arXiv subjects

Yucheng Guo

Publications and source records attributed to Yucheng Guo.

At least 19 recordsLinked to original sources

arXiv2026

ODIN: Confirmation and 3D Reconstruction of Six Massive Protoclusters at Cosmic Noon

Protoclusters represent sites of accelerated galaxy formation and extreme astrophysical activity characteristic of dense environments. Identifying massive protoclusters and mapping their spatial structures are therefore crucial first steps in understanding how the large-scale environment influences galaxy evolution. We combine wide-field Ly$α$ imaging from the ODIN survey with extensive DESI and ancillary spectroscopy across the extended COSMOS and XMM-LSS fields ($\approx$14 deg$^2$) to search for massive protoclusters. We confirm six systems at $z\approx 2.4$ and $z\approx 3.1$, reconstruct their three-dimensional structures, estimate descendant halo masses, and, for one structure at $z\approx 3.12$, demonstrate that overlapping narrowband filters ($NB497$ and $N501$) provide accurate redshift tomography for emission-line galaxies. One protocluster at $z\approx 2.45$ overlaps with one of the LATIS tomographic fields, enabling direct comparison between galaxy and H {\sc i} overdensities traced by Ly$α$ forest absorption. Another at $z\approx 3.12$ hosts a massive quiescent galaxy ($M_{\ast} \approx 1.2 \times 10^{11}M_\odot$), indicating early quenching in a dense environment. By comparing Ly$α$ emission properties across environments, we find that protocluster galaxies exhibit higher median line fluxes and a deficit of faint emitters relative to the field. The effect is strongest when both 2D and 3D density information are combined, indicating that galaxies in the densest protocluster cores are most affected by environmental processes. This effect is stronger at $z\approx3.1$ than at $z\approx2.4$, suggesting possible redshift evolution.

astro-ph.GA↗

arXiv2026

Two domains of extended Lyman-alpha emission around galaxies: from local radiation to environmental regulation

We examine the relation between extended Ly$α$ halos around high-redshift galaxies and the main factors responsible for driving the emission in such halos, in particular at distances around and beyond one virial radius $r_\mathrm{vir}$. To reach the required surface brightness sensitivity we take advantage of the MUSE eXtremely Deep Field (MXDF) survey, allowing us to probe levels as faint as $\sim 10^{-20}$ erg cm$^{-2}$ s$^{-1}$ arcsec$^{-2}$ in individual Ly$α$ halos. Our sample consists of the 21 apparently core- and halo-brightest (yet intrinsically low luminosity $\log_{10}$L$_{\mathrm{Ly}α} < 42.3$ erg s$^{-1}$) Ly$α$ emitters (LAEs) in the MXDF at $3<z<4$, with typical virial radii around 20 kpc. We measure their radial surface brightness profiles out to 50 kpc (more than $2r_{\mathrm{vir}}$) and investigate the correlations between surface brightness and internal (star formation rates of the host galaxies, SFR) or external influences (environmental density, $δ+1$). We find a clear break in these correlations at radii around or just below $1r_{\mathrm{vir}}$. Below this break the emission correlates tightly with SFR (as expected) and not at all with $δ+1$. Beyond $\sim 1r_\mathrm{vir}$(20 kpc) we observe the opposite trend with no dependence on SFR, but an emerging correlation with $δ+1$. We compare our measurements with the expected integrated surface brightness from ultrafaint, individually undetected LAEs and find that the latter is insufficient to drive the observed correlation. We conclude that Ly$α$ emission from the outer halos is regulated by the surrounding environment, but originates mostly from diffuse gas rather than discrete sources.

astro-ph.GA↗

arXiv2026

The Small-scale Structures in the Wind of Messier 82

Small-scale multiphase structure plays a central role in galactic-wind evolution, yet the parsec-scale morphology and excitation of the warm ionised gas remain poorly constrained. We present deep HST narrow-band imaging of the southern wind of Messier~82 (M82) in Halpha, [OIII], [SII], and [NII], designed to resolve the warm ionised phase on parsec scales. The Halpha emission is detected to 2.1 kpc above the disk, while the fainter emission lines are detected over smaller radial extents, with [OIII] reaching 1.5kpc. We develop a filament-finding pipeline for the Halpha image and construct a quantitative catalogue of the filamentary structures. The wind forms a highly connected network of strands and knots, dominated by compact filaments with typical projected widths of 5.3pc and lengths of 9.5pc. Both the projected covering fraction and the line-flux contribution of the filamentary component decline with height, showing that the outer wind becomes increasingly dominated by diffuse emission. Optical line-ratio diagnostics indicate that the warm ionised gas occupies an intermediate excitation regime: photoionisation by the central starburst can energetically power the observed Halpha luminosity, while the systematic separation between filamentary and diffuse emission, together with the evolution of the line ratios with vertical distance, suggests an increasing contribution from shocks or other similar heating in the diffuse outer wind. These results show that separating filamentary and diffuse emission in high-resolution imaging provides a powerful way to connect the morphology, excitation, and multiphase structure of galactic winds.

