Computer Vision
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Saturday 12 September
Counterfactual world models (CWM) extract motion from pretrained video predictors by comparing factual and intervened predictions, but uniform aggregation weights responses equally without explicitly incorporating physical priors. Our key insight is to incorporate physical priors into candidate reliability learning, motivating LPA-CWM with a lightweight Learned Physical Adjudicator (LPA). Trained on dense MOVi-F trajectories, the 3.0M-parameter LPA compares visual context and response structure across an unordered candidate set to predict relative weights; windowed localization and one paired re-evaluation recover motion with the CWM frozen. Existing video-level benchmarks do not directly assess motion correspondence, where low localization error can conceal missing trajectory segments. We introduce Completeness-aware Motion Correspondence (CMC), a ground-truth-anchored protocol jointly measuring localization, completeness, visibility, and continuity, counting missing predictions as failures on visible dynamic points. Across DAVIS, Kinetics, and RoboTAP, LPA-CWM improves all main CMC measures over Uniform CWM, with relative gains of 18.1%–60.0% in average Dynamic Correspondence Accuracy ($\mathrm{DCA}_{\mathrm{avg}}$), and improves TAP-Vid First tracking accuracy (overview: https://LPA-CWM.github.io).
Machine Learning
Multi-Modal Tumor Survival Prediction via Graph-Guided Mixture of Experts
Large Language Models (LLMs) have displayed impressive capabilities in handling tasks that require few demonstration examples, making them effective few-shot learners. Despite their potential, LLMs face challenges when it comes to addressing complex real-world tasks that involve multiple modalities or reasoning steps. For example, predicting cancer patients' survival period based on clinical data, cell slides, and genomics poses significant logistical complexities. Although several approaches have been proposed to tackle these challenges, they often fall short in achieving promising performance due to their inability to consider all modalities simultaneously or account for missing modalities, variations in modalities, and the integration of multi-modal data, ultimately compromising their effectiveness. This thesis proposes a novel approach for multi-modal tumor survival prediction to address these limitations. Taking inspiration from recent advancements in LLMs, particularly Mixture of Experts (MoE)-based models, a graph-guided MoE framework is introduced. This framework utilizes a graph structure to manage the predictions effectively and combines multiple models to enhance predictive power. Rather than training a single foundation model for end-to-end survival prediction, the approach leverages a MOE-guided ensemble to manage model callings as tools automatically. By leveraging the strengths of existing models and guiding them through a MOE framework, the aim is to achieve better performance and more accurate predictions in complex real-world tasks. Experiments and analysis on the TCGA-LUAD dataset show improved performance over the individual modal and vanilla ensemble models.
Read the full abstract
Astrophysics of Galaxies
The New Dark Matter Density Profile from JWST JADES Galaxies
We present the first dark matter density profile derived directly from James Webb Space Telescope JADES Data Release~3 NIRSpec observations of $N = 587$ galaxies spanning the cosmic noon epoch $1.5 \leq z \leq 3.5$. From each NIRSpec G235M/G395M spectrum we extract the H$\alpha$ emission-line velocity dispersion $\sigma_{\rm ha}$, stack galaxies in four redshift bins of $\Delta z = 0.5$, and reconstruct representative group rotation curves within the General Theory of Relativity. Fitting the four-parameter modified exponential model and deriving the exact GTR energy density, we reduce the profile to a compact $[2/3]$ Padé approximant with all coefficients determined in closed form from JWST observations. The profile is cusp-free, fully analytic, and redshift-dependent. Energy conditions, causality, and orbital stability are all satisfied. Most strikingly, the central density $\rho(0)$ varies by less than $15\%$ across $z = 1.5$–$3.5$ despite a $35\%$ decline in the asymptotic rotation velocity, revealing a universal dark matter core saturation density at cosmic noon decoupled from baryonic evolution.
