Peer Revenue Now accepting positions

Papers

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

All of arXiv

Saturday 7 November

Machine Learning

Finite Time Analysis of Risk-Sensitive RL via Noisy Power Iteration

Estimating the principal eigenpair of a positive operator from stochastic evaluations gives rise to a noisy form of normalized power iteration: each step applies a sampled operator estimate and then renormalizes. We establish finite-time stochastic-approximation guarantees for this procedure, working in Hilbert's projective metric, which is well-suited to the multiplicative geometry of Perron–Frobenius eigenproblems. Our main application is risk-sensitive average-cost reinforcement learning (RL) with exponential utility. In this setting the Bellman equations are multiplicative, and value functions are characterized by nonlinear eigenvalue problems rather than additive fixed points. We cast both policy evaluation and control in this framework, obtaining risk-sensitive TD- and Q-learning algorithms that learn from a single Markovian trajectory. Under explicit positivity, mixing, and linear function approximation assumptions, we prove finite-time bounds on the recovered eigenvector and eigenvalue with $\tilde{O}(\epsilon^{-2})$ sample complexity up to problem-dependent constants. The results cover both tabular and linear function approximation regimes, and through the duality between exponential utility and KL-robust control yield finite-time guarantees for KL-robust average-cost RL as well.
Read the full abstract

All authors Waqar Mirza, Yashaswini Murthy, Laixi Shi, Eric Mazumdar, Adam Wierman

Computer Vision

Tiresias: Towards Language-Conditioned World Models

World models enable planning through internal simulation, while natural language provides intuitive task specification. Yet, language-conditioned world modeling remains fragmented across domains with disparate action spaces, datasets, and evaluation protocols. We introduce LCWM, a shared task that maps an initial agent-view observation and a language instruction to an action sequence that fulfills the instruction. Its accompanying benchmark, LCWM-Bench, integrates 15 datasets across navigation, robotics, 3D gaming, and driving under a shared action interface and evaluation protocol, pairing 142,987 episodes with 428,961 instructions in three styles. We further propose Tiresias, an agent that plans via selective imagination: a language-conditioned diffusion policy generates candidate trajectories, a trajectory ranker shortlists them, and a video-pretrained world model imagines their visual consequences. A visual ranker evaluates these rollouts, complementing trajectory scores to select the final plan. Tiresias achieves competitive planning and imagination performance on LCWM-Bench. Ablations demonstrate the complementary benefits of trajectory ranking and visual verification, while experiments with alternative proposal methods show that the trained rankers remain effective beyond our diffusion policy. Together, LCWM-Bench and Tiresias provide a unified evaluation setting and an effective starting point for studying language-guided planning through world models.
Read the full abstract

All authors Fengyi Wu, Yifei Dong, Yetong Sha, Lingdong Kong, Qiyu Hu, Bohan Xiong, Qi Dai, Alexander G. Hauptmann, Zhi-Qi Cheng

Algebraic Geometry

Metric nefness of direct image sheaves under weakly semi-stable Kähler morphisms

We investigate the nefness of the direct image sheaves of relative adjoint bundles under weakly semi-stable Kähler morphisms between complex spaces. We show that the direct image sheaves are locally free and that their tautological line bundles are metrically nef, strengthening the previously known results on torsion-freeness and certain weak positivity.
Read the full abstract

Artificial Intelligence

A Computational Map of the Disaster Mental Health Research Landscape

When disasters strike, they not only damage infrastructure, buildings, and the economy, but also affect the mental health of affected populations. Research on the mental health effects of disasters has expanded rapidly and has become an important, well-funded area of psychology, public health, and related social sciences. However, the overall structure of this research landscape has not been systematically characterized. Using large language models, we conducted a review of all articles related to terms 'disaster' and 'mental health' indexed in PubMed. 7,859 articles' metadata were retrieved using the NCBI API, and then Claude Sonnet was used to extract structured variables from the abstracts. The results indicated that earthquakes and epidemics were the most commonly studied disasters while many other disasters such as heat and cold waves have received less attention. PTSD, depression, and anxiety were well investigated but many other mental health issues (e.g., aggression, suicide) are relatively ignored. Despite enjoying relatively large sample sizes (median = 452, IQR = 142 - 1,568), the field rarely utilizes machine learning algorithms. Hypothesis significance testing, reviews, qualitative methods, and structural equation modeling are respectively most common methods while agent-based modeling, machine learning, casual inference, and NLP each appear in less than 1% of articles. Studies are limited to a small number of countries, particularly the United States, Japan, and China, while many disaster-prone regions remain substantially underrepresented. This LLM-assisted computational map provides a broad overview of the disaster mental health literature while highlighting important methodological gaps, with particular emphasis on the underutilization of computational methods that could help advance the field.
Read the full abstract

Machine Learning

From Task Affinities to Optimal Allocations: Probabilistic Graphical Modelling for Supervised Finetuning