astro-ph.GA↗

arXiv2026

Observation of mirror-odd and mirror-even spin texture in ultrathin epitaxially strained RuO2 films

Recently, rutile ruthenium dioxide (RuO$_2$) has attracted renewed interest due to expectations of prominent altermagnetic spin splitting. However, accumulating experimental evidence suggests that, in its bulk and thick-film forms, RuO$_2$ does not display any form of magnetic ordering. Despite this, the spin structure of RuO$_2$ remains largely unexplored in the ultrathin limit, where substrate-imposed epitaxial strain can be substantial. Here, we use spin-resolved angle-resolved photoemission spectroscopy, supported by ab initio calculations, to reveal the electronic structure of 2-nanometer-thick epitaxial RuO$_2$ heterostructures. We observe an unconventional spin texture characterized by the coexistence of mirror-even and mirror-odd momentum-dependent components. A comprehensive symmetry analysis rules out nonmagnetic origins of this spin texture. These findings suggest an emergent nonrelativistic spin structure enabled by epitaxial strain in the ultrathin limit, marking a distinct departure from the behavior of relaxed or bulk RuO$_2$. Our work opens previously unexplored perspectives for exploring symmetry-breaking mechanisms and spin textures in oxide heterostructures.

cond-mat.mtrl-sci↗

arXiv2026

Lyman-alpha haloes in the aftermath of reionisation

We present a comparative study of Ly$α$ haloes (LAHs) around low-luminosity (L$_{\mathrm{Ly}α}\lesssim 10^{42}$ erg s$^{-1}$) Ly$α$-emitting galaxies (LAEs) at very high redshifts $z\geq6$ and a reference sample at $z\sim 3$ covering a similar Ly$α$ luminosity and host galaxy stellar mass range. Using data from the Multi-Unit Spectroscopic Explorer (MUSE) at the ESO VLT, we extracted the samples such that at the different redshifts we obtain the same intrinsic surface brightness sensitivity, accounting for cosmological dimming. We detect extended Ly$α$ emission around 6 out of 18 high-$z$ LAEs in the MUSE eXtremely Deep Field (MXDF), more than doubling the number of known such objects at $z\geq6$. We obtain an only slightly higher individual LAH detection fraction of 40% among the lower redshift comparison sample. Yet the typical exponential scale lengths at $z\geq6$ are three times smaller than those at $z\sim3$. Stacking the LAEs with undetected haloes gives again drastically different results for the two samples, with a highly significant halo detection at $z\sim 3$ but no trace of extended Ly$α$ emission at $z\geq6$. We also find the Ly$α$ spectral line widths of the high-$z$ sample to be $\sim$2.5 smaller in comparison to the lower redshift objects. We discuss the potential mechanisms driving such strong changes. In a reionisation-driven scenario the higher neutral fraction in the intergalactic and circumgalactic media might lead to substantial scattering losses of escaping Ly$α$ radiation, leaving detectable only emission from the vicinity of the star-forming regions. In an alternative scenario the LAH properties might be linked more closely to the evolution of their host galaxies than previously thought.

astro-ph.GA↗

arXiv2026

Embodied Operators and Benchmarking: Toward Reusable and Deployable Embodied Intelligence Systems

Embodied intelligence systems require not only end-to-end policy models, but also reusable functional modules that transform multimodal observations, robot states, human demonstrations, and task contexts into structured representations, decisions, trajectories, control references, and system services. This work defines these modules as embodied operators and studies them as independent yet composable units in embodied intelligence pipelines. We clarify their definition boundary, emphasizing task semantics, standardized input-output contracts, deployability, reusability, and multi-layer optimizability. We further construct a taxonomy covering five categories: detection and segmentation, spatial localization and 3D understanding, hand motion recovery, embodied foundation models and task-decision operators, and planning, control, and system support operators. For each category, we summarize representative functions, technical paradigms, application roles, and practical limitations. Beyond taxonomy, we propose a multi-dimensional benchmark framework that evaluates embodied operators in terms of correctness, end-to-end efficiency, resource usage, temporal stability, portability, interface compatibility, deployment reliability, and downstream task utility. We also discuss workflow-level operator acceleration and open challenges in operator composition, data standardization, world models, VLA safety, edge deployment, and real-world application value. Overall, this work argues that embodied operators should be optimized and evaluated as holistic deployable components, providing a foundation for reusable, scalable, and verifiable embodied intelligence systems.

