Search arXiv⌕ Search

SEARCH · Search arXiv

Search Search arXiv

Search indexed arXiv papers on artificial intelligence, large language models, computer vision and robotics. Read source abstracts and follow links to arXiv.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 145 records · Page 8Linked to original sources
arXiv2026

Curvature-stabilized three-dimensional spiral states and local-field nucleation of skyrmions on spherical caps

The combination of polystyrene particles and multilayers with chiral interfacial interactions offers an interesting platform to study the interplay between topology, chirality, and curvature. Here, we investigate Pt/Co/Ta multilayers sputtered onto self-assembled polystyrene particles with a diameter of 919 nm, which show a magnetic reversal distinctly different from that of planar films. Using non-invasive scanning transmission X-ray microscopy and in-vacuum magnetic force microscopy (MFM) with low-moment tips, we observe maze domains in the planar films, whereas the curvilinear films host three-dimensional spiraling stripe states. Repeated scanning with a conventional higher-moment MFM tip ruptures these spirals into metastable skyrmions at the top of the particles. Micromagnetic simulations of spherical caps reveal that increasing curvature renders a single spiral stripe energetically favorable over an isolated skyrmion, and that this stabilization is predominantly of magnetostatic origin. Our results identify the spiral states as intrinsic to the curvilinear geometry and the skyrmions as metastable states nucleated by local magnetic fields, underlining the value of non-invasive imaging to distinguish intrinsic curvature effects from probe-induced magnetic textures.

cond-mat.mes-hall↗
arXiv2026

Bubble curtains in a lock-exchange flow: the importance of transient dynamics in the curtain-driven regime

Bubble curtains are line bubble plumes that are used to mitigate saltwater intrusion in ship locks. When the lock gate that separates saline seawater from fresh river water is opened, a lock-exchange flow develops. Installing a bubble curtain at the gate location disrupts this flow and reduces saltwater infiltration. For real-world applications, it is important to quantify how effective the bubble curtain is as a function of the key governing parameters. To this end, we performed multiphase large-eddy simulations that faithfully reproduce earlier experimental results including the two distinct operating regimes: the breakthrough regime and the curtain-driven regime. This paper focuses on the curtain-driven regime and seeks to clarify how the effectiveness of bubble curtains evolves over time. The detailed spatial and temporal data from the simulations, together with the ability to systematically vary the governing parameters, enabled us to overcome several limitations inherent in previous experiments. Furthermore, the simulations were used to obtain parameter values to build a semi-analytical model. Both the simulations and the semi-analytical model successfully capture and elucidate the time evolution of the density field and of the bubble curtain's effectiveness. The findings highlight that the time elapsed since the gate opening and the transient dynamics play a crucial role in determining the performance of bubble curtains for mitigation of salt intrusion.

physics.flu-dyn↗
arXiv2026

FermiGrad-SVD: An LLM Compression Pipeline with Global Rank Selection and Gauge Fixing

Low-rank factorization of weight matrices is a promising approach to compress large language models (LLMs), but its practical effectiveness depends on allocating ranks across layers and efficiently representing the resulting factors. We introduce three complementary methods for improving SVD-based LLM compression. First, FermiGrad formulates global layer-wise rank allocation as a differentiable optimization problem by replacing discrete singular-value truncation with continuous Fermi gates. Because FermiGrad optimizes only rank parameters over precomputed low-rank factors, it avoids repeated activation decompositions and backpropagation through SVD operations. Second, the Targeted Compressor extends FermiGrad with a task-specific objective that directs limited model capacity toward a selected domain. At 60% parameter density for Llama-3.1-8B-Instruct, SciQ-targeted compression achieves 60% MMLU accuracy, exceeding the reported quantized Dobi-SVD result by 10 percentage points, while deliberately trading performance on unrelated tasks for domain specialization. Third, PivGa provides a lossless secondary compression of low-rank factors by exploiting the gauge freedom in their parametrization. The proposed framework is modular, computationally efficient, and applicable at scale, as demonstrated by compressing Llama-3.3-70B-Instruct to 50% parameter density.