Read the full abstract
Quantum Physics
Formulating Multistage Cutting Stock Problems as QUBO
Cutting stock problems are of large relevance to a variety of industry branches. Here, we present a linear programming formulation for a restricted version of the multistage 2D cutting stock problem with Guillotine cuts and test it on published benchmark instances. After this, we derive a non-exact reformulation in quadratic unconstrained binary optimization (QUBO) form via unbalanced penalization, which we show to have possible benefits in modeling the problem with usable leftovers. This work mainly considers the restricted formulation, in which the dimensions of each cut are determined by those of a single required piece. We further discuss how the approach can be extended to the unrestricted problem, which allows more general cutting patterns involving cuts with dimensions of multiple required pieces and naturally gives rise to quadratic terms in the inequalities of the problem formulation. We showcase solutions via a Simulated Annealing solver for the restricted problem class, where we mainly study the performance of the iterative augmented Lagrangian method. We also discuss possible applications of Quantum Annealing in this setting, as a strong motivation to research QUBO formulations.
Read the full abstract
Econometrics
What Variation Identifies Payoffs in a Dynamic Game?
Observed choice in a dynamic game mixes current profit with continuation value. A rival adds a second problem: the same comparison averages over the rival's equilibrium policy. Changing the primitive transition rewrites continuation technology; changing the rival's Markov policy, holding that law fixed, rewrites the mixture over rival-contingent payoffs. The two are not substitutes. For a rival-feature payoff of rank $K$, rank identification up to location requires $\Ephi=\lceil(MK-1)/(M-1)\rceil$ policy environments, and a second kernel when payoffs are saturated. Rank can still be restored by arbitrarily small policy differences. Independent private shocks force mixed rival actions to factor, so a payoff that depends jointly on $d$ rivals is visible only at order $\eta^{d}$ near a common interior baseline. Either rank fails or the smallest identified singular value is at most $\kappa\eta^{\dPhi}$, independently of how many kernels are stacked. Oracle-GLS variance in that direction vanishes only if $n\eta^{2\dPhi}$ diverges. An Anderson–Rubin set that carries first-stage error in the design matrix covers without a vanishing-risk condition. On U.S.\ airline entry, even among rank-identified directions, the most favorable rival-dependent contrast is several times wider than observed behavior.
Read the full abstract
Probability
Functional limiting behaviour of non-stationary marked Hawkes processes under multi-scaling high-intensity regime
This paper studies the non-stationary marked Hawkes processes in which the base intensity function is time-dependent and the kernel function is governed by an external random factor called mark. The marks are assumed to be determined by random events occurrence over time resulting in a non-identical distribution. These facts introduce two kinds of difficulties: one from the absence of the i.i.d. (independent and identical distributed) and the other is from the non-stationarity. In this framework, an asymptotic regime, often referred to as the high intensity regime, is considered, under which the intensity increases with time. This can be obtained by multiplying the time parameter by n alpha, for some alpha > 0. Under the high intensity regime, the functional law of large numbers and the functional central limit theorem are established. The rescaled (centered and normalized) marked Hawkes process is proved to converge in distribution to a Gaussian process, which is the accumulation of Gaussian noise and a diffusion process. In addition, a shot-noise process is studied, and the functional Limit Theorems under appropriate hypotheses are established.
Read the full abstract
Systems and Control
Fixed-Time Resilient Integral Reinforcement Learning for Input-Constrained Unknown Nonlinear Systems Under FDI Attacks and Disturbances: A Data-Driven Admissible Warm Start
This paper develops a resilient learning controller for unknown nonlinear systems operating under actuator limits, false-data-injection attacks, and external disturbances. The key idea is to learn a saturated secure policy directly from finite trajectory data while guaranteeing that both the learning error and the closed-loop state converge to compact neighborhoods within a uniform fixed time independent of initial conditions. An integral formulation removes the unknown drift from the implementable learning law, while stored informative data sustain learning after online excitation fades. To mitigate the closed-loop sensitivity to arbitrary critic initialization, pre-deployment data, which may also be reused from the replay stack, are lifted through a finite-dimensional Koopman representation to construct a stabilizing initial policy, whose inverse saturated-policy map provides a data-driven critic-weight warm start. The resulting controller preserves input constraints by construction and guarantees practical fixed-time robustness under persistent attacks and disturbances. The proposed learning and initialization architecture is further verified through a two-link robot stabilization example, where the results demonstrate rapid state recovery, bounded critic learning, reliable actuator-constraint satisfaction, and improved closed-loop behavior under informed critic initialization.