cs.AI↗

arXiv2026

Building a Scalable, Reproducible, Evaluatable, and Closed-Loop Simulation Environment Foundation for Embodied Intelligence

This paper presents a cloud-native simulation infrastructure framework for embodied intelligence that supports large-scale training, standardized evaluation, and simulation-based data collection. The framework unifies simulation environment generation, task execution, trajectory collection, model evaluation, data management, and cloud services into a scalable and reproducible platform. To address the high cost, limited scalability, and poor reproducibility of real-world robotic data collection, the framework adopts cloud-native technologies including elastic resource scheduling, containerized simulation, unified data management, and service-oriented system design, enabling efficient large-scale simulation for multi-model and multi-task workloads. Built on a four-layer architecture, the framework provides standardized environment assets, automated task generation, trajectory collection, benchmark evaluation, and closed-loop data optimization. It further integrates representative systems including D-VLA, RL-VLA3, Sword, and Pre-VLA to support scalable simulation, dynamic scheduling, visual augmentation, and real-time data filtering. We argue that cloud-native simulation infrastructure provides a unified foundation for data generation, model training, standardized evaluation, and real-world deployment, and will play a key role in the future development of embodied intelligence.

cs.RO↗

arXiv2026

Optimal Loss Allocation in a Mean-Field Model of Systemic Risk

We study a systemic-risk control problem in which a central planner allocates losses generated by bank defaults across the surviving institutions. Banks are modeled through their distances to default, evolving as absorbed Brownian motions with downward jumps induced by redistributed default losses. Unlike bailout models, the planner cannot inject external capital or reduce the aggregate loss, and the only admissible intervention is to decide how each endogenous loss is assigned among solvent banks. The objective is to maximize terminal system health, including survival mass as a leading special case and, more generally, increasing concave welfare functionals of the terminal distribution. Our main result identifies an optimal allocation rule with a simple economic interpretation: losses should be concentrated on the currently healthiest institutions. In discrete time, this rule takes the form of a cutoff or taxing-the-richest policy, which reduces banks above an endogenous threshold down to that threshold while leaving weaker banks untouched. We prove convergence of the time-discretized mean-field control problem as the allocation time step tends to zero and characterize the limiting problem as a singular mean-field control problem. The optimally controlled law is described by a reflected free-boundary formulation, in which the cutoff becomes the moving upper edge of the support, and the associated value function satisfies a Hamilton-Jacobi equation on Wasserstein space. Finally, we formulate the corresponding finite-particle control problem and show, under suitable assumptions, that the cutoff-controlled particle system converges to the continuous-time mean-field model. This provides a finite-system foundation for the optimal mean-field loss-allocation rule.

math.OC↗

arXiv2026

MusE GAs FLOw and Wind (MEGAFLOW) XIV: Background-Galaxy Absorption Reveals Kiloparsec-Scale Structure in the Cool Circumgalactic Medium

The properties of the cool ($T\sim10^4$~K) gas in the circumgalactic medium (CGM) are closely linked to the physical mechanisms that create and maintain this multiphase medium. The cool CGM is thought to consist of discrete clouds, whose characteristic size is unknown. Here we present a geometric and direct approach to constrain the coherence scale of these cool structures using stacked MgII absorption lines measured against extended background galaxies and effectively point-like background quasars, whose sizes are a few kpc and $\lesssim$ 0.01 pc, respectively. When the background-source size is smaller than the coherence scale of the foreground clouds, incomplete covering lowers the detection fraction and causes the median stacked absorption to differ from the mean. For stacked MgII absorption against background galaxies, the mean and median equivalent width (EW) profiles are broadly consistent. For stacked MgII absorption against background quasars, by contrast, the median and mean EW profiles differ significantly, and more so as the impact parameter increases beyond 100 kpc. Furthermore, we find a tentative trend that the median and mean EW profiles are broadly consistent for large background galaxies (median half-light radius $\approx 6.6$ kpc), but differ for small background galaxies ($\approx 1.5$ kpc). This indicates that MgII clouds have a coherence length of $\sim$2-7~kpc. Using a toy model in which the CGM is populated with discrete cool clouds, we show that the observed differences arise naturally from the combination of partial covering and beam averaging. Our results provide a new geometry-based measure of the small-scale structure of cool CGM gas.

