arXiv subjects

Zhen Guo

Publications and source records attributed to Zhen Guo.

At least 37 records · Page 2Linked to original sources

arXiv2026

MICo-150K: A Comprehensive Dataset Advancing Multi-Image Composition

In controllable image generation, synthesizing coherent and consistent images from multiple reference inputs, i.e., Multi-Image Composition (MICo), remains a challenging problem, partly hindered by the lack of high-quality training data. To bridge this gap, we conduct a systematic study of MICo, categorizing it into 7 representative tasks and curate a large-scale collection of high-quality source images and construct diverse MICo prompts. Leveraging powerful proprietary models, we synthesize a rich amount of balanced composite images, followed by human-in-the-loop filtering and refinement, resulting in MICo-150K, a comprehensive dataset for MICo with identity consistency. We further build a Decomposition-and-Recomposition (De&Re) subset, where 11K real-world complex images are decomposed into components and recomposed, enabling both real and synthetic compositions. To enable comprehensive evaluation, we construct MICo-Bench with 100 cases per task and 300 challenging De&Re cases, and further introduce a new metric, Weighted-Ref-VIEScore, specifically tailored for MICo evaluation. Finally, we fine-tune multiple models on MICo-150K and evaluate them on MICo-Bench. The results show that MICo-150K effectively equips models without MICo capability and further enhances those with existing skills. Notably, our baseline model, Qwen-MICo, fine-tuned from Qwen-Image-Edit, matches Qwen-Image-2509 in 3-image composition while supporting arbitrary multi-image inputs beyond the latter's limitation. Our dataset, benchmark, and baseline collectively offer valuable resources for further research on Multi-Image Composition.

cs.CV↗

arXiv2026

The long-term outburst(s) of GPSV16: from an intermediate to a FUor classification

FU Ori outbursts are thought to play a key role in stellar mass assembly and in the chemistry of protoplanetary disks during the early formation of stars. However, uncertainties remain regarding the universality of these events and the physical mechanism driving the high-amplitude variability. In this work, we present an analysis of optical, near- and mid-IR photometry (ZTF, UKIDSS GPS, NEOWISE) and near-IR spectra (IRTF, Gemini) of the eruptive variable Class I YSO GPSV16. The YSO, associated with the HII region G71.52$-$00.38 ($d=3.61$~kpc), showed two outbursts, one with $ΔK_{\rm s}=2.2$~mag (2005-2012) and a second starting in 2016 with $ΔK_{\rm s}=5.6$~mag and accretion luminosity of $\sim$130 L$_{\odot}$. The outbursts displayed distinct spectroscopic characteristics: the first showed emission lines associated with a hot inner disk surface, whereas the second showed absorption lines arising from the cooler upper layers of a viscously heated disk. These features likely arose due to the different accretion rates reached during each outburst. The second outburst showed a two-stage mid-IR rise, requiring $\approx8.4$ years to reach peak brightness. The mid-IR rise also started 8 years before the onset of the optical outburst. The wavelength-dependent light curve points to an instability that is triggered at larger distances within the accretion disk and propagates inward. Assuming a propagation time of 8 years for the accretion wave, we estimate that the second outburst started at a distance of $r\sim0.4$~AU. These results show how long-term, multi-wavelength photometric monitoring can help identify the disk instabilities that trigger eruptions in YSOs.

astro-ph.SR↗

arXiv2026

Quantum Many-Body Simulations of Catalytic Metal Surfaces

Quantum simulations of metal surfaces are critical for catalytic innovation. Yet existing methods face a cost-accuracy dilemma: density functional theory is efficient but system-dependent in accuracy, while wavefunction-based theories are accurate but prohibitively costly. Here we introduce FEMION (Fragment Embedding for Metals and Insulators with Onsite and Nonlocal correlation), a systematically improvable quantum embedding framework that resolves this challenge by capturing partially filled electronic states in metals. FEMION combines auxiliary-field quantum Monte Carlo for local catalytic sites with a global random phase approximation treatment of nonlocal screening, yielding a scalable approach across diverse catalytic systems. Employing FEMION, we address two longstanding challenges: determining the preferred CO adsorption site and quantifying the H2 desorption barrier on Cu(111). Furthermore, our calculations demonstrate that the recently discovered 10-electron-count rule can also be extended to the single-atom catalysis processes on 3d metal surfaces, resolving the controversies arising from density functional theory calculations. We thus open a predictive, first-principles route to modeling complex catalytic systems.

