglement only after one or more interaction steps. A symmetry-protected three-qubit model realizes $m_\star=2$, and a three-mode bosonic example shows that genuine tripartite entanglement can arise even when the two newly formed reduced pairs remain separable.
Read the full abstract
Cryptography and Security
Detokenization Leaks: Reconstructing Local LLM Outputs From Cache Traces
We present a new attack that reconstructs the text generated by locally hosted LLMs by observing CPU cache activity during detokenization. Unlike prior attacks that rely on deployment-specific assumptions, such as shared data memory, CPU offloading, or Mixture-of-Experts architectures, our approach targets the detokenizer, a component used in default LLM inference pipelines. To obtain clean signals, we use Flush+Reload on shared tokenizer code to detect when decoding occurs, which lets us perform Prime+Probe at the right moment and isolate token-dependent cache activity. We then apply a clustering-and-language-model pipeline to recover text from noisy cache observations. We evaluate the attack across multiple datasets, hardware platforms, inference frameworks, and model families, and show that it can recover semantically accurate outputs from real-world local LLM deployments, including agentic systems. This vulnerability is particularly significant because the most widely used tokenizer implementations are susceptible to the attack and are embedded in many popular local LLM products and agent frameworks, including systems such as OpenClaw (which we demonstrate), substantially broadening the practical attack surface.
Read the full abstract
Analysis of PDEs
Singular Extremal Solutions on Thin Ellipsoids with Varying Nonlinearities
Let $N=m+1$ and consider the thin ellipsoid \[ \Omega_\varepsilon=\{(y,x_N)\in\mathbb R^m\times\mathbb R:\ |y|^2+\varepsilon^{-2}x_N^2<1\}. \] We prove that in every sufficiently large dimension, for every sufficiently small $\varepsilon>0$, there exists a smooth positive, strictly increasing, strictly convex, superlinear nonlinearity $f_\varepsilon$ for which the extremal solution in $\Omega_\varepsilon$ is an unbounded $H^1_0$ solution. Combining this result with Dancer's thin-domain regularity theorem for the Gelfand nonlinearity $f(t)=e^t$, we obtain on the same sufficiently thin ellipsoids a bounded Gelfand extremal solution and an unbounded extremal solution for another nonlinearity. Thus, in this two-part sense, Brézis' Open Problem~6.1 is resolved in every sufficiently large dimension.
Read the full abstract
Quantum Physics
Efficient Representation of multicategorical local Hilbert spaces: nonlinear Restricted Boltzmann Machines to Kolmogorov-Arnold Networks
Neural network quantum states (NQS) for representing spin-$\frac{1}{2}$ systems are typically built from Multi-Layer Perceptrons (MLPs) with binary visible variables, in which the input spins are first combined through linear affine maps before more expressive nonlinear transformations are applied. The standard way for representing multi-categorical systems, namely spin-1 and spin-2, or the q-state quantum Potts systems with more than two local degrees of freedom, is a unary or one-hot encoded MLP architecture. Following a few similar studies, I observe that while the one-hot construction is the mathematically faithful encoding for systems with more than two local states such as spin-$S$ models with $S > 1/2$ or $q$-state Potts models, its parameter count grows with the number of local states, and the resulting optimization landscape can hinder convergence. I find numerically that allowing $\log \Psi$ to be a nonlinear function of the raw multi-valued spin variable is a natural categorical generalization: it preserves the labelling freedom of the local basis and reproduces the one-hot model with strictly fewer parameters, often with improved trainability. I first benchmark this idea on shallow Restricted Boltzmann Machines (RBMs), equipping them with several distinct nonlinear connections, for spin-1, 2, and 3 Heisenberg chains. I then turn to Kolmogorov-Arnold Networks (KANs), where each edge carries a learnable univariate nonlinearity, and show that they provide a strictly more expressive realization of the same principle. Finally, I demonstrate that this framework captures the critical behaviour of the quantum Potts Hamiltonian, recovering its phase transition.