Supervised fine-tuning over a large pool of heterogeneous tasks is often computationally expensive, leading practitioners to train language models on a fixed training budget of carefully chosen SFT tasks. Thus, achieving strong downstream performance hinges on selecting an effective subset of tasks, or equivalently, identifying an optimal task mixture under a fixed budget constraint. In practice, supervised finetuning mixtures are often determined using simple heuristics (e.g., uniform or size-proportional sampling) that ignore task interactions, which can hurt transfer and waste budget on redundant task sources. We introduce TASKMIXPGM, a framework for learning continuous task mixtures through energy-based modelling over SFT tasks. Each task forms a node in a Markov random field, with unary potentials capturing task utility and pairwise potentials modelling inter-task relationships via behavioral divergences between predictive distributions of single-task fine-tuned models. Optimizing this objective yields task mixtures that balance utility, coverage, and redundancy. We further show that the induced set function defined over the set of such tasks is weakly submodular under budget constraints, enabling approximation guarantees for discrete task selection. Empirically, we provide a fixed, benchmark-independent mixture that improves over heuristic mixing across multiple model families and different training budgets. TASKMIXPGM also uncovers interpretable structure in task interactions, revealing complementarities and redundancies that inform effective SFT mixtures.
Read the full abstract

All authors Prateek Chanda, Saral Sureka, Krishnateja Killamsetty, Ganesh Ramakrishnan

Statistical Methodology

Causal Discovery Grounding and the Naturalness Assumption

Standard causal discovery postulates that structure can be described by a directed acyclic graph (DAG), which can be partially reconstructed by using the faithfulness principle. According to this principle, conditional independencies in the observational distribution correspond exactly to those encoded in the unknown DAG. However, such algorithms are challenged by the fact that, without even stronger assumptions, approximate failures of faithfulness are empirically undistinguishable from exact ones. This leaves an inferential gap which can lead to major implications to causal effect identification. We introduce a different perspective, not reliant on exact independencies. By considering the specialized problem of discovering whether unmeasured confounding can be ruled out for a particular cause-effect pair, we introduce a principle we call naturalness, in which the strength of observational associations bounds the strength of causal dependencies. The bounding function is unknown, with degrees of freedom not present in faithfulness. We model uncertainty about this function using a Bayesian framing which allows for the use of experiments from separate studies, grounding an adaptive causal discovery principle on historical experimental data. We demonstrate its usefulness in partial identification of causal effects, and discuss how the promising results can motivate new principles for general causal discovery.
Read the full abstract

All authors Guilherme P. Soares, Andreas Koukorinis, Sander Beckers, Jake Fawkes, David S. Watson, Ricardo Silva

Operator Algebras

Weighted Logarithm Laws in non-commutative Probability

We prove analogues of weighted laws of iterated logarithm and laws of logarithm for non-commutative martingales and independent random variables respectively. First we obtain a weighted LIL for non-commutative martingales, extending the sharp unweighted result of arXiv:2509.22037v2 to the case for polynomial weights, then by approximation to a broader class of functions. Then we proved a weighted law of logarithm for triangular array of weights and replacing the coefficient in the denominator by regularly varying functions with index $\alpha$. We show that the normalised sums converge to 0 for $\alpha> 1/2$ and for the case $\alpha=1/2$, we obtain the sharp upper bound with constant $\sqrt{2}$.
Read the full abstract

Machine Learning

Pruning DNA Sequence

Data pruning can substantially reduce training cost by removing examples that contribute little additional learning signal, but it has not been studied for DNA language-model pretraining. We introduce Sequence Pruning (SP), a dynamic pruning framework that removes genomic sequences as they become easy for the model to predict. SP uses per-sequence pretraining accuracy computed directly from the model's forward pass, requiring no separate scoring model or additional pass over the corpus. We evaluate SP across five pretraining settings spanning two corpora, three model families, and both MLM and CLM objectives, with downstream evaluation on five genomic benchmarks. SP reduces pretraining data processing by up to 96% while largely preserving downstream performance. At moderate savings, pruning improves performance by up to 1.2 MCC points over full-data pretraining; at the highest savings, pretraining uses only 4% of the baseline GPU hours with little performance loss. SP also outperforms the dynamic pruning methods InfoBatch and BPAS across most comparisons. Beyond accuracy, our framework supports alternative criteria such as perplexity and EL2N, although accuracy yields the strongest downstream performance at matched savings. These results show that genomic pretraining does not require repeatedly processing all available sequences and that dynamic sequence-level pruning can substantially improve its data and computational efficiency. Code is available at https://github.com/FaiasPromit/pruning-dna-sequence.
Read the full abstract

Artificial Intelligence

A Topological Encoder Decoder Framework for Temporal Graph Learning

Most temporal graph learning methods reduce future prediction to discriminative inference over past interactions, focusing on edges or node labels for a largely persistent set of nodes. This view breaks down when new nodes and edges appear, since the future graph becomes a new object with changing size, composition, and structure. We address this gap by framing temporal graph prediction as an inverse topology problem. Instead of predicting edges directly, we first predict a multiscale topological descriptor of the future graph and then reconstruct a plausible future snapshot that realizes this descriptor under inductive constraints. This approach makes global structure and node churn explicit prediction targets and produces forecasted graphs that downstream tasks can evaluate without retraining. Across 14 temporal graph datasets, we evaluate our method, TOPOGED, on node, link, and graph property prediction and compare it against state-of-the-art temporal graph models. TOPOGED significantly improves node-level forecasting accuracy over the strongest baseline, with a macro average of 0.49 versus 0.05. It also outperforms baselines in 60% of graph-structure metric evaluations and substantially increases macro-average edge prediction metrics, from near-zero to 0.12. Our results show that topology-guided graph forecasting can predict inductive future snapshots whose structure supports multiple downstream evaluations.
Read the full abstract