physics.chem-ph↗

arXiv2026

InnoAds-Composer: Efficient Condition Composition for E-Commerce Poster Generation

E-commerce product poster generation aims to automatically synthesize a single image that effectively conveys product information by presenting a subject, text, and a designed style. Recent diffusion models with fine-grained and efficient controllability have advanced product poster synthesis, yet they typically rely on multi-stage pipelines, and simultaneous control over subject, text, and style remains underexplored. Such naive multi-stage pipelines also show three issues: poor subject fidelity, inaccurate text, and inconsistent style. To address these issues, we propose InnoAds-Composer, a single-stage framework that enables efficient tri-conditional control tokens over subject, glyph, and style. To alleviate the quadratic overhead introduced by naive tri-conditional token concatenation, we perform importance analysis over layers and timesteps and route each condition only to the most responsive positions, thereby shortening the active token sequence. Besides, to improve the accuracy of Chinese text rendering, we design a Text Feature Enhancement Module (TFEM) that integrates features from both glyph images and glyph crops. To support training and evaluation, we also construct a high-quality e-commerce product poster dataset and benchmark, which is the first dataset that jointly contains subject, text, and style conditions. Extensive experiments demonstrate that InnoAds-Composer significantly outperforms existing product poster methods without obviously increasing inference latency.

cs.CV↗

arXiv2026

Error-Aware Knowledge Distillation via Targeted Revision for Customer-Service Summarization

We introduce an Analyze-Revise-Finetune (ARF) pipeline that enables smaller open-source language models (LLMs) to surpass substantially larger proprietary models in customer service summarization tasks. The pipeline first analyzes and categorizes common errors in summaries produced by a teacher model (GPT-3.5), then performs a targeted revision using a compact editor model (Llama 3.1 70B) to generate high-quality, refined training data. Fine-tuning smaller student models (e.g., Llama 3.1 8B, QWen3 4B) on this refined data resulted in superior summarization performance compared to GPT-3.5. The ARF pipeline improves cost efficiency and data privacy while maintaining competitive accuracy, illustrating a generalizable framework for enhancing open-source LLMs across diverse downstream applications.

cs.CL↗

arXiv2026

Searching for White Dwarf Candidates Formed Through Binary evolution in Star Clusters

White dwarfs (WDs), the evolutionary endpoints of most stars, can form through both single-star and binary channels. While single-star evolutionary models enable reliable WD age estimates, binary evolution introduces interactions that can accelerate WD formation and result in a variety of exotic WDs, which may exhibit strong magnetic fields, rapid rotation, or even serve as potential gravitational wave sources. Such systems offer valuable insights into magnetic field generation, angular momentum evolution, and compact object physics. Star clusters, with their approximately coeval populations, allow precise age determination of member WDs. If a WD's total age derived from single-star evolution exceeds that of its host cluster, it likely indicates a binary origin. In this study, we use \textit{Gaia} 5D astrometry to identify 439 WD candidates in 117 open clusters, with 244 likely formed via binary evolution. We discuss the possibility of dynamical ejection for WDs meeting only 2D (proper motion space) membership criteria. Spectroscopic observations further reveal a subset with strong magnetic fields and rapid rotation, supporting their binary evolutionary origin.

astro-ph.SR↗

arXiv2026

Gaia24ccy: An outburst followed the footsteps of its predecessor

Accretion-driven outbursts in young stellar objects remain poorly understood, largely limited by a statistically small sample of closely followed-up events. This underscores the importance of a thorough exploration of each outbursting object. We studied a peculiar outbursting system, Gaia24ccy, which exhibited two $Δg \sim$ 3.8 mag outbursts in 2019 and 2024. The system consists of two unresolved, nearly identical, and rapidly rotating young stars: Gaia24ccy A (1.1419 days) and Gaia24ccy B (1.7898 days). Periodogram analyses just before the onset of the outbursts suggest Gaia24ccy B to be the outbursting component. Unlike any previously known EXor sources, the two outburst profiles show a very similar evolution: both rose at the same rate for the first 15 days, followed by many 'sub-bursts' on the timescale of 10-20 days. The 2019 outburst lasted 145-255 days, while the 2024 outburst persisted for 367 days. We infer the unstable region to lie at $r_{\rm trigger} \simeq 0.019-0.047$ au ($\sim5-12.3 R_\star$). The accreted mass per event $M_{\rm acc}\sim10^{-5} M_\odot$ can be provided by a compact inner-disk reservoir. The photometric rise/decay timescales and the mid-infrared color evolution favor a thermal-viscous trigger in a hot inner disk, while the appearance of rich emission-line spectra indicates concurrent magnetospheric compression - together best described by a hybrid picture. Finally, we explain the reddening of the mid-infrared color observed during the outburst as a consequence of the competing emission from the viscous disk and the photosphere.