Read the full abstract
Computation and Language
GraMRAG: Orchestrating Multi-Agent Multi-Step Reasoning via Graph Memory with Reinforcement Learning
Although existing multi-agent Retrieval-Augmented Generation (RAG) systems have demonstrated promise on complex multimodal reasoning tasks, they remain fundamentally limited in reasoning depth and memory structure, suffering from inadequate retrieval and state blindness when answering knowledge-intensive questions. To address these limitations, we propose GraMRAG, a graph memory-guided multi-agent RAG framework that integrates a dynamic multimodal memory graph to enable stable, multi-step multimodal reasoning. We introduce a vision-text bridged reasoning paradigm that unifies multi-scale entity cropping with a ReAct-style visual toolchain, enhancing the long-horizon cross-modal reasoning capability. We further construct a multimodal memory graph that formalizes agent reasoning as a dynamic directed acyclic graph (DAG), explicitly modeling action-observation dependencies to mitigate state blindness and suppress redundant retrieval. Moreover, we propose Topology-Aware Policy Optimization (TAPO) that leverages graph topology for critical path identification and targeted node pruning, enabling fine-grained credit assignment across multi-step reasoning trajectories. Extensive experiments on challenging multimodal benchmarks demonstrate that our approach consistently outperforms existing baselines and achieves state-of-the-art performance on complex long-horizon reasoning tasks.
Read the full abstract
Machine Learning
Stabilizing Performative Feedback Loops with Minimal Model Deployments
When algorithmic predictions inform people's decisions, the models we deploy are performative and actively shape the data we see. This feedback loop between algorithms and their broader environments introduces a challenge in the mechanics of social prediction: If different predictive models induce different distributions, is it possible to efficiently learn a prediction rule that is optimal for the distribution that it induces? Formally, this solution concept is known as performative stability. A core challenge in learning a performatively stable predictor is that, unlike supervised learning where distributions are fixed, the learner must deploy different predictors and observe their induced distributions. The main contribution of our work is a new algorithmic procedure that, in the high-accuracy regime, finds a performatively stable model in nearly the minimum number of model deployments without making any assumptions regarding how predictions shape distributions. In particular, our procedure succeeds at finding a randomized performatively stable predictor using exponentially fewer model deployments than prior approaches. Our second main contribution is a structural result showing how this recent randomized notion of stability achieved by our algorithm can be derandomized into a single predictor satisfying the prior deterministic notion if one is willing to assume that the loss is well-conditioned and that performative effects are weak, as in early work in this area. On a technical level, our results come from building on an underexplored technical connection between performative stability and expected variational inequalities.
Read the full abstract
Rings and Algebras
Matrices over non-commutative rings as sums of fifth and seventh powers
Let $R$ be non-commutative ring with unity. In this paper, we prove that a given matrix in $M_n(R)$ is a sum of $p$-th powers in $M_n(R)$, where $n < p$ if and only if its trace can be written as a sum of $p$-th powers in $M_n(R)$, for $n = 2,3,4$ and $p = 5$ and $n = 4,5,6$ for $p = 7$. This paper extends the results of S. A. Katre and Kshipra Wadikar, $(p=3)$ and Garge $(n \geq p), (p = 5, 7)$, and S. A. Katre and Deepa Krishnamurthi $(n \geq p)$.
Read the full abstract
Geometric Topology
Brown's Asymptotic Limit Functor and Proper Homology
In 1974, E. M. Brown introduced the $\wp$-functor to study the proper homotopy groups of an end, suggesting a parallel proper homology theory that has since remained undeveloped. In this paper, we construct this missing proper homology and relate it to proper homotopy by developing a proper Hurewicz theorem. Bypassing abstract pro-categorical machinery, we show that $\wp$ has major advantages over the classical limits $\varprojlim$ and $\varprojlim^1$: it is exact, detects pro-triviality, satisfies a cardinality dichotomy, and connects to these classical limits through a 4-term exact sequence. As an application, we prove that the Brown–Grossman proper fundamental group of any open contractible $3$-manifold other than $\mathbb{R}^3$ is uncountable and perfect.