astro-ph.GA↗

arXiv2026

Revealing Cosmic Ecosystems with the Hubble Space Telescope in 2030s and Beyond

Ultraviolet spectroscopy with the Hubble Space Telescope (HST) provides the most direct and sensitive probe of the disk-circumgalactic medium (CGM) interface at radii of 20 kpc, where galaxies exchange gas, metals, and energy with their surroundings. Many of the key diagnostics of the multiphase circumgalactic medium -- including H I, O VI, C II-IV, Si II-IV, N V, Ne VIII, and other metal transitions -- lie in the ultraviolet and are inaccessible from the ground, making HST the only observatory capable of making the required observations. By measuring the physical (column density, density), chemical (metallicity, ionization structure), and kinematical properties of the gas at the disk-CGM interface, UV absorption-line spectroscopy reveals how galaxies acquire fresh fuel, recycle enriched material, and drive feedback into their halos. When combined with spectroscopic characterization of the host galaxy's stellar populations and the feedback they generate (outflow velocity, mass loading), we will establish a direct understanding of how stellar populations enable circulation of gas and metals through the galactic ecosystem. HST's ultraviolet (UV) spectroscopic capability provides the only comprehensive observational pathways for uncovering the physical drivers that regulate galaxy growth and evolution in the low-redshift Universe.

astro-ph.IM↗

arXiv2026

LCEz4-M1: A Lyman Continuum Emitter Candidate at z = 4.444 in the MUSE Hubble Ultra Deep Field

High-redshift Lyman continuum emitters (LCEs) are crucial for understanding how galaxies ionize the neutral hydrogen in the epoch of reionization. However, detected LCEs at $z>4$ are quite rare. Here we report an LCE candidate at $z = 4.444$, dubbed LCEz4-M1, which is one of the highest-redshift LCE candidates currently known. The redshift is determined from the Ly$α$ emission line detected in the VLT/MUSE spectrum. The Lyman continuum (LyC) signal is detected independently in the \emph{Hubble Space Telescope} (HST) F435W image and the VLT/MUSE spectrum at significances of $\simeq3.7~σ$ and $\simeq2.8-3.0~σ$, respectively. The LyC centroid is spatially consistent with the JWST/NIRCam continuum within the astrometric uncertainty. Adopting the maximum IGM transmission, we infer conservative lower-limit escape fractions of $f_{\rm esc}({\rm F435W}) = 0.82^{+0.13}_{-0.17}$ and $f_{\rm esc}({\rm MUSE}) = 0.75^{+0.18}_{-0.28}$. Using the combined JWST and MUSE data set, we characterize the physical properties and morphology of LCEz4-M1. In our fiducial JWST-only SED fit, the galaxy is compact but has a moderate current galaxy-integrated star formation surface density, $Σ_{\rm SFR}=0.38~M_{\odot}\,{\rm yr^{-1}\,kpc^{-2}}$, suggesting that it is not an extreme compact starburst under this fiducial interpretation. While we find no clear evidence for an ongoing major merger for LCEz4-M1, the presence of a faint companion ($\sim 0''5$) detected in the F277W band suggests a potential minor interaction. We also find that LCEz4-M1 may lie in a locally overdense region, although the environmental interpretation remains tentative. Finally, the low ${\rm SFR}_{10\,{\rm Myr}}/{\rm SFR}_{100\,{\rm Myr}}$ ratio, low Ly$α$ EW, and relatively weak rest-frame optical emission lines of LCEz4-M1 may indicate a post-burst LyC-leaking phase.

astro-ph.GA↗

arXiv2026

AdaptiveLoad: Towards Efficient Video Diffusion Transformer Training

In video generation models, particularly world models, training large-scale video diffusion Transformers (such as DiT and MMDiT) poses significant computational challenges due to the extreme variance in sequence lengths within mixed-mode datasets. Existing bucket-based data loading strategies typically rely on "equal token length" constraints. This approach fails to account for the quadratic complexity of self-attention mechanisms, leading to severe load imbalance and underutilization of GPU resources. This paper proposes \textit{AdaptiveLoad}, an integrated optimization framework consisting of two core components: (1) A dual-constraint adaptive load balancing system, which eliminates long-sequence bottlenecks by simultaneously limiting memory consumption and computational load ($B \times S^p \le M_{\text{comp}}$); (2) A fused LayerNorm-Modulate CUDA kernel, which utilizes a D-tile coalesced reduction strategy to increase throughput and alleviate memory pressure. Experimental results on the Wan 2.1 world model demonstrate that our method reduces the computational imbalance rate from 39\% to 18.9\%, improves peak VRAM utilization efficiency by 22.7\%, and achieves an overall training throughput increase of 27.2\%.