Read the full abstract
Machine Learning
Tracking the Moving Frontier: Long-Short Term Advantage Estimator
Group-based RLVR methods estimate advantages by repeatedly sampling multiple trajectories for each prompt, making long-horizon agent training expensive and discarding useful experience accumulated across iterations. We ask whether historical experience can replace these repeated within-iteration comparisons without directly optimizing on stale trajectories. We introduce Long-Short Term Advantage Estimator (LSTAE), a single-stream RL algorithm that uses history for advantage estimation while updating the policy only with the current rollout. LSTAE maintains a persistent tracker for each task anchor. At the trajectory level (long term), a drift-aware historical baseline tracks the anchor's moving success frontier and measures the relative contribution of each new trajectory. At the step level (short term), a recent state-experience buffer exploits recurrent states to estimate localized action advantages. This two-timescale design converts accumulated experience into multi-granular credit signals, requiring only one rollout per anchor. Across agentic and mathematical reasoning benchmarks, LSTAE matches or improves upon strong group-based baselines while substantially reducing rollout cost.
Read the full abstract
Machine Learning
Data Efficient Sample Selection for In-Context Learning
The In-context learning (ICL) paradigm aids large language models (LLMs) to adapt to new tasks without need for fine-tuning. However, selecting an optimal combination of demonstration examples from a large pool of example subsets is a challenging problem. Existing approaches for selection do not model the complex relationship between ICL samples and downstream LLM performance. They typically perform static task-level selection, choosing subsets once offline, which can fail to generalize to unseen queries. We introduce DearICL (Data Efficient Algorithm for Ranking) ICL samples, a new framework that models demonstration example selection as a subset ranking problem. DearICL employs a non-linear surrogate employing a differentiable sorting objective within a gap-index bandit algorithm. The gap-index based approach enables fine-grained separation of good arms and borderline arms, which is used as an auxiliary objective to train the non-linear surrogate through sufficient sampling of borderline arms, supporting instance-level subset ranking. On exemplar selection benchmarks with open-source LLMs, DearICL achieves 8.08-15.9% accuracy gains over strong linear bandit baselines, with low sample complexity. Code and data: https://github.com/VenkteshV/DearICL.
Read the full abstract
Data Structures and Algorithms
Tight Bounds on the Cost of Adaptivity for the Meyerson Sketch
In online facility location, points arrive one at a time, and the algorithm must either open a facility at the arriving point or route the point to an existing facility. The Meyerson sketch opens a facility at each arriving point with probability proportional to the point's distance to the closest open facility, and requires no state beyond the set of open centers. Due to its simplicity, space efficiency, and strong guarantees against the offline optimum, the Meyerson sketch has become a workhorse of streaming and online clustering. In many such applications, however, the set of open facilities are visible to the process that generates the stream, which can adaptively select future points based on the algorithm's past random choices, voiding its classical guarantees. In this work, we quantify the effect of such adaptivity. We compare an adaptively generated run of the sketch against an oblivious replay, an independent execution, with fresh coins, on the very same generated sequence, and study the adaptivity ratio of expected adaptive cost to expected replay cost. We determine the worst-case ratio in both directions: adaptivity can neither inflate nor deflate the expected cost, or the number of open facilities, by more than an $O(\log\Delta/\log\log\Delta)$ factor, where $\Delta$ is the aspect ratio of the input points (the ratio of the largest to the smallest pairwise distance). This is asymptotically tight as there are deterministic generators on the real line that inflate or deflate the cost by an $\Omega(\log\Delta/\log\log\Delta)$ factor. We show that these robustness guarantees carry over to Meyerson-based sketches for approximate $k$ clustering with sketch size $O(k\,\mathrm{polylog}(n))$.
Read the full abstract
Machine Learning
Towards Unified Multimodal Graph Foundation Model: A Bridge-Router-Adapter Based Approach
Multimodal graphs couple node attributes in different modalities, such as text and images, with relational structure, enabling topological structure and cross-modality attributes to be modeled jointly. Multimodal graph foundation models seek unified representations from such data that transfer across different graph domains and downstream tasks. However, existing methods exhibit two fundamental limitations. (1) Cross-Scope Context Entanglement. They merge scope-specific graph contexts into a unified representation, obscuring their distinctions during multimodal construction. (2) Scope-Ignorant Modality Routing. They route modalities within a fixed graph scope, overlooking how modality relevance varies across neighborhood ranges. To address these challenges, we propose BRAIN, a unified model that focuses on graph context that combines neighborhood scope with modality composition. BRAIN comprises a scope-conditioned Bridge that combines structural information spanning local-to-global neighborhood scopes with different modality compositions; a hierarchical Router that estimates the relevance between the scope and the task, and selects compositions separately within each scope, allowing modality utility to vary with graph range; and a lightweight residual Adapter that further specializes the routed embedding for downstream prediction. BRAIN is trained through multi-graph pretraining followed by task-specific adaptation. Experiments across nine datasets and four task families demonstrate its broad effectiveness, improving node-classification and link-prediction performance by up to 4.73% relative to the strongest baseline, while achieving an average relative improvement of 14.72% across four graph-to-text and two graph-to-image metrics.