All authors Ronan Buck, Kiarash Shamsi, Tran Gia Bao Ngo, Astrit Tola, Baris Coskunuzer, Cuneyt Gurcan Akcora

Category Theory

Generalized near-group tensor categories

Let $\mathcal{D}$ be a finite tensor category with exactly one isomorphism class of simple objects which are non-invertible. Semisimple tensor categories $\mathcal{D}$ with this property are called near-group categories, and have been studied by Evans, Gannon, Siehler and others. In this paper, we study such categories in the non-semisimple setting, with the additional assumption that the non-invertible simple object is projective. Specifically, we consider the case when $\mathcal{D}$ admits a fiber functor and has a unique semisimple abelian block (in the general case that $\D$ is not semisimple). We describe the structure of such categories and their corresponding Hopf algebras, and construct examples of such categories with a cyclic Picard group. We obtain additional results when the Frobenius-Perron dimension of the unique non-invertible simple object is a prime number.
Read the full abstract

Econometrics

Asymptotic Theory of Integrated Pairwise-Difference Functionals of Kernel Estimators, with Applications to Treatment Effect Heterogeneity

This paper develops asymptotic theory for integrated pairwise-difference statistics constructed from kernel estimators, with applications to treatment effect heterogeneity. The statistics integrate absolute or positive-part differences between kernel estimates evaluated at pairs of points. Their two-point structure falls outside existing one-point $L_1$-type functional theory. We establish asymptotic normality using a Hájek-type projection of a Poissonized statistic and develop feasible centering and variance estimators. We consider two applications. First, we test whether treatment effects vary with one covariate block while holding another fixed. Second, we establish root-$N$ inference for the directional Gini mean difference of the conditional treatment effect function, which measures heterogeneity in outcome units. The framework also accommodates tests of monotone effect modification and estimated pseudo-outcomes. Simulations corroborate the asymptotic theory and show good finite-sample performance. An application to 401(k) eligibility and net financial assets illustrates the usefulness of the proposed tests.
Read the full abstract

Computer Vision

From Reflection to Co-Evolution: Self-Evolving Agents with Locally Verified Memory

LLM agents have made impressive progress in coding and mathematical reasoning, motivating growing interest in learning from long-horizon interaction. Memory supports this process, but plausible lessons can be redundant or reinforce inappropriate actions. Can a memory's source decision provide a practical test for whether to retain it? We localize consequential trajectory steps, extract candidate memories, and replay each source decision with and without its candidate under the same context. This reveals two properties: most candidates barely shift the probed distribution, and both successful and failed source actions can be reinforced or suppressed. These observations motivate Anchor, which tests shift magnitude and source-role alignment before retention. Admitted memories guide later trajectories, enabling experience-driven improvement under a frozen policy. On held-out ALFWorld and WebShop tasks, Anchor improves success rate and mean score by up to 19.9% and 17.2%, respectively, relative to the strongest baseline for each backbone. With Qwen3.5-27B, it retains 41% fewer ALFWorld cards than unfiltered retention and outperforms all four WebShop cost-comparison baselines while using fewer tokens.
Read the full abstract

All authors Xinxiang Yin, Xinglang Zhang, Haodong Chen, Zhiqing Cui, Weidong Tang, Yuxuan Wang, Junrui Wang, Jingyi Zhang, Jianfeng Zhang, Jian Qian

Geometric Topology

Polynomial conjugacy invariants of pseudo-Anosov maps derived from train track representatives

We investigate two kinds of polynomials associated with pseudo-Anosov maps. Each kind consists of three polynomials that are conjugacy invariants of pseudo-Anosov mapping classes and these polynomials contain more information than pseudo-Anosov stretch factors. The first kind is defined by the train track representatives. The second kind is obtained from the first kind by incorporating additional data that encode the transitions among the unpunctured singularities of the invariant foliations. The Birman-Brinkmann-Kawamuro polynomials coincide with those of the second kind. As an application, we prove that the Birman-Brinkmann-Kawamuro polynomials are conjugacy invariants of pseudo-Anosov mapping classes which is also proved by Dodge-Kawamuro, independently.
Read the full abstract

Machine Learning

Quantitative Assessment of Crystal Structure Prediction

Crystal Structure Prediction (CSP) aims to predict a material's 3D structure from its chemical composition, with applications from energy storage to biomedical devices. Despite its importance, existing evaluation frameworks suffer from key limitations: they rely on threshold-sensitive or incomparable metrics, focus predominantly on geometric performance while overlooking physical plausibility, and fail to account for polymorphism, where a chemical composition can crystallize into multiple stable structures. In parallel, recent Symmetry-Informed CSP (SICSP) methods often report results alongside standard CSP baselines while receiving crystallographic templates at inference, which constrain the search space and make SICSP a fundamentally different task. To address these limitations, we introduce an Orientation-invariant, Polymorph-Aware, Large Benchmark (OPAL-Bench) that evaluates CSP and SICSP as separate tasks. We evaluate 7 CSP and 2 SICSP methods under OPAL-Bench across 4 datasets using threshold-free geometric and physical distance metrics. Our results show that no method dominates across all metrics, and that geometric performance is not strongly correlated with physical plausibility. We also release OPAL-CSP, an SE(3)-equivariant flow-matching baseline for the CSP track that performs competitively on geometric metrics and ranks among the best on physical ones. By unifying geometric and physical metrics under a polymorph-aware protocol, OPAL-Bench enables a more comprehensive assessment of generative quality in CSP.
Read the full abstract