Read the full abstract
Human-Computer Interaction
When Single-User-Oriented LLM-based Assistants Involve Others: A Scoping Review of Pathways, Risks, and Responses
LLM-based assistants are increasingly extending into multi-party contexts, while core operational processes for context management, personalization, identity attribution, authority attribution, and action execution often remain organized around a single user. Existing work examines particular multi-party settings, but lacks a systematic account of how these single-user-oriented assistants begin to involve additional human parties and what risks emerge. To address this gap, we conducted a scoping review of 58 studies. We identify five operational pathways spanning direct and indirect involvement, five recurring risk domains, and five areas of implemented and proposed responses. Based on these findings, we argue for governance that attends to changing cross-person roles and relationships in practice, and for assistant designs that preserve person-specific boundaries throughout interaction.
Read the full abstract
Computational Physics
Exceptional-point conjugate symmetry and migration in fluid-loaded elastic waveguides: restructuring the physical dispersion spectrum
The dispersion spectrum of a fluid-loaded elastic waveguide is a non-Hermitian system with eigenmodes on a two-sheeted Riemann manifold. Although exceptional points (EPs) organize avoided crossings, the topological rules by which fluid loading restructures the observable spectrum (both physically admissible modes and their connectivity) remain unknown, and conventional solvers fail to recover complete branches. Here we establish these rules. We prove that, for real elastic moduli and real fluid parameters, EPs obey a conjugate-pair symmetry, and that a pair on the physical sheet enforces an avoided crossing of real wavenumbers between the two modes it controls, yielding a topological criterion for mode identification. We then identify two independent reconstruction mechanisms. First, the physical observation space expands continuously: a mode observable above its vacuum cut-off can extend downward to the sound line, admitting solutions in a frequency band empty in vacuum; vacuum-anchored seeds above cut-offs continued downward capture such branches systematically. Second, EP migration rewires mode connectivity: as fluid density increases, conjugate pairs leaving the physical sheet (signaled by a sign reversal of the imaginary part of either the in-plane or the vertical wavenumber) leave the two previously coupled physical branch segments unconnected by any EP within the observation space; they become independent curves that may intersect freely on the real frequency axis, with the narrowest veerings losing their topological protection first. Mirror-symmetry breaking births additional EPs generically in the trapped regime but only conditionally in the leaky regime. Numerical computations on symmetric and asymmetric composite laminates under single- and double-sided water loading validate the framework, recovering multiple leaky branches missed by standard solvers.
Read the full abstract
Cryptography and Security
AGENTQ: Quantization-Conditioned Backdoor Attacks on LLM Agents
Quantization is one of the default deployment paths for open-weight LLM agents, but it is not behavior-preserving: an adversary can release a full-precision checkpoint that passes audits yet misbehaves once quantized, termed as quantization-conditioned attack (QCA). Prior QCA work targets free-text generation, where harm is mediated by a human reader. In contrast, the agentic setting poses a more severe risk: the triggered payload is a structured function that can be executed without human oversight. We present the first study of QCA against LLM agents. We find that directly adapting prior backdoor-injection methods can produce malicious behavior after quantization, but substantially degrades benign utility, rendering the resulting attacks impractical. To understand the true upper bound of the threat, we propose AGENTQ, an attack framework that combines layer-banded LoRA injection with partial-PGD repair over a multi-codebook quantization-equivalence class. AGENTQ preserves normal agentic capability while concentrating malicious behavior in the quantized model. Across three trigger-action pairs and three codebooks (NF4, FP4, INT8), AGENTQ reaches up to 100% post-quantization attack success rate with minimal loss of benign utility, underscoring the need to make quantization-aware safety evaluation a standard requirement before open-weight agents are deployed.
Read the full abstract
Theoretical Economics
Strategy-Proof Set-Valued Choice on Single-Dipped Domains
We study strategy-proof social choice where agents have single-dipped preferences and outcomes may be sets of alternatives. While point-valued strategy-proof rules select an endpoint, we ask which outcomes survive when sets are compared through responsive extensions. Under the standard responsive extension, preferences over fixed-length intervals remain single-dipped, so the endpoint restriction persists and anonymity yields a single-threshold rule. Under responsive dominance, the incomplete order shared by all responsive extensions, an interior interval can survive because the two extreme intervals need not be comparable. For arbitrary sets, requiring every chosen alternative to be most preferred by some agent yields voting-by-committees rules, with one committee per endpoint; anonymous rules have a two-threshold quota form. The characterization holds on the entire endpoint-peaked domain, and every such rule is group-strategy-proof.