Read the full abstract
Machine Learning
Behavioral Cloning Outperforms Entropy-Regularized RL: Critic-Driven Failure of Actor-Critic Methods on Adaptive Tumor Treatment
Adaptive dosing requires policies that reduce tumor burden without excessive toxicity. Learned dosing policies are typically judged against historical or heuristic comparators, which cannot show whether a policy has found the best behavior available. We instead study a three-population tumor-control ODE in which optimal-control analysis fixes the form of a good schedule – bang-bang dosing punctuated by a singular arc – and construct a numerical controller of that form as a proxy for near-optimal behavior. Judged against this reference under a sustained-cure criterion – 200 consecutive days below 5% carrying capacity – Soft Actor-Critic (SAC) trained from scratch never reaches cure. Behavioral cloning (BC) of the reference reproduces it (100% sustained cure, 30/30 seeds), but SAC fine-tuning of the cloned policy destroys it across five entropy coefficients, and TD3 and BC-regularized SAC fail identically; the pattern persists under multiplicative pharmacokinetic action noise. Along curative trajectories the post-collapse critic ranks the collapsed-policy action above the reference action in 96% of states, concentrated in the maintenance phase, and the policy settles into a non-curative adaptive-therapy equilibrium. The reference is what makes this legible: against a heuristic comparator the fine-tuned policy would read as a competent controller rather than a failure.
Read the full abstract
Signal Processing
Computation-Limited Signals in Practice: A Throughput Anomaly in Software-Defined O-RAN PHY Layers
We formalize a performance anomaly in software-based wireless physical-layer (PHY) implementations, a key enabling technology for Open Radio Access Network (O-RAN) systems. For a fixed offered load, the number of PHY symbols required to transmit the load decreases as the nominal bit rate increases because each symbol conveys more bits. Shorter bursts, however, provide fewer opportunities to exploit pipelined modulation, transmission, and demodulation. Processing latency therefore represents a larger fraction of the total delivery time and may cause a nominally faster PHY configuration to deliver lower effective throughput. We call the operating condition in which computational latency governs effective throughput the computation-limited regime. Its most pronounced manifestation occurs when the offered load fits in a single symbol: intersymbol pipelining cannot be exploited, and modulation and demodulation latencies cannot be amortized. Their combined latency may exceed the symbol duration even when transmitter and receiver processing independently meet their real-time deadlines. Because PHY processing complexity often scales superlinearly with the resources allocated to each symbol, a higher-rate configuration may consequently deliver lower throughput than a slower one. We validate this anomaly using processing-latency measurements reported for established software-defined PHY implementations. The results show that PHY processing latency should be incorporated into link-adaptation algorithms for very-high-throughput services in future 6G wireless networks.
Read the full abstract
Computer Vision
TransNormal-2: Geometry-Grounded Rectified Flow with Edge-Aware Decoding for Precise Normal Estimation
Diffusion-based models enable monocular geometry estimation, yet their pixel-space precision is limited by a shared, under-studied error source: VAE reconstruction degradation. The 8x spatial compression in the VAE encoder-decoder degrades surface normals at object boundaries; even encoding and decoding ground-truth normals introduces 1.3–8.5{\deg} of mean angular error (MAE), with edge MAE reaching 2.8x the global MAE. We present TransNormal-2, a FLUX.2-based rectified-flow framework with single-step deterministic inference that addresses this degradation on both sides of the VAE decoder: in how latent predictions are supervised during training, and in how decoded normals are corrected at inference. First, geometry-aware pixel-space losses, including inverse rendering self-consistency, von~Mises-Fisher angular loss, and wavelet edge-aware regularization, complement latent MSE by enforcing spherical normal geometry and diffuse image-formation cues after VAE decoding. Second, a lightweight Geometric Refinement Module (GRM) applies an RGB-guided residual correction to reduce boundary-localized decoding errors without freely rewriting the coarse prediction. On general-scene benchmarks, TransNormal-2 matches or exceeds MoGe-2 on all eight reported metrics while using only 1.4% as many task-specific normal annotations. The gains are clearest for transparent objects, reducing MAE by 4.2{\deg} on ClearGrasp and 3.1{\deg} on ClearPose over the strongest prior baselines. Code will be released at https://longxiang-ai.github.io/TransNormal-2.