All authors Sergio Rincón, Gabriel González, Nicolás Andrade, Rafael Velásquez, Sebastián Ojeda, Juanita Puentes, Nicolás Aparicio, Paula Cárdenas, Pablo Arbeláez

Computer Vision

RefineSplat: Reconstruction-Guided Object-Centric Gaussian Splatting for Open-Vocabulary Querying

Open-vocabulary 3D scene understanding aims to segment and retrieve arbitrary objects in 3D scenes using natural language queries. Most existing methods typically follow a splat-then-segment paradigm, where 3D Gaussian Splatting is first reconstructed and then augmented with labels or language features for subsequent segmentation. However, this pipeline often suffers from view inconsistency, ambiguous object boundaries, and weak object-level correspondence, especially in complex static and dynamic scenes. Building on our prior Segment-then-Splat framework, we present RefineSplat, a reconstruction-guided extension that retains its object-centric paradigm: it first tracks objects in 2D and then reconstructs each object with a dedicated set of 3D Gaussians, preserving explicit object-Gaussian correspondence for unified open-vocabulary segmentation and retrieval. To improve robustness, we further close the loop between 2D tracking and 3D reconstruction. Since off-the-shelf video trackers can produce unreliable masks that propagate errors into object reconstruction, we introduce a reconstruction-guided feedback mechanism that refines object tracking using rendered masks and multi-view consistency. Moreover, instead of assigning each object a single language embedding, we propose a viewpoint-aware embedding assignment strategy, where each object maintains multiple view-clustered embeddings to capture view-dependent semantic cues. We also introduce Blender-OVS, a synthetic benchmark of static and dynamic sequences with exact object meshes and point clouds, enabling direct evaluation of extracted geometry. Experiments on static and dynamic benchmarks show improved rendered-view segmentation and lower object-level Chamfer distances, with competitive scene reconstruction quality.
Read the full abstract

All authors Yiren Lu, Disheng Liu, Sreya Srinidhi, Yunlai Zhou, Yiran Qiao, Chaoda Song, Tuo Liang, Jing Ma, Huan Wang, Yu Yin

Chemical Physics

Can AI Agents Reliably Optimize Computational Chemistry Codes?

Artificial intelligence (AI) approaches have opened new possibilities in computational and theoretical chemistry. AI agents can now execute computational workflows autonomously, and their use in scientific software development represents another promising direction. Herein, we present a general strategy for the controlled optimization of scientific software using a dual-agent harness. By using separate worker and controller agents for optimization and evaluation, respectively, we find that this approach can reduce computational costs while preserving numerical accuracy for the computational chemistry methods examined. We apply this strategy to TIP4P water simulations in LAMMPS, as well as to Coulomb and exchange matrix construction and linear-response time-dependent density functional theory (TDDFT) calculations in PySCF. Among these applications, the largest speedups are obtained for TDDFT. We then extend this strategy from optimizing established packages to developing efficient parallel CPU- and GPU-accelerated implementations of an initially serial molecular dynamics code, resulting in speedups of two to three orders of magnitude. Finally, we apply the protocol to open-ended algorithm development by constructing an adaptive self-consistent field approach that coordinates the use of multiple solvers. We find that the resulting approach can converge some challenging systems and occasionally reach lower-energy solutions than those obtained with the standalone solvers tested. These applications demonstrate that a strategy developed for improving code efficiency may also support the systematic development of computational methods. Overall, our benchmark study provides an example of AI-assisted advances in computational chemistry.
Read the full abstract

Computer Vision

Vision Language Models for Primary Structure Threat Identification in Wildland Urban Interface Fires

Wildland urban interface (WUI) fires increasingly threaten communities and infrastructure, motivating the need for rapid visual assessment. Recent research has focused primarily on detecting fire and assessing post-event damage. However, structure-level fire threat identification during active WUI fire events remains underexplored. This study evaluates Qwen3-VL-8B-Instruct for this active-fire decision using two separately scored tracks: 1,824 synthetic images from 491 designed WUI scenarios and 45 frames from a controlled-burn experiment conducted in collaboration with Columbus Fire & EMS Department. Under zero-shot prompting, the model returned valid target fields for 1,808 synthetic images and 39 controlled-burn frames. Accuracy among valid outputs was 94.0% for the synthetic track and 69.2% for the controlled-burn track, compared with a 16.0% random-selection baseline for the latter. On the synthetic track, the model generally identified the intended target across varied designed WUI scenarios when the relevant fire-spread cues were visually distinct. Controlled-burn performance was well above random selection, but errors showed difficulty interpreting spatial relations in complex scenes. The dual-track protocol provides scalable stress testing in designed scenarios and a controlled-burn reference for assessing model behavior under physical fire conditions. Future work can use the synthetic data for VLM post-training and then evaluate the adapted models on larger and more diverse real WUI fire datasets.
Read the full abstract