astro-ph.SR↗

arXiv2026

CL-bench: A Benchmark for Context Learning

Current language models (LMs) excel at reasoning over prompts using pre-trained knowledge. However, real-world tasks are far more complex and context-dependent: models must learn from task-specific context and leverage new knowledge beyond what is learned during pre-training to reason and resolve tasks. We term this capability context learning, a crucial ability that humans naturally possess but has been largely overlooked. To this end, we introduce CL-bench, a real-world benchmark consisting of 500 complex contexts, 1,899 tasks, and 31,607 verification rubrics, all crafted by experienced domain experts. Each task is designed such that the new content required to resolve it is contained within the corresponding context. Resolving tasks in CL-bench requires models to learn from the context, ranging from new domain-specific knowledge, rule systems, and complex procedures to laws derived from empirical data, all of which are absent from pre-training. This goes far beyond long-context tasks that primarily test retrieval or reading comprehension, and in-context learning tasks, where models learn simple task patterns via instructions and demonstrations. Our evaluations of ten frontier LMs find that models solve only 17.2% of tasks on average. Even the best-performing model, GPT-5.1, solves only 23.7%, revealing that LMs have yet to achieve effective context learning, which poses a critical bottleneck for tackling real-world, complex context-dependent tasks. CL-bench represents a step towards building LMs with this fundamental capability, making them more intelligent and advancing their deployment in real-world scenarios.

cs.CL↗

arXiv2026

$π_\texttt{RL}$: Online RL Fine-tuning for Flow-based Vision-Language-Action Models

Vision-Language-Action (VLA) models enable robots to understand and perform complex tasks from multimodal input. Although recent work explores using reinforcement learning (RL) to automate the laborious data collection process in scaling supervised fine-tuning (SFT), applying RL to large-scale flow-based VLAs (\eg, $π_0$, $π_{0.5}$) remains challenging due to intractable action log-likelihoods raised from flow matching. We address this challenge with $π_{\texttt{RL}}$, featuring two technical approaches: (1) \textbf{Flow-Noise} models the denoising process as a discrete-time MDP with a learnable noise network for exact log-likelihood computation. (2) \textbf{Flow-SDE} integrates denoising with agent-environment interaction, formulating a two-layer MDP that employs ODE-to-SDE conversion for efficient RL exploration. We evaluate $π_{\texttt{RL}}$ across various benchmarks, with experiments demonstrating that RL yields significant performance improvements in both in-distribution and out-of-distribution settings.

cs.LG↗

arXiv2026

A near-infrared stellar atlas of the Galactic plane from the VVVX survey

The VISTA Variables in the Via Lactea eXtended (VVVX) ESO public survey observed the Galactic plane and the outer Galactic bulge in the near-infrared to mitigate the effects of extinction that severely limit optical observations of these regions. By significantly expanding the area covered by the original VVV survey, VVVX enables a deeper and broader exploration of the most obscured and crowded regions of the Milky Way. We aim to extend and complete our photometric catalogs of the entire Galactic plane region accessible from the southern hemisphere, focusing on the areas newly covered by the VVVX survey. Building on previous work, we applied point-spread function fitting techniques to detect point sources and extract their deep J, H, and Ks photometry across the VVVX footprint. The resulting catalogs were calibrated using astrometric and photometric reference data. Cross-matching between filters and epochs was used to ensure a high level of reliability and completeness. We produce a deep, highly complete near-infrared catalog of more than 700 million sources in the Galactic plane and outer Galactic bulge. When combined with our previous VVV atlas, the full catalog includes over 1.5 billion sources. The derived density maps and color-magnitude diagrams enable detailed studies of Galactic structure, extinction, and stellar populations, and highlight features such as the Carina arm tangency, the Sagittarius stream, and numerous star clusters. This extended atlas provides an unprecedented view of the innermost regions of the Milky Way. It is now publicly available through the VISTA Science Archive, offering a valuable resource for the astronomical community to investigate the structure and evolution of the Galactic disk and bulge.