cs.DC↗

arXiv2026

D-VLA: A High-Concurrency Distributed Asynchronous Reinforcement Learning Framework for Vision-Language-Action Models

The rapid evolution of Embodied AI has enabled Vision-Language-Action (VLA) models to excel in multimodal perception and task execution. However, applying Reinforcement Learning (RL) to these massive models in large-scale distributed environments faces severe systemic bottlenecks, primarily due to the resource conflict between high-fidelity physical simulation and the intensive VRAM/bandwidth demands of deep learning. This conflict often leaves overall throughput constrained by execution-phase inefficiencies. To address these challenges, we propose D-VLA, a high-concurrency, low-latency distributed RL framework for large-scale embodied foundation models. D-VLA introduces "Plane Decoupling," physically isolating high-frequency training data from low-frequency weight control to eliminate interference between simulation and optimization. We further design a four-thread asynchronous "Swimlane" pipeline, enabling full parallel overlap of sampling, inference, gradient computation, and parameter distribution. Additionally, a dual-pool VRAM management model and topology-aware replication resolve memory fragmentation and optimize communication efficiency. Experiments on benchmarks like LIBERO show that D-VLA significantly outperforms mainstream RL frameworks in throughput and sampling efficiency for billion-parameter VLA models. In trillion-parameter scalability tests, our framework maintains exceptional stability and linear speedup, providing a robust system for high-performance general-purpose embodied agents.

cs.AI↗

arXiv2026

An Updated Characterization of Luminous Ly{\alpha} emitters at the End of Reionization

We present a multi-wavelength physical characterization of 14 luminous Ly$\alpha$ emitters (LAEs) at $z\approx6$, integrating deep ground-based Magellan/M2FS spectroscopy with heterogeneous JWST/NIRCam broad- and medium-band imaging. Identified via strong Ly$\alpha$ lines with extreme Ly$\alpha$ luminosities of ${>}10^{42.6}$ erg s$^{-1}$, the sample exhibits very large rest-frame equivalent widths (${\gtrsim}100$ \AA) and steeply blue UV continua ($\beta_{\rm median}\simeq-2.2$, $-18.2>M_{\rm 1500}>-20.2$ mag). Crucially, the integration of NIRCam medium-band photometry (F410M) breaks the degeneracy between strong rest-optical nebular emission and Balmer breaks, resolving prior mass overestimations. The tightly constrained spectral energy distribution modeling demonstrates that these luminous LAEs tend to be unequivocally low-mass, ultra-young dwarf starbursts; half the sample is characterized by stellar masses of $M_* < 10^8 M_{\odot}$, ages $\lesssim10$ Myr, and negligible dust attenuation. We also map the production efficiency of ionizing photons and Ly$\alpha$ escape fractions ($f_{\rm esc}^{\rm Ly\alpha}$). The $f_{\rm esc}^{\rm Ly\alpha}$ values are exceptionally high, with a median of ${\gtrsim}40$%, increasing for the bluer UV continua. Finally, analyzing spatial offsets between the Ly$\alpha$ centroid and the stellar counterpart, we demonstrate empirically that internal dust content, rather than neutral hydrogen gas, dominate the suppression of Ly$\alpha$ radiative transfer. Our study reveals that strong Ly$\alpha$ emission of the luminous LAEs are generally attributed to both the vigorous starburst activities and the high $f_{\rm esc}^{\rm Ly\alpha}$. Resembling Lyman continuum leakers, these extreme dwarf systems function as highly efficient ionizing engines at the conclusion of the Epoch of Reionization.

astro-ph.GA↗

arXiv2026

NoiseGate: Learning Per-Latent Timestep Schedules as Information Gating in World Action Models