Computer Vision

PhysDNet: Physics-Driven Gradient Amplification for Real-Time Image Dehazing

Modern dehazing networks are often computationally expensive and overfit to synthetic haze, limiting real-time deployment. This underscores an urgent demand for efficient, high-fidelity restoration methods that can operate on resource-constrained edge devices in practical, dynamic environments. We introduce PhysDNet, a lightweight multimodal architecture that fuses NIR imagery with sparse LiDAR depth and is trained with a physics-constrained dual-head objective. Differentiating through a fixed Koschmieder inversion induces density-adaptive gradient amplification, acting as an implicit resource allocation mechanism that concentrates learning on heavily degraded regions without increasing inference cost. PhysDNet delivers superior restoration with minimal capacity and enhanced training efficiency, facilitating real-time deployment on edge devices without increasing inference overhead. On the Seeing Through Fog benchmark, PhysDNet outperforms competitors by 1.67 dB at 5.2 times lower compute while recovering 54.9% of fog-induced detection loss. Ablations and cross-domain evaluations show that physics-guided training improves generalization, including under non-homogeneous haze that violates the Koschmieder assumption. PhysDNet further demonstrates compatibility with fixed-function NPUs, running at 29/45/99 FPS at <4 W across L/M/S variants. Code, deployment scripts and pre-trained checkpoints are available at https://github.com/Heatdh/PhysDnet.
Read the full abstract

All authors Rayen Dhahri, Lei Sun, Victor Javier Kartsch Morinigo, Alice Pagnoux, Luca Benini, Danda Pani Paudel, Luc Van Gool

Machine Learning

CHARM: Certified Homology-Aware Reverse-scale Mapping via Diffusion-Guided Quotient Towers

Conventional dimensionality-reduction methods optimize geometric fidelity but do not return auditable homology certificates or identify the input observations associated with nontrivial classes. Direct high-dimensional persistent-homology computation is costly, and its representatives are rarely linked to a shared planar view. We introduce Certified Homology-Aware Reverse-scale Mapping (CHARM), which augments two-dimensional coordinates with algebraically certified, sample-traceable structural records. Diffusion-guided cover proposals and homotopy-safe contractions compress the input neighborhood flag complex. A Cover Quotient Tower supports cross-scale representative tracking and exact $H_k$ certification when algebraic tests complete. Reverse-scale layout connects quotient-chain preimages to the same coordinates used for geometric exploration. Contraction preserves the input neighborhood flag-complex homotopy type; the quotient tower inherits a multiplicative 3-interleaving with coarse-hop Rips, and conditional distortion bounds connect it to original Euclidean Rips. In a fixed-seed benchmark of 22 datasets (41,637 samples), CHARM achieves the highest share of first- or second-place scores across 15 metrics. On nonsynthetic data, it ranks first or second on every metric and first on all four $H_0/H_1$ persistence-diagram metrics. CHARM couples strong nonsynthetic geometric performance with an auditable structural layer.
Read the full abstract

Solar and Stellar Astrophysics

Evidence for Ambipolar Diffusion from Ion-Neutral Velocity Drift in HH 212

We present the Atacama Large Millimeter/submillimeter Array Band 7 observations of the hyperfine structure of C$^{17}$O (3-2) and the H$^{13}$CO$^{+}$ (4-3) molecular line emissions of the Class 0 protostellar system HH 212 at $\sim$ 44 au spatial resolution. Radiative transfer models interpreted in the framework of non-ideal MHD of the envelope allow us to estimate an envelope mass of $M_{env}\sim0.12\,M_\odot$ and a mass infall rate of $\dot{M}_{env}\sim4.4-5.7\times10^{-6}\,M_\odot\,\mathrm{yr}^{-1}$, consistent with values reported in the literature. The models also reveal a systematic radial velocity drift between the neutral tracer (C$^{17}$O (3-2)) and the ionized species (H$^{13}$CO$^{+}$ (4-3)) of $0.1-0.3$ km s$^{-1}$, exceeding the spectral resolution of 0.105 km s$^{-1}$ and consistent with possible ambipolar diffusion in the envelope. The ambipolar diffusion fraction, $f_{AD}\sim 0.1-1$, and the Hall parameter, $\beta_{i,H_2}\gg1$, indicate partial ion-neutral decoupling, with non-ideal effects becoming more significant toward the inner regions. In the envelope, our analysis shows that rotational energy decreases from $\sim 50\,\%$ of the gravitational energy at $R_{cen}$ to $\sim 25\,\%$ at $2R_{cen}$ and $\sim4\,\%$ at 1100 au. In contrast, for $B_0=1$ mG, magnetic energy increases from $\sim 0.4\,\%$ of the gravitational energy at $R_{cen}$ to $\sim2.6\,\%$ at $2R_{cen}$ and $\sim20\,\%$ at 1100 au. The drift velocity expected from radial force balance is consistent with the modeled drift velocity within $\sim300-400$ au of the central source but is lower at larger radii. These results are consistent with a scenario in which ambipolar diffusion leads to partial ion-neutral decoupling, allowing neutrals to move faster than ions within the envelope.
Read the full abstract