Read the full abstract
Information Retrieval
EviMap: Evidence-Grounded Hierarchical Topic Maps for Exploring Unlabeled Corpora
Research teams and organizations often explore unfamiliar free-text collections, from survey comments and reviews to reports and domain documents, before labels, queries or coding schemes exist. At this stage, the first thematic map shapes what users notice, prioritize and carry into downstream analysis, so it should be trusted only insofar as it can be verified. Existing options force a trade-off between scale and verifiability. Qualitative coding preserves evidence but is slow. Search presupposes a query. Clustering and topic models scale but produce labels users must interpret. One-shot large language model (LLM) summaries are fluent yet difficult to reproduce or audit. We present EviMap, an interactive system providing researchers and practitioners with an auditable thematic overview of such corpora. Guided by model-generated context describing the corpus and hypothesized stakeholder concerns, EviMap extracts within-document evidence phrases and organizes them, rather than whole documents, into a three-level map of aspects, groups and fine-grained topics. Embedding-based clustering narrows the search space for finer semantic judgments by the LLM. Each node traces back to supporting phrase spans, so documents link to topics through evidence they contain and users can audit labels against the original text. Users can start from a top-level corpus map, drill into topics, inspect highlighted evidence in original documents, and combine two topics to find documents discussing both. We demonstrate this workflow across six heterogeneous corpora spanning 2,108 to 101,699 documents, with a comparison against flat and hierarchical LLM baselines. By grounding every label in verbatim source spans, EviMap makes a topic map not just readable, but verifiable. Code, demo video, and interactive dashboard are available at https://github.com/zhiyintan/EviMap.
Read the full abstract
Computation and Language
ECOKV: Geometry-Aware KV Cache Eviction via Complementary Diversity Metrics
Although multimodal Large Language Models (MLLMs) excel in diverse tasks, their scalability remains limited by the memory and computational overhead of KV cache storage. Recent KV cache eviction approaches incorporate a cosine similarity-based diversity metric with importance metrics to selectively retain critical key-value pairs. However, cosine similarity involves normalization that discards magnitude information, and it often yields uniformly high similarity values across layers due to the anisotropy property of hidden representations. In our study ECOKV, we rigorously deconstruct the capabilities of existing diversity metrics. Moving beyond simple measurement, we propose a geometry-aware composite metric that jointly leverages Euclidean distance and cosine similarity to capture token diversity from complementary perspectives. Furthermore, we use these two metrics to estimate the redundancy level of each attention head, allowing adaptive weighting between diversity and importance scores during token selection. Finally, we demonstrate that the observation window commonly employed to preserve recent tokens can be substantially reduced, thereby allocating more cache capacity to informative tokens and yielding consistent improvements. Extensive experiments demonstrate that ECOKV achieves state-of-the-art performance under various compression ratios and can be seamlessly integrated with existing KV cache eviction methods. We further analyze the relationship between importance and diversity, and examine redundancy patterns across layers and attention heads.