World Action Models (WAMs) are an emerging family of policies that tie robot action generation to future-observation modeling. In this work, we focus on the joint video--action modeling paradigm, where actions and imagined future observations are co-generated along a shared denoising or flow trajectory, so that perception, prediction, and control are coupled within one generative process. Existing WAMs typically realize this paradigm with a Mixture-of-Transformers (MoT), where video and action tokens interact through shared self-attention. This architecture can in principle assign a separate timestep $t_f$ to each predicted latent frame, yet current systems collapse this degree of freedom onto a single shared scalar $t$. Under the noise-as-masking view of Diffusion Forcing, this shared schedule imposes the unjustified prior that every predicted latent is equally reliable for action generation. We instead view the per-latent schedule as a \emph{learnable information-gating policy}: by changing a latent frame's noise level, the policy modulates the reliability of its Key/Value contribution to the action tokens. We propose \textbf{NoiseGate}, which combines independent per-latent timestep sampling during backbone training, a lightweight Gating Policy Network that emits per-latent time increments during denoising, and task-reward optimization that trains the schedule policy without hand-crafted shape priors. Built on a joint video--action MoT backbone, NoiseGate delivers consistent gains on diverse RoboTwin random-scene manipulation tasks.

cs.RO↗

arXiv2026

JoyAI-RA 0.1: A Foundation Model for Robotic Autonomy

Robotic autonomy in open-world environments is fundamentally limited by insufficient data diversity and poor cross-embodiment generalization. Existing robotic datasets are often limited in scale and task coverage, while relatively large differences across robot embodiments impede effective behavior knowledge transfer. To address these challenges, we propose JoyAI-RA, a vision-language-action (VLA) embodied foundation model tailored for generalizable robotic manipulation. JoyAI-RA presents a multi-source multi-level pretraining framework that integrates web data, large-scale egocentric human manipulation videos, simulation-generated trajectories, and real-robot data. Through training on heterogeneous multi-source data with explicit action-space unification, JoyAI-RA effectively bridges embodiment gaps, particularly between human manipulation and robotic control, thereby enhancing cross-embodiment behavior learning. JoyAI-RA outperforms state-of-the-art methods in both simulation and real-world benchmarks, especially on diverse tasks with generalization demands.

cs.RO↗

arXiv2026

Thousand-GPU Large-Scale Training and Optimization Recipe for AI-Native Cloud Embodied Intelligence Infrastructure

Embodied intelligence is a key step towards Artificial General Intelligence (AGI), yet its development faces multiple challenges including data, frameworks, infrastructure, and evaluation systems. To address these issues, we have, for the first time in the industry, launched a cloud-based, thousand-GPU distributed training platform for embodied intelligence, built upon the widely adopted LeRobot framework, and have systematically overcome bottlenecks across the entire pipeline. At the data layer, we have restructured the data pipeline to optimize the flow of embodied training data. In terms of training, for the GR00T-N1.5 model, utilizing thousand-GPU clusters and data at the scale of hundreds of millions, the single-round training time has been reduced from 15 hours to just 22 minutes, achieving a 40-fold speedup. At the model layer, by combining variable-length FlashAttention and Data Packing, we have moved from sample redundancy to sequence integration, resulting in a 188% speed increase; π-0.5 attention optimization has accelerated training by 165%; and FP8 quantization has delivered a 140% speedup. On the infrastructure side, relying on high-performance storage, a 3.2T RDMA network, and a Ray-driven elastic AI data lake, we have achieved deep synergy among data, storage, communication, and computation. We have also built an end-to-end evaluation system, creating a closed loop from training to simulation to assessment. This framework has already been fully validated on thousand-GPU clusters, laying a crucial technical foundation for the development and application of next-generation autonomous intelligent robots, and is expected to accelerate the arrival of the era of human-machine integration.

cs.RO↗

arXiv2026

Particle Systems with Local Interactions via Hitting Times and Cascades on Graphs

We introduce a family of particle systems on sparse graphs where local interactions occur via hitting times, providing a dynamic and tractable model for default cascades in large sparsely-connected financial networks. Building on the framework of Lacker, Ramanan and Wu (2023), we extend convergence theory to systems with singular interactions, capturing the abrupt and discontinuous nature of systemic events. We establish conditions for well-posedness through a minimality principle and connect fragility to dynamic percolation thresholds. Our analysis demonstrates continuity of the joint law of defaults with respect to local graph convergence, establishes convergence of empirical distributions, and characterizes the default time distribution in tree-like networks. This framework offers a rigorous and flexible foundation for modeling systemic risk in evolving financial systems, featuring continuous-time dynamics, heterogeneous and local interactions, and instantaneous default cascades.

math.PR↗