All authors J. A. López-Vázquez, Chin-Fei Lee, Carlos Tapia, Jyun-Heng Lin, Claudio Codella, Linda Podio, Jesús Miguel Jáquez-Domínguez

Statistical Methodology

SAMBoost: Spatially Aware Mesh Boosting

We propose SAMBoost (Spatially Aware Mesh Boosting), a gradient boosting method that combines tree-based covariate modelling with a stochastic partial differential equation (SPDE) latent spatial field through a single penalised objective. At each boosting iteration, the predictor is decomposed into separate covariate and spatial components. The method is motivated by applications in which the spatial effect is not a nuisance to be removed but a quantity of interest that must be recovered alongside the covariates. In insurance, for example, practitioners need both flexible nonlinear covariate effects and an interpretable geographic risk surface that can be reported on its own. Existing approaches often involve trade-offs, offering flexible covariate modelling but no explicit spatial field, rigorous spatial inference with a fixed additive covariate structure, or a combination of both at computational costs that limit practicality for large datasets. SAMBoost addresses these goals within the boosting loop, and the underlying mesh produces a sparse precision structure whose cost is largely independent of sample size. On simulations and on an open benchmark, SAMBoost matches or improves on the predictive accuracy of joint Gaussian process and tree boosting approaches. It is also faster and recovers the spatial field and covariate effects more accurately. When covariate effects include interactions, it outperforms additive Bayesian spatial models. Applied to a flood insurance portfolio with 4.9 million records, SAMBoost trains on the full training set without subsampling and runs about sixteen times faster than a Bayesian reference, with a consistent spatial risk surface. A public Python implementation is available.
Read the full abstract

All authors Mulah Moriah, Franck Vermet, Pierre Ailliot, Juliette Legrand

Econometrics

Panel Local Projections with Missing Data

A panel local projection can remain identified even when missing data leave no complete regression case. We recover the projection over specified dates by combining factor-based conditional means with idiosyncratic covariances from observed pairs. Under strong factors in the auxiliary panel, sufficient factor variation at observed dates and enough observations of the required pairs, a second-order expansion yields bias-corrected moments and a joint score for inference. The localized multiplier bootstrap provides pointwise confidence intervals and simultaneous bands under additional dependence restrictions. We revisit the investment response to geopolitical risk studied by Caldara and Iacoviello (2022) with a larger quarterly firm sample and extend the analysis to monthly projections of trailing-three-month investment intensity. In the monthly analysis, the six-month response difference between high- and low-exposure firms becomes more negative as firms with incomplete histories enter the sample.
Read the full abstract

Numerical Analysis

How good are spline spaces for the Kolmogorov $n$-width of a Sobolev ball?

This paper studies the approximation power of polynomial splines on a uniform one-dimensional mesh. The quality of the best approximation of the unit ball of $H^{s}$ by an $n$-dimensional subspace, measured in the $L^{2}$ norm, is quantified by the sup-inf of that space: for each function in the Sobolev ball, take its minimal approximation error over the space, and then take the supremum of these errors over the ball. The infimum of this sup-inf over all $n$-dimensional spaces is the well-known Kolmogorov $n$-width $d_{n}$. Choosing the $n$-dimensional subspace to be the piecewise polynomial space of degree $k$ with smoothness $C^{\mu}$ ($-1\le\mu\le k-1$) gives our sup-inf $E$. The approximation accuracy of spline spaces is then examined by comparing $d_{n}$ and $E$. The two quantities are determined by two eigenvalue problems, through which we obtain a lower bound for $d_{n}$ and an upper bound for $E$. Under the restriction $k\ge s-1$, the upper bound for the error of the maximally smooth spline space ($\mu=k-1$) is asymptotically equal to the lower bound for $d_n$, showing that uniform $B$-splines are asymptotically optimal among all $n$-dimensional spaces. For $\mu<k-1$, we prove that $E/d_n$ is bounded below by a constant strictly larger than $1$, except when $k>s-1$ and $\mu=k-2$ hold simultaneously. This reveals the superiority of $B$-splines in the sense of $n$-widths. Numerical experiments corroborate these bounds and, in some cases (e.g. the exception above), show that they are sharp.
Read the full abstract