cs.LG↗
arXiv2026

Benchmarking and Learning Object-Centric Procedural Understanding in Instructional Videos

Procedural activities are fundamentally driven by object state transitions, yet existing instructional video benchmarks remain action-centric and cannot evaluate whether models reason about how objects evolve toward task completion. In this work, we introduce ProcObject-10K, the first benchmark that jointly evaluates object-centric reasoning and temporal evidence grounding in instructional videos, across both egocentric and exocentric views. It comprises 10,522 open-ended VideoQA pairs grounded in 1,799 videos, spanning 137 tasks across 9 domains and five reasoning types covering preconditions, state evolution, counterfactuals, mistakes, and readiness. Benchmarking 13 leading MLLMs reveals a substantial answering-grounding gap: models produce plausible answers while failing to localize the supporting evidence (mIoU < 45%), exposing their reliance on linguistic priors rather than recognizing and reasoning about the object dynamics. As a step toward closing this gap, we further provide an object-centric supervised fine-tuning baseline with pseudo object-level supervision and spatial-temporal constraints. Models fine-tuned on ProcObject-10K not only improve on the benchmark itself, but also transfer effectively to other grounded VideoQA and embodied planning tasks. The dataset, annotations, and evaluation toolkit have been publicly released on our project page to support future research on object-centric procedural understanding: https://wenliangguo.github.io/ProcObject-10K-Page.

cs.CV↗
arXiv2026

Hereditary History-Preserving Bisimilarity: Characterizations via Backward Ready Multisets

We devise two complementary characterizations of hereditary history-preserving bisimilarity (HHPB): a denotational one, over stable configuration structures, and an operational one, in a reversible process calculus. Our characterizations rely on forward-reverse bisimilarity (FRB) augmented with backward ready multiset equality, which is able to distinguish between autoconcurrency and autocausation. With respect to previous work, we thus shift the emphasis from uniquely identifying events, as done in other characterizations, to counting occurrences of identically labeled events associated with incoming transitions. This yields a more lightweight behavioral equivalence than HHPB: our bisimulations are still binary relations instead of ternary ones and do not need to incrementally build any labeling- and causality-preserving bijection for occurred events. We show that our characterizations are valid in the absence of non-local autoconflicts. We then study the logical foundations of these characterizations by relating event identifier logic, which captures HHPB, and backward ready multiset logic, developed for our extension of FRB.

cs.LO↗
arXiv2026

Automorphisms of Sedenions

This paper extends the octonion calibration in Clifford algebra to two related sedenion calibrations. Associative calibrations map quaternion rings in Clifford algebra to those in Cayley-Dickson algebras, with octonions consisting of seven rings and sedenions having 35 rings. A new non-associative calibration is found that is related to the sedenion associative calibration but is invertible, providing a classification of the ideals of Clifford algebra and hence of 8 x 8 matrices.This leads to new subalgebras of Clifford algebras and a thorough analysis of the possible automorphisms of sedenions is applied using the Spin and Pin groups. While the calibrations and automorphisms of octonions have minimal divergence, it is found sedenions extend Lie exceptional algebra, G2, to a Lie algebra of rank 49. This work clarifies and extends Schafer's and Brown's sedenion automorphisms. Brown introduced automorphisms that change the sedenion representation, unlike G2, but Clifford algebra analysis finds there are 92 such pure representations. The non-associative calibrations generate an infinite series of new Clifford subalgebras that match the finite geometry PG(N,2) series, discovered by Gino Fano, and connect the projections of simplices with those from hyper-cubes. The visual representation of sedenions as the PG(3,2) tetrahedron, representing the 15 Fano planes of sedenions, allows the sedenion power-associative loops to be visually distinguished from the octonion loops. The visual representation of sedenions may be important for particle physics and Clifford algebra provides analysis tools for the PG(N,2) series that uncovers the hidden structure of Cayley-Dickson algebras within an associative algebra.