Read the full abstract
Robotics
Mobile Multi-Robot Navigation under Runtime Uncertainty via Koopman Operator Learning and Nonlinear Model Predictive Control
In this work, we developed a nonlinear model predictive control (NMPC) framework that employs learned dynamics via the Koopman Operator theory for mobile multi-robot navigation. We formulated and solved NMPC problems using a lifted bilinear Koopman-based model that accurately predicts affine input systems affected by perturbations and uncertainties. Two exemplary multi-robot navigation problems are considered: target reaching and formation control. The output of our method enables closed-loop multi-robot navigation and formation control in environments populated with obstacles, whereby the Koopman operator-based model used in the NMPC formulation addresses runtime uncertainties, namely, various degrees of random wheel slipping. We validated the effectiveness of our method for both problems via extensive numerical simulations in different environments with wheeled robots affected by different amounts of slip and without knowledge of their true dynamic models.
Read the full abstract
Computation and Language
Measuring the Creativity of Frontier LLMs in Automated Research
Frontier LLMs are increasingly capable of conducting automated research, yet their creativity in this setting has not been systematically evaluated. We propose a set of metrics to evaluate creativity along the two dimensions of valueness and novelty. Valueness assesses whether each proposed idea is useful, while novelty is evaluated from three perspectives: whether the same idea has appeared before (Exact-Match P-Novelty), whether the modified variable or variable combination has been explored before (Variable-level P-Novelty), which reflects the breadth of research-space exploration, and whether the proposed idea is explicitly attributed to external knowledge in the model's reasoning (H-Novelty). Our evaluation shows that the models achieve relatively similar Valueness and Exact-Match P-Novelty scores, while differing substantially in Variable-level P-Novelty. H-Novelty is also consistently high among the models for which it can be evaluated. Notably, further correlation and idea-level performance analyses reveal a strong positive correlation between Variable-level P-Novelty and research performance.
Read the full abstract
Soft Condensed Matter
Spatially Resolved Reconstruction of Ising Couplings in Tunable Colloidal Artificial Spin Lattices
Buckled colloidal monolayers constitute a versatile soft-matter platform for engineering artificial spin lattices, with each particle serving as a single Ising spin. While the average Ising coupling energy has been approximately derived for perfect particle lattices, extracting the complete set of coupling parameters from real samples remains inaccessible. Here, we apply an inverse method to reconstruct all nearest-neighbor effective Ising coupling energies from measured colloidal spin configurations. We design multiple experimental protocols to control the thermodynamic state of the colloidal system, from isotropic compression and shear deformation to modulation of interparticle attraction, each giving rise to distinct spin configurations. Using spin configuration data, we reconstruct all nearest-neighbor effective Ising coupling energies using maximum likelihood estimation. To assess inference reliability without ground-truth model parameters, we propose to use the convergence of the standard deviation of the estimated couplings as a practical, ground-truth-free criterion, and validate its reliability using simulation data with known parameters. The extracted spatially resolved couplings reveal how each control protocol influences the sign, magnitude, statistical distribution, and spatial arrangement of the microscopic coupling parameters: isotropic compression strengthens antiferromagnetic couplings and enhances quenched disorder; shear deformation generates direction-dependent anisotropic couplings; and increased interparticle attraction drives a crossover from antiferromagnetic to paramagnetic and then to ferromagnetic couplings. This work establishes a practical inference framework for estimating effective model parameters of colloidal artificial spin lattices.