Computer Vision

ClearForm: Shared-Surface Gaussian Splatting for Transparent and Opaque Scenes

Transparent-interface reconstruction must recover glass surfaces even when most observed color comes from the scene behind them. We present ClearForm, a shared-planar-Gaussian method for jointly reconstructing transparent interfaces and the appearance seen through them. We identify and analyze the Geometry-Appearance Compositing Dilemma (GACD): when the same transparency-dependent alpha controls both color and geometry, revealing the background can weaken a transparent surface's contribution to geometric reconstruction. We derive its effect on plane-based depth estimation, identify a mechanism that can weaken geometric learning, and quantify the loss of interface support through controlled comparisons. To address this conflict, we develop geometry-preserving dual-opacity surface reconstruction. Shared planar Gaussians retain interface contributions in geometry compositing while modeling transmission and sparse reflection in RGB, supported by transparency-aware geometric constraints. To manage the different primitive-visibility requirements of these compositors, we introduce Dual-Evidence Lifecycle (DEL), which retains Gaussians supported by multi-view geometry or RGB reconstruction evidence. On 3D-FRONT-T, ClearForm achieves 27.05 mm Chamfer distance (CD) and 0.928 F1, compared with 59.18 mm CD for our PGSR reproduction and the reported 34 mm CD and 0.836 F1 for GLINT. On TransLab, it achieves 1.65 mm CD, 0.963 F1, and 40.14 dB PSNR. In matched ablations, dual-opacity compositing reduces CD from 59.25 to 27.09 mm, while DEL improves PSNR by 2.06 dB. Our project page is available at https://hajeongx.github.io/clearform/.
Read the full abstract

Computation and Language

FEROM: Frontier Endogenous Reveal-Order Marginal Policy Optimization for Masked Diffusion LMs

Masked diffusion language models generate text by iteratively unmasking positions, with practical samplers often selecting reveal positions from the model's own logits. The reveal order is therefore an endogenous latent variable of the rollout process with an enormous discrete latent space. Policy optimization objectives relying on pre-defined schedules or sampled trajectories introduce either bias or high variance. In this paper, we propose FEROM, Frontier Endogenous Reveal-Order Marginal Policy Optimization, which approximates policy gradients for the rollout-induced marginal response policy of masked diffusion LMs. FEROM derives a Rao-Blackwellized policy-gradient identity over latent reveal orders and expresses the resulting estimator as posterior edge occupancy on a reveal-state DAG. To make marginalization practical, we introduce Frontier Reveal Marginalization, a budgeted estimator that approximates this marginal edge structure by combining CDF-based computation of local probabilities of the endogenous reveal process with frontier expansion of high-mass partial states. Integrated into a GRPO-style objective, FEROM replaces single-path log-scores with a detached token-score surrogate over posterior-relevant reveal edges. Experiments on math and coding tasks show comparable or improved results over existing methods under matched scoring-state budgets. An offline proxy study further shows improved marginal-gradient approximation with increased scoring budgets.
Read the full abstract

All authors Zian Su, Ziyang Huang, Kaiyuan Zhang, Xiangyu Zhang

Strongly Correlated Electrons

Discovery of Kondo-Kramers Nodal Line Topological State in strongly correlated material CeAuAl$_3$

Kramers nodal lines (KNLs) are predicted in all achiral, non-centrosymmetric crystal structures, and yet there are very few experimental realizations of KNLs, only in weakly correlated systems. Especially, their realization in strongly correlated systems remains unexplored. Here we investigate KNL topology in the non-centrosymmetric, achiral Kondo lattice CeAuAl$_3$, using magnetotransport, thermodynamic measurements, density functional theory (DFT), and tight-binding toy model calculations. We find that CeAuAl$_3$ hosts KNLs in the $\mathit{\Gamma}\text{-}Z$ and $X\text{-}P$ directions, protected by time-reversal and mirror symmetries. Breaking of these crystalline symmetries, along with time-reversal symmetry via spin polarization, lifts the KNL degeneracy, which, in turn, generates Berry curvature resulting in an intrinsic anomalous Hall effect (AHE). More importantly, CeAuAl$_3$ shows large AHE not only below the magnetic ordering temperature $T_N$ but rather below the Kondo temperature. This underscores the role of the Kondo correlations in bringing the KNLs at the Fermi surface rather than at accidental energies, as would be expected without correlations. Our work identifies a new topological state, Kondo-Kramers Nodal Line, with CeAuAl$_3$ being the first compound explicitly identified to host this new correlated topological state, and demonstrating the expected AHE associated with it. More generally, our work outlines the pathway for finding Kondo-KNL systems in all compounds that combine achiral, non-centrosymmetric structure with Kondo correlations.
Read the full abstract

All authors Hari Bhandari, Kevin Allen, Kuan-Sen Lin, Yuan Fang, Arushi, Karthik Rao, Mounica Mahankali, Sanu Mishra, Jian-Xin Zhu, Christopher Lane, Qimiao Si, Emilia Morosan

Physics and Society

Telling coordination from coincidence: a no-interaction twin for collective systems

Measures of emergence based on information theory or causal closure ask whether a group variable has a predictable law of its own, but not whether that law was produced by the members' interactions: the centre of mass of non-interacting random walkers passes the standard criterion $\Psi$. We introduce the interaction share S, the fraction of the forecast error of a system's no-interaction twin (the same members with every interaction removed) that the real group law eliminates, and a projected estimator with expectation zero without interaction. For Gaussian forecasts, $-\frac{1}{2}\ln(1-S)$ is the expected Kullback-Leibler divergence of the real group law from the twin's. S is exactly zero for the centre of mass of linearly coupled walkers, and a machine-checked proof extends this to any pairwise force obeying Newton's third law. Across magnets, oscillators and flocking models it rises from 0 to about 1 where collective order appears. With time-shifted virtual groups and positive controls, it separates real fish schools (0.54-0.91, 17 recordings, two species) from groups that never interacted (-0.10 to 0.07), and it falls to zero when zebrafish are placed in the dark. In 221 mouse brain areas and in US stock markets the shares are small (about 1-3%) and were confirmed in pre-registered tests on unseen data. For linear forecasts the share is a memory gap, computable from two autocorrelation curves without fitting a twin; it matches measured shares (Spearman 0.97) and predicted fresh markets within 0.004. All code, data pipelines and hashed pre-registrations are public.
Read the full abstract