math.RA↗
arXiv2026

Quantifying displacement: an urban expansion consequence via persistent homology

Population displacement is a housing-related involuntary residential dislocation. It has become increasingly widespread in many cities, particularly in neighbourhoods undergoing rapid economic and demographic change, and measuring it is essential to assess the social consequences of urban transformation and housing market pressures. Despite its relevance, quantifying displacement presents difficulties due to limited replicability across cities and time periods and the need to analyse long time spans: displacement is a gradual process, impossible to capture in one data snapshot. We introduce a novel tool to overcome these difficulties. Using publicly available address change data, we construct four cubical complexes simultaneously incorporating geographical and temporal information of people moving, and analyse using Topological Data Analysis tools. Finally, we demonstrate this method through a 20-year case study in Madrid, Spain. The results demonstrate its ability to reveal patterns consistent with population displacement and identify the neighbourhoods and years potentially affected--patterns not observable from yearly maps of the raw address change data.

cs.CG↗
arXiv2026

Symmetry-Aware Steering of Equivariant Diffusion Policies: Benefits and Limits

Equivariant diffusion policies (EDPs) combine the expressivity of diffusion models with the generalization and sample efficiency afforded by geometric symmetries. While steering these policies with reinforcement learning (RL) offers a promising fine-tuning mechanism beyond demonstration data, directly applying standard (non-equivariant) RL ignores the symmetries that EDPs are designed to exploit, leading to sample inefficiency and instability. We theoretically establish that the diffusion process of an EDP is equivariant, inducing a group-invariant latent-noise MDP which is well-suited for equivariant diffusion steering. Building on this, we introduce a principled symmetry-aware steering framework and compare standard, equivariant, and approximately equivariant RL strategies across tasks with varying degrees of symmetry. While we identify the practical boundaries of strict equivariance under symmetry breaking, exploiting symmetry during the steering yields substantial benefits---enhancing sample efficiency, preventing value divergence, and achieving strong policy improvements even when EDPs are trained from extremely limited demonstrations.

cs.LG↗
arXiv2026

Safe Learning for Contact-Rich Robot Tasks: A Survey from Classical Learning-Based Methods to Safe Foundation Models

Contact-rich tasks pose significant challenges for robotic systems due to inherent uncertainty, complex dynamics, and the high risk of damage during interaction. Recent advances in learning-based control have shown great potential in enabling robots to acquire and generalize complex manipulation skills in such environments, but ensuring safety, both during exploration and execution, remains a critical bottleneck for reliable real-world deployment. This survey provides a comprehensive overview of safe learning-based methods for robot contact-rich tasks. We categorize existing approaches into two main domains: safe exploration and safe execution. We review key techniques, including constrained reinforcement learning, risk-sensitive optimization, uncertainty-aware modeling, control barrier functions, and model predictive safety shields, and highlight how these methods incorporate prior knowledge, task structure, and online adaptation to balance safety and efficiency. A particular emphasis of this survey is on how these safe learning principles extend to and interact with emerging robotic foundation models, especially vision-language models (VLMs) and vision-language-action models (VLAs), which unify perception, language, and control for contact-rich manipulation. We discuss both the new safety opportunities enabled by VLM/VLA-based methods, such as language-level specification of constraints and multimodal grounding of safety signals, and the amplified risks and evaluation challenges they introduce. Finally, we outline current limitations and promising future directions toward deploying reliable, safety-aligned, and foundation-model-enabled robots in complex contact-rich environments. More details and materials are available at our \href{ https://github.com/jack-sherman01/Awesome-Learning4Safe-Contact-rich-tasks}{Project GitHub Repository}.

cs.RO↗
arXiv2026

How to Tropicalize a non-Archimedean Lattice

We present a method for computing the tropicalization of a lattice over the valuation ring of a non-Archimedean field. Our construction relies on the entropy polynomial, a tropical polynomial whose coefficients are determined by the valuations of minors of a matrix representing the lattice. We show that the tropicalization of the lattice is the support of a polyhedral complex defined through this polynomial. Moreover, we prove that this tropicalization is a finitely generated tropical semimodule over the nonnegative tropical scalars and provide explicit generators in terms of the entropy vector of the lattice. Finally, we extend our construction to valuated bimatroids, establishing analogous results beyond the realizable setting.