Read the full abstract
Human-Computer Interaction
The Effectiveness of Virtual Patient Simulation Versus Peer Simulation in Providing Sexual Counseling During Pregnancy: A Randomized Controlled Trial
Although sexual health counseling is one of the important responsibilities of healthcare professionals, effective educational methods are needed to develop students' counseling skills in this field. The aim of this study was to compare the effectiveness of virtual patient simulation and peer simulation methods in developing sexual counseling skills during pregnancy among nursing faculty students. This randomized controlled trial included 51 participants assigned to one of three groups: virtual patient simulation, peer simulation in a virtual environment, or face-to-face peer simulation. In the study, all groups received face-to-face theoretical instruction on sexual counseling during pregnancy. Following the theoretical training, students participated in virtual patient simulation, peer simulation in a virtual environment, or face-to-face peer simulation practices according to the groups to which they were assigned by randomization. Outcome measures included the Sexual Attitudes and Beliefs Scale, the Student Satisfaction and Self-Confidence in Learning Scale, and the Sexual Counseling Skills Evaluation Form. It was determined that the participants' SABS scores decreased significantly after the intervention. According to the results of the mixed repeated-measures ANOVA, the effect of time was statistically significant. However, the group effect and the group x time interaction were not significant. There was no difference between the groups in terms of Satisfaction in learning, Self-confidence in learning, and Skill scores. This study showed that different simulation methods used in sexual counseling education during pregnancy were effective in reducing students' negative attitudes and beliefs and provided similar educational outcomes. Therefore, virtual patient and peer simulations may be recommended as feasible approaches for improving sexual counseling skills.
Read the full abstract
Numerical Analysis
Dilation-Covariant Hankel Pencils for Multiscale Recovery of Sparse Mellin Spectra
We study sparse Mellin spectral recovery from geometric samples. Dilation covariance is shown to generate the Hankel structure directly, rather than only after reduction to a classical exponential-sum model. For complex exponents, incommensurate sampling scales remove the logarithmic aliasing that persists at a single scale. A second recovery channel is obtained from the minimal Hankel pencil: contour winding counts spectral nodes regionally, while Rouché-type bounds certify isolated nodes and unresolved clusters under noise. The associated contour margin has an explicit degeneration rate as the sampling ratio approaches one. Numerical experiments show that the auxiliary scale improves both identifiability and resolution, and that certified contour counts can remain reliable beyond the regime of accurate pointwise recovery.
Read the full abstract
Computer Vision
FSAN: Flow State Attention Network for Aerodynamic Prediction
Accurate aerodynamic prediction is critical for designing fuel-efficient and safe transportation systems such as aircraft and automobiles, yet traditional computational fluid dynamics (CFD) simulations remain computationally expensive and expertise-intensive, severely limiting their use in iterative design and real-time analysis. Existing deep learning surrogates suffer from two major limitations: (i) they are evaluated on datasets with narrow flow-condition ranges, leaving their performance under complex flow conditions undemonstrated; (ii) they treat global flow conditions as a single vector injected uniformly across all surface points, ignoring that different geometric regions experience distinct local flow phenomena, which degrades prediction accuracy under complex flow conditions. To address these limitations, we propose the Flow State Attention Network (FSAN). FSAN separately encodes point cloud and flow conditions, then partitions the geometry into multiple flow states via learnable soft assignments, and uses flow features to update these state representations, which in turn influence point cloud features through state changes. This enables fine-grained, state-specific interaction between geometry and flow information. Extensive experiments on two well-recognized aerodynamic benchmarks demonstrate that FSAN achieves the highest accuracy among the methods compared in this work at a higher computational cost. On Emmi-Wing, FSAN reduces the Relative L2 (REL-L2) error by over 20% compared to the strongest baseline (Transolver), and on DrivAerNet++, it achieves a 10% reduction compared to the strongest baseline (AdaField). These results establish FSAN as a promising neural surrogate on public benchmarks with diverse flow conditions and geometries.
Read the full abstract
Differential Geometry
Nonnegative curvature on vector bundles over ${CP}^3$ and its blow-up
Every real vector bundle of rank at least seven over $CP^3$, or every real vector bundle of rank at least eight over $CP^3\#CP^3$, admits a complete metric of nonnegative sectional curvature on its total space. The corresponding unit sphere bundles also admit nonnegative sectional curvature. These are fixed-rank results: no additional trivial summand is required, and there is no restriction on any characteristic class. The proof combines low-dimensional bundle classification with the Grove–Ziller lifting theory over $S^4$ and $CP^2$.