Read the full abstract
Fluid Dynamics
Retraction dynamics of polymer-laden droplets impacting a smooth solid surface
Hypothesis: The presence of a polymer in a droplet plays a significant role in contact line dynamics following droplet impact onto a solid substrate. We hypothesize that the polymer concentration modifies the post-impact wetting behavior. Experiments: In our experiments, we visualize the contact line dynamics of polymer-laden droplets impacting onto a flat sapphire substrate. Using high-speed side and bottom-view imaging, we analyze droplet spreading and retraction over a broad range of Weber numbers (10-700) and polyethylene oxide (PEO) concentrations (0-400 ppm). To systematically characterize the contact line dynamics, we used image processing to quantify the maximum spreading factor, receding velocity, and receding contact angle. Findings: The effect of the polymer on the spreading phase is negligible, as the maximum spreading factor follows classical inertial-capillary scaling, showing that polymer addition produces no measurable change in the maximum spreading under the present conditions. In contrast, the receding phase exhibits a strong dependence on the polymer concentration: increasing the PEO content leads to a significant reduction in the receding velocity and dynamic contact angle, while the spreading dynamics remain water-like. These results are consistent with an additional resistance associated with the receding contact-line region, while the microscopic origin of this resistance remains unresolved. To capture this behavior, we propose a scaling model that relates the normalized retraction velocity to the receding contact angle and impact conditions. The model effectively collapses the experimental data across all investigated concentrations and Weber numbers.
Read the full abstract
Combinatorics
Simplicial Complexes of Antichains in Root Posets and Related Combinatorics of Dyck Paths
For a crystallographic root system ${\mathfrak D}$ we consider the simplicial complex $\Delta_{\mathfrak D}$ of all antichains in the root poset of ${\mathfrak D}$. We show that $\Delta_{\mathfrak D}$ is shellable if and only if ${\mathfrak D}$ is $A_n$, $B_n$, $D_3$ or $G_2$. Since antichains in types $A_n$ and $B_n$ can be identified with Dyck paths and symmetric Dyck paths, respectively, this yields a simplicial complex on Dyck paths. Indeed, in type $A_n$, shellability can be extended to rational Dyck paths. The $f$- and $h$-triangles then yield statistics on (symmetric/rational) Dyck paths. We determine these statistics for $A_n$ and $B_n$ and leave the case of rational Dyck paths as an open problem.
Read the full abstract
Systems and Control
Characterizing Identifiability and Generalization for Inverse Receding-Horizon Linear-Quadratic Regulator Problems
We consider the problem of objective inference in the context of receding-horizon linear-quadratic regulator (LQR). In this setting, we are given sequential state-action observations, where each observed action is the first control of a newly solved finite-horizon LQR problem. We characterize when the objective of that problem is uniquely identifiable from these observations and when additional observations provide no new information about the objective. We then analyze action prediction at unseen states and show that all objectives reproducing the observed actions yield identical actions throughout the affine hull of the observed states; outside this hull, we derive an upper bound on the prediction error. % and deriving a prediction-error bound outside this hull. Additionally, we show that, when only the linear objective terms are unknown, exact prediction holds at every state. Finally, numerical results show that, even under stochastic observation noise, re-optimizing an inferred objective enables accurate action prediction at unseen states across different planning horizons.
Read the full abstract
Cryptography and Security
Symmetric Models for Syndrome Decoding
This paper introduces a new polynomial model for the exact variant of the Syndrome Decoding Problem (SDP) in the binary case. The model is based on elementary symmetric polynomials. We estimate the computational complexity of solving the corresponding polynomial system by establishing bounds on the degree of regularity and on the solving degree of the ideal associated to the model. The complexity estimate is lower than for previous polynomial models. We also provide a variant of the model whose complexity depends directly on the specific instance of the SDP and is lower than for the first model. Finally, we discuss how to apply our approach to solve other variants of the SDP.