math.CO↗
arXiv2026

Irreducible groups and ergodicity in the boundary

We show that if $G$ is a real semi-simple Lie group, and $Γ$ is a discrete subgroup of $G$ containing a subgroup $Σ$ acting ergodically (in a strong sense) on the Furstenberg boundary of $G$, then $Γ$ is not isomorphic to a free product of $Σ$ with $\mathbb{Z}$. Moreover, if $Σ$ has algebraic entries, then $Γ$ has algebraic entries as well. As a consequence, we show that if all irreducible discrete subgroups of ${\rm SL}_2(\mathbb{R}) \times {\rm SL}_2(\mathbb{R})$ act ergodically on $\mathbb{S}^1\times \mathbb{S}^1 $, such groups cannot be free groups (or even Gromov hyperbolic). In the appendix, we discuss a connection between the existence of discrete irreducible groups and diophantine properties of Lie groups.

math.GR↗
arXiv2026

From Many Models, One: Macroeconomic Forecasting with Reservoir Ensembles

Model combination offers an alternative to selecting a single forecasting model, particularly when individual models are sensitive to random initialization and hyperparameter choices. We study online aggregation through Follow-the-Leader (FTL) and Hedge-type algorithms, with an application to ensembles of Multi-Frequency Echo State Networks (MFESNs) for macroeconomic forecasting. On the theoretical side, we derive Hoeffding- and Bernstein-type expected-regret bounds for FTL and decreasing Hedge under independent and $φ$-mixing expert losses. Under a positive gap between the best expert's expected loss and those of its competitors, the bounds are uniform in the number of forecasting rounds and exhibit logarithmic dependence on the number of experts. Monte Carlo experiments investigate the effects of temporal dependence and the separation between experts' expected losses. Empirically, we construct MFESN ensembles for forecasting U.S. quarterly GDP growth using mixed-frequency macroeconomic and financial indicators. The results show that adaptive aggregation, particularly FTL and AdaHedge, achieves substantial reductions in out-of-sample mean squared forecasting error relative to individual MFESN models and conventional forecast combinations.

econ.EM↗
arXiv2026

From ASTRID to BRAHMA -- The role of overmassive black holes in little red dots in cosmological simulations

We investigate the origin of Little Red Dots (LRDs) using a realization of the BRAHMA cosmological hydrodynamic simulation that produces overmassive black holes at $z \sim 4\text{-}8$. We generate mock JWST observations by modeling galaxy spectral energy distributions (including dust attenuation) together with AGN emission using cloudy-based templates, and test whether such systems reproduce the observed LRD population. We consider both a standard low-redshift AGN emission model and a gas-enshrouded AGN emission model, in which the AGN is embedded within unresolved dense gas that produces a strong Balmer break, and find that LRDs emerge in significantly larger numbers with the latter model. Assuming that all AGN reside in this gas-enshrouded phase, we recover LRD number densities ($Φ\sim 2 \times 10^{-4}~\mathrm{Mpc}^{-3}$), close to the observed abundance. This implies an effective gas-enshrouded duty cycle approaching unity, which is not predicted by current theoretical models. Current JWST constraints on the high-z black hole mass function suggest that this tension may arise from a deficit of low-mass ($M_{\rm BH}\lesssim10^7~M_\odot$) LRDs in BRAHMA. Applying the same gas-enshrouded model to the ASTRID simulation, which lacks overmassive black holes, we find that the LRD abundance remains $\gtrsim2~\rm dex$ below observations. Overall, while the overmassive black holes in this BRAHMA realization substantially enhance the production of LRD-like systems, reproducing the observed LRD abundance requires them to reside in dense gas environments with unrealistically high duty cycles. We therefore conclude that overmassive black holes are unlikely to be the sole origin of the observed JWST LRD population, and may instead account for only a subset of these systems.