Read the full abstract
Dynamical Systems
Equidistribution Measures of infinite entropy for Transcendental Functions
In the 1980s Lyubich and Freire-Lopes-Mañé proved that for any rational function of degree d \geq 2, both preimages and periodic points equidistribute to the unique measure of maximal entropy log(d). Their results provide a fundamental understanding of the dynamics of iterated rational functions, and have since been generalized to many different contexts, including classes of higher-dimensional polynomial and rational maps. In the current paper we depart from the algebraic category and aim to prove analogous statements for transcendental functions in the complex plane, which have infinite topological entropy. We introduce two different methods for constructing invariant measures in the transcendental setting, namely via embedded symbolic dynamical systems and via transfer operators associated to suitably chosen weights. In the latter case we isolate three properties of the weights -normality, tightness, and irreducibility- which together imply convergence to an invariant measure. We provide examples for each method, given by three classes of transcendental entire functions: disjoint-type maps, strongly polynomial-like maps, and a class of maps inspired by Baker's construction of multiply connected wandering domains and by Bishop's construction of Julia sets of Hausdorff dimension 1, which we call Baker-Bishop maps. For each of these classes we prove that with respect to carefully chosen weights, preimages equidistribute to an invariant mixing probability measure of infinite entropy. For Baker-Bishop maps and disjoint-type maps we also prove equidistribution of periodic points. In contrast to the rational setting, the measures we construct are not unique: by varying the weights one obtains infinitely many distinct measures.
Read the full abstract
General Physics
Algebraic solutions for propagation of elastic SH-waves through an isotropic inhomogeneous geometry
The algebraic structure of the equation governing the propagation of an elastic SHwave through an isotropic inhomogeneous layer surrounded by two homogeneous half-spaces is revisited. It is shown that the problem under a quartic variation has a hidden sl(2) symmetry based on which quasi-exact solutions are available via a completely analytical approach. This is of particular interest as the approach proposed for the corresponding tri-confluent Heun equation can be generalized to some other classes of geometries as well and the complicated considerations in case of Heun equations are absent.
Read the full abstract
Machine Learning
Assessing Covariate-Informed Grid Load Forecasting with a Time-Series Foundation Model
Modern power systems are growing increasingly complex as they integrate diverse generation sources to meet rising demand, making accurate load forecasting challenging. Recent advances in time-series foundation models (TSFMs) resulted in promising performance in zero-shot univariate load forecasting tasks. However, real-world load forecasting often involves multiple target variables and requires the integration of exogenous variables, raising important questions about the utility of TSFMs in realistic settings. In this study, we position Chronos-2, a recently developed model by Amazon, as a representative multi-channel TSFM that supports univariate, multivariate, and covariate-informed forecasting, and conduct a systematic investigation of how such models can be used for real-world load forecasting. While prior work has evaluated Chronos-2 on a limited number of energy-related tasks in a zero-shot setting, its performance relative to established task-specific deep learning models and its behavior when adapted using task-specific historical data remains insufficiently understood. In this work, we evaluate Chronos-2 on two real-world utility datasets, ISO New England and ENTSO-E, and benchmark it against widely used task-specific deep learning models. Our results show that Chronos-2 benefits substantially from task-specific fine-tuning and achieves strong short-horizon forecasting performance, but its zero-shot accuracy lags behind task-specific models and its forecasting error grows more rapidly with increasing forecast steps. Overall, this study provides a detailed characterization of the strengths and limitations of TSFMs such as Chronos-2 in grid load forecasting and offers practical insights into how a pretrained TSFM can be effectively adapted for operational load forecasting applications.
Read the full abstract
Chemical Physics
Out-of-Distribution Inverse Design of Elastic Networks with Differentiable Graph Neural Network Molecular Dynamics
Machine-learning-based inverse design can accelerate the discovery of materials with targeted properties, but conventional structure–property models often require large training datasets and generalize poorly beyond their training distribution. Here, we present a differentiable inverse design framework based on a graph neural network molecular dynamics simulator. By combining a short dynamical initialization with physics-based refinement during simulation, the framework enables optimization well beyond the conditions represented in the training data. Using disordered elastic networks, we show that a simulator trained only on non-auxetic systems with Poisson's ratios between 0.1 and 0.4 can design strongly auxetic networks with values as low as -0.3. The framework also produces stress–strain responses outside the range observed during training, creates localized mechanical defects absent from the training data, and generalizes across system size, enabling optimization of networks tested up to 5000 nodes despite being trained only on networks with fewer than 200 nodes.