astro-ph.GA↗

arXiv2026

Observational Signatures of Planetary Tidal Disruption Events Around Solar-Mass Stars

The tidal disruption of planets by their host stars represents a growing area of interest in transient astronomy, offering insights into the final stages of planetary system evolution. We model the hydrodynamic evolution and predict the multi-wavelength observational signatures of planetary TDEs around a solar-mass host, focusing on Jupiter-like and Neptune-like progenitors and examining how different eccentricities of the planet's pre-disruption orbit shape the morphology and emission of the tidal debris. We perform 2D hydrodynamic simulations using the FARGO3D code to model the formation and viscous evolution of the resulting debris disk. We employ a viscous alpha-disk prescription and include a time-dependent energy equation to compute the disk's effective temperature and subsequently derive the bolometric and multi-band photometric light curves. Our simulations show that planetary TDEs produce a diverse range of luminous transients. A Jupiter-like planet disrupted from a circular orbit at the Roche limit generates a transient peaking at $L_{bol} \approx 10^{38}$ erg s$^{-1}$ after a 12-day rise. In contrast, the same planet on an eccentric orbit (e=0.5) produces a transient of comparable peak luminosity but on a much shorter timescale, peaking in only 1 day and followed by a highly volatile light curve. We find that the effect of eccentricity is not universal, as it accelerates the event for Jupiter but delays it for Neptune. A robust "bluer-when-brighter" colour evolution is a common feature as the disk cools over its multi-year lifetime. The strong dependence of light curve morphology on the initial orbit and progenitor mass makes these events powerful diagnostics. This framework is crucial for identifying planetary TDEs in time-domain surveys.

astro-ph.EP↗

arXiv2026

Towards Understanding and Characterizing Vulnerabilities in Intelligent Connected Vehicles through Real-World Exploits

Intelligent Connected Vehicles (ICVs) are a core component of modern transportation systems, and their security is crucial as it directly relates to user safety. Despite prior research, most existing studies focus only on specific sub-components of ICVs due to their inherent complexity. As a result, there is a lack of systematic understanding of ICV vulnerabilities. Moreover, much of the current literature relies on human subjective analysis, such as surveys and interviews, which tends to be high-level and unvalidated, leaving a significant gap between theoretical findings and real-world attacks. To address this issue, we conducted the first large-scale empirical study on ICV vulnerabilities. We began by analyzing existing ICV security literature and summarizing the prevailing taxonomies in terms of vulnerability locations and types. To evaluate their real-world relevance, we collected a total of 649 exploitable vulnerabilities, including 592 from eight ICV vulnerability discovery competitions, Anonymous Cup, between January 2023 and April 2024, covering 48 different vehicles. The remaining 57 vulnerabilities were submitted daily by researchers. Based on this dataset, we assessed the coverage of existing taxonomies and identified several gaps, discovering one new vulnerability location and 13 new vulnerability types. We further categorized these vulnerabilities into 6 threat types (e.g., privacy data breach) and 4 risk levels (ranging from low to critical) and analyzed participants' skills and the types of ICVs involved in the competitions. This study provides a comprehensive and data-driven analysis of ICV vulnerabilities, offering actionable insights for researchers, industry practitioners, and policymakers. To support future research, we have made our vulnerability dataset publicly available.

cs.CR↗

arXiv2025

RLinf: Flexible and Efficient Large-scale Reinforcement Learning via Macro-to-Micro Flow Transformation

Reinforcement learning (RL) has demonstrated immense potential in advancing artificial general intelligence, agentic intelligence, and embodied intelligence. However, the inherent heterogeneity and dynamicity of RL workflows often lead to low hardware utilization and slow training on existing systems. In this paper, we present RLinf, a high-performance RL training system based on our key observation that the major roadblock to efficient RL training lies in system flexibility. To maximize flexibility and efficiency, RLinf is built atop a novel RL system design paradigm called macro-to-micro flow transformation (M2Flow), which automatically breaks down high-level, easy-to-compose RL workflows at both the temporal and spatial dimensions, and recomposes them into optimized execution flows. Supported by RLinf worker's adaptive communication capability, we devise context switching and elastic pipelining to realize M2Flow transformation, and a profiling-guided scheduling policy to generate optimal execution plans. Extensive evaluations on both