astro-ph.GA↗
arXiv2026

Next-to-Leading Order corrections to the Next-to-Eikonal DIS structure functions

We compute next-to-leading order (NLO) corrections to next-to-eikonal (NEik) quark background contributions to DIS structure functions. Among NEik corrections, $t$-channel quark exchanges provide the lowest order contributions in $α_s$, and can be represented as insertions of the quark background field of the target. At NLO, we compute NEik corrections induced by both quark and gluon background fields, and suppressed by an explicit factor of $α_s$. We show that the NLO corrections to the NEik longitudinal structure function are finite, while those to the NEik transverse structure function exhibit rapidity and UV divergences. These divergences are analyzed, and the finite contributions are extracted.

hep-ph↗
arXiv2026

Consistent Bayesian meta-analysis on subgroup specific effects and interactions

Clinical trials commonly report effects on the full study population as well as for patient subgroups. Random-effects meta-analyses of subgroup-specific effects and treatment-by-subgroup interactions might be misaligned, especially when subgroups contribute differently to study populations. We show that, after contribution adjustment, subgroup meta-analysis can recover the same interaction inference regardless of whether interaction or subgroup data are used. Our framework for subgroup-data interaction meta-analysis has four advantages. First, it adjusts for variation in relative subgroup contribution, quantified by the within-trial information fraction. Second, it is robust to prevalence imbalance and variation. Third, it provides a self-contained model-based approach. Fourth, it can incorporate prior information into Bayesian interaction meta-analysis. The method demonstration uses an example application including 7 trials of disease-modifying therapies in relapsing-remitting multiple sclerosis; the contribution-adjusted model manages to reproduce the results from the interaction meta-analysis while yielding matching estimates for the subgroup effects. By controlling subgroup contribution while retaining subgroup interpretability, this approach allows for interaction decision-making when published subgroup data are available. Although the proposed model is presented in the Bayesian context, it can also be implemented in frequentist or likelihood frameworks.

stat.ME↗
arXiv2026

Welfare at Risk: Distributionally Robust Bounds for Policy Evaluation

This paper develops a framework to study how the welfare effects of policy interventions are distributed across individuals when those effects are not directly observed. We bound the superquantile of individual welfare changes by the superquantile of conditional average welfare changes, which is typically identified, revealing how gains and losses are spread across the population. We also bound the share of the population whose welfare loss from the policy exceeds any given threshold --- a distribution-free counterpart to average cost--benefit analysis, closer to a poverty-rate measure of policy harm than to a mean welfare effect. We further extend both bounds to allow for ambiguity in the analyst's knowledge of the distribution of observable characteristics: a non-robust bound estimated on one population can be systematically overoptimistic when applied to a different target population, while the robust version, with its ambiguity radius chosen commensurate with the shift, restores a valid guarantee. We illustrate the method in generalized Roy models with subjective participation costs, bounding not only the distribution of welfare changes in the overall population but also, in a result new to this literature, the entire distribution of welfare (and of participation costs) among those who select into treatment --- the treatment-on-the-treated distribution, not merely its mean. We connect these results to the multi-outcome IV framework of \cite{Heckman_Urzua_Vytlacil_2006_REStat}, showing our robust bounds are immune to the instrument-sensitivity problem they document

econ.EM↗
arXiv2026

Embodied Learning of Reward for Musculoskeletal Control with Vision Language Models

Discovering effective reward functions remains a fundamental challenge in motor control of high-dimensional musculoskeletal systems. While humans can describe movement goals explicitly such as "walking forward with an upright posture," the underlying control strategies that realize these goals are largely implicit, making it difficult to directly design rewards from high-level goals and natural language descriptions. We introduce Motion from Vision-Language Representation (MoVLR), a framework that leverages vision-language models (VLMs) to bridge the gap between goal specification and movement control. Rather than relying on handcrafted rewards, MoVLR iteratively explores the reward space through iterative interaction between control optimization and VLM feedback, aligning control policies with physically coordinated behaviors. Our approach transforms language and vision-based assessments into structured guidance for embodied learning, enabling the discovery and refinement of reward functions for high-dimensional musculoskeletal locomotion and manipulation. These results suggest that VLMs can effectively ground abstract motion descriptions in the implicit principles governing physiological motor control.

cs.RO↗