Read the full abstract
Materials Science
Large spin splitting metallic altermagnets from machine-learned design rules
Altermagnets combine compensated magnetic order with momentum-dependent spin splitting, providing spin-polarized electronic states without a net magnetization. Metallic $d$-wave altermagnets are particularly promising because they can support time-reversal-odd spin currents in linear response, yet experimentally validated bulk realizations remain scarce. Here, we use high-throughput density-functional-theory data to relate the magnitude of altermagnetic band splitting to compositional, structural, and DFT-derived magnetic descriptors. An interpretable gradient-boosted model identifies two candidate-prioritization criteria: compact unit cells and magnetic sublattices that sustain sizable local moments. Guided by these trends, we screen tetragonal $A_2XY$ Heusler compounds in space group $P4/mmm$. Among 307 structures, symmetry identifies 169 altermagnetic arrangements, of which 157 remain metallic altermagnets in DFT. Sixteen realize an altermagnetic collinear ground state, of which 15 are dynamically and mechanically stable and 10 also lie on or below the calculated thermodynamic hull. Six candidates exceed the CrSb splitting obtained under the same computational protocol, led by Co$_2$AlSc ($\Delta_\mathrm{max}=2.24$~eV) and Fe$_2$AlGe ($2.07$~eV). A Julliere-model estimate gives a tunneling magnetoresistance of up to $203\%$ at the Fermi level for Co$_2$AlSc. These results identify a chemically tunable family of metallic $d$-wave altermagnets and demonstrate how interpretable machine learning can guide targeted first-principles searches.
Read the full abstract
Analysis of PDEs
Parabolic defect measures for Harmonic Maps with Free Boundary
We introduce defect measures for the heat flow of a class of harmonic maps with free boundary and initiate their analysis. Our flow is motivated by singular perturbations on the boundary of the domain, emphasizing structural differences between harmonic maps with free boundary and their classical counterpart. The harmonic maps we consider are a reformulation of maps introduced by Da Lio and Rivière and arise in several problems in the spectral geometry of Steklov eigenvalues. Contrary to the classical case, the flow we consider exhibits features of global nature and the blow-up analysis turns out to be much more challenging.
Read the full abstract
Econometrics
Homophily and transitivity in dynamic network formation
In social and economic networks linked agents often share connections in common. There are two competing explanations for this phenomenon. First, agents may have a structural taste for transitive links - the returns to linking may be higher if two agents share a common connection. Second, agents may assortatively match on unobserved attributes, a process called homophily. We study parameter identifiability in a simple model of dynamic network formation with both effects. Agents form, maintain, and dissolve links over time to maximize utility. The return to linking may be higher if agents share connections in common. A pair-specific utility component allows for arbitrary homophily on time-invariant agent attributes. We derive conditions under which it is possible to detect the presence of a taste for transitivity in the presence of assortative matching on unobservables. We leave the joint distribution of the initial network and the pair-specific utility component, both high dimensional nuisance parameters, unrestricted. Our identification result is constructive, suggesting an analog estimator, whose finite and (single) large network properties we characterize. We show, via examples, the delicacy of information accumulation in the single (large) network setting.
Read the full abstract
High Energy Physics (Phenomenology)
Reliable Hybrid Neural Surrogates for Multidimensional Grids: Application to Double Parton Distributions
The conventional grid-based approach used for storing and evaluating ordinary parton distribution functions (PDFs), as implemented for example in LHAPDF, becomes much more demanding for double parton distributions (DPDs). While a collinear PDF depends on one longitudinal momentum fraction and one factorization scale, an unequal-scale DPD depends on two momentum fractions and two independent factorization scales for every parton-flavor combination. Direct tabulation of this four-dimensional dependence therefore requires considerably more disk space and runtime memory. In this work, we develop a hybrid neural-network method for compact storage and fast evaluation of $y$-independent unequal-scale DPDs. The network compresses the reference grid by reproducing most of its values, while the original values are stored only at points where the neural prediction does not reach the required accuracy. In this way, the neural model captures the bulk of the grid, with a sparse table covering the remaining difficult points. For the GS09-based dense grid studied here, only $0.694\%$ of the active flavor values need to be stored in the sparse table. Compared with the compressed dense grid, the hybrid method reduces the disk footprint by a factor of $10.6$, peak memory use by a factor of $31.6$, and initialization time by about a factor of $10.6$. At the same time, its evaluation throughput reaches about $85\%$ of the dense-grid rate. The method is implemented natively in C++ within PDFxTMDLib and requires neither Python nor PyTorch at runtime. The hybrid scheme therefore offers much lower memory and startup costs, with only a modest reduction in evaluation speed.