Robotics

CPNav: Calibrated Vision-Language Navigation for Object-Goal Search in Unseen Environments

Directing a robot to locate a target object in unseen environments remains a central challenge in embodied AI. Vision-Language Model (VLM)-based methods provide rich semantic reasoning but often produce uncalibrated and overconfident action decisions. We introduce CPNav, a navigation framework that combines collision-aware planning with statistically calibrated VLM action scoring for map-free object-goal navigation. From RGB-D input, CPNav constructs a navigability map, generates collision-free motion primitives, and queries two VLMs for goal detection and action scoring. The resulting scores are calibrated offline via conformal prediction (CP) using oracle actions obtained through a tree-structured exploration procedure with backtracking, providing finite-sample guarantees at a user-specified risk level. The CP layer is model-agnostic and can wrap any VLM policy. Online, calibrated VLM scores guide action selection, while a tree-structured exploration history records alternative actions and allows backtracking. We evaluate CPNav in both simulation and real-world hardware settings. In AI Habitat simulation, compared with the uncalibrated baseline method, CPNav achieves relative SR and SPL improvements of 6.2% and 11.4% on HM3D v0.2, and 12.4% and 8.8% on MP3D, respectively. In real-world deployment on a modified Hiwonder MentorPi robot, the corresponding improvements in SR and SPL are 24.9% and 15.2%, respectively. Code is available at https://github.com/practice-lab-ucla/CPnav.
Read the full abstract

All authors Qizhao Chen, Yuanhong Zeng, Shoh Nishino, Anushri Dixit

Fluid Dynamics

Tuning fluid frictional drag through flow-induced aggregation of wormlike micelles

Charged amphiphilic surfactants are versatile polymorphic molecules which can form supramolecular structures, such as wormlike micelles. We study the rheology and non-Newtonian fluid dynamics of the cationic surfactant trimethyltetradecylammonium chloride (TTAC) combined with the aromatic counter-ion sodium salicylate (NaSal). In the non-dilute regime, this solution exhibits a linear viscoelastic response consistent with a single-mode Jeffreys model, and in capillary pinch-off experiments, fluid filaments form extended cylindrical threads, indicating strong strain-stiffening. In the dilute regime the linear viscoelasticity and extensional rheology become immeasurably small, but the TTAC/NaSal mixture develops flow-induced structures at high shear rates, as the result of an end-to-end concatenation of micelles produced by shear-induced flow alignment. These flow-induced structures augment both the measured viscosity and the viscous drag in laminar flows at high shear rates (by as much as 50 %). In turbulent Taylor-Couette flow, the flow-induced structures produce a sudden and dramatic destabilizing influence on the stable toroidal Taylor vortices, resulting in upwards of a 200 % enhancement in turbulent drag. However, with increasing Reynolds number, the turbulent flow transitions smoothly to a state of drag reduction, generating up to a 30 % reduction in fluid frictional drag but without the shear-induced degradation that plagues drag reduction strategies based on solutions of high molecular weight polymer, such as polyacrylamide. These results show that surfactants are a unique and effective modifier of fluid frictional drag in both laminar and turbulent flow regimes, capable of producing both significant levels of drag enhancement and drag reduction by tuning the composition and operating conditions.
Read the full abstract

All authors Lucas Warwaruk, Amanda Tiano, Berith Isaac, Gareth H. McKinley

Machine Learning

Bias, Measurement Error, and Double-Dipping: When Can GNN Convolutions Help Brain Connectome Prediction?

Graph Neural Networks (GNNs) are an important tool for fMRI prediction, where empirical functional connectivity matrices serve both as features and for graph construction. However, recent studies and our own experiments show that these methods often fail to outperform simpler graph-agnostic baselines. We propose a statistical explanation for this failure by formulating connectome prediction as an errors-in-variables (EIV) problem: the true connectome is latent, while the observed covariance matrix is a noisy finite-time-series estimate. In a linear asymptotic setting, we decompose the effect of graph convolution into three mechanisms: smoothing-induced bias and variance reduction, graph estimation error, and a double-dipping effect arising when the same noisy matrix determines both graph and features. Under a latent community model adapted to neuroscience, we identify a narrow favorable regime, set by the brain community structure, the predictive signal, the graph construction and the recording length, in which message passing can improve prediction. Simulations validate this decomposition, while real-data experiments illustrate that connectome GNN pipelines often fall outside this regime. Our results provide a first step towards explaining when graph convolutions can help brain connectome prediction, and why they often do not.
Read the full abstract

All authors Tommaso Castellani, Jiaqi Li, Muriah D. Wheelock, Rezwana Rosen Razzaque, Helmut Laufs, Hong Chen, Claire Donnat