Read the full abstract
Astrophysics of Galaxies
Recovering turbulent velocity statistics from noisy integral-field spectroscopy
Turbulence plays a critical role in the evolution of H II regions, yet recovering its statistical properties from observations remains challenging, particularly when centroid velocities are affected by instrumental noise. We present a methodology for recovering the second-order velocity structure function from intermediate spectral-resolution integral-field spectroscopy and apply it to VLT MUSE observations of the Orion Nebula. The plane-of-sky velocity field is characterized by fitting a simple parametric model to the observed structure function, allowing the turbulent velocity variance, correlation length, power-law slope, and noise level to be estimated simultaneously. We investigate the trade-off between instrumental noise and spatial resolution through systematic spatial binning and validate the recovered turbulent parameters against previously analyzed high spectral-resolution KPNO echelle observations. We find that spatial binning improves the empirical structure function provided that the bin size remains below approximately $0.05 r_0$, while the fitted model recovers consistent turbulent parameters across all binning levels. The MUSE results agree closely with those from the KPNO data, demonstrating that intermediate-resolution integral-field spectroscopy can recover reliable turbulence statistics despite its relatively poor velocity resolution. We also identify modest but systematic differences in the turbulent properties traced by emission lines of different ionization potential, reflecting the geometric and ionization structure of the nebula. Our methodology provides a robust framework for extracting turbulent velocity statistics from noisy integral-field spectroscopic observations.
Read the full abstract
Software Engineering
Yesterday's Interviews for Today's Engineers: Retrospective Perceptions and a Work-Aligned Hiring Framework (2005-2026)
Technical hiring has increasingly become a skill system of its own, requiring candidates to prepare for processes that only partially resemble the position being filled. This paper reconstructs how the software-engineering community remembers that process across four eras from 2005 to 2026. Retrospective ratings from 911 respondents show a history of expanding interview work and burden, while an analysis of public preparation questions finds that most tested knowledge has a plausible use in an ordinary software-engineering project. The contradiction is only apparent: relevant topics can still form an invalid process when they are accumulated beyond the vacancy, presented under artificial constraints, or selected and interpreted without a common job-derived standard. The paper responds with the Ticket-Based Pragmatic Assessment Framework, a prescriptive reference framework that derives a bounded assessment from declared responsibilities, near-term tickets, and project-entry work. It preserves normal tools, defines three explicit evidence gates, limits repeated sessions, and stops when the decision is supported. The objective is to make an interview a compressed version of the job rather than a parallel profession that candidates must learn merely to obtain it.
Read the full abstract
Machine Learning
Data-Efficient Agentic Graph Domain Adaptation via Reliability-Aware Prototype Learning
Agentic learning systems are often required to adapt after deployment by observing new data and reusing prior knowledge under limited supervision or feedback. For graph-structured prediction, Graph Domain Adaptation (GDA) naturally instantiates this setting by transferring knowledge from labeled source graphs to unlabeled target graphs under distribution shifts. However, most GDA methods assume sufficient labeled source graphs, which becomes restrictive in data-efficient agentic settings where only limited source evidence can be retained. Under such constraints, source semantics become unreliable, leading to unstable source anchoring, uncertain target association, and fragile targetmarginal calibration. To address these challenges, we propose DEAG, a reliability-aware prototype learning framework for data-efficient agentic GDA. DEAG estimates class reliability from retained source support and embedding compactness, and constructs stable reusable source anchors by blending empirical prototypes with classifier directions. Guided by these anchors, DEAG performs prototype-aware soft target association and aligns confidence-weighted target centers with source semantics. A source-prior regularizer further sharpens target predictions while keeping the target marginal consistent with retained source evidence. Experiments on graph benchmarks with diverse domain shifts show that DEAG improves average adaptation performance over competitive GDA baselines under the same source-data budget.
Read the full abstract
Differential Geometry
Invariant almost Hermitian structures on the maximal real flag manifold of type $\mathcal{A}_3$
We study the Gray-Hervella classes of invariant almost Hermitian structures on the maximal real flag manifold of type $\mathcal{A}_3$. We first obtain an explicit parametrization of all invariant almost Hermitian structures and establish necessary and sufficient conditions for an invariant almost Hermitian structure to belong to each Gray-Hervella class. For each nonempty class, we give an explicit parametrization of the corresponding set of structures, describe it as a smooth manifold and determine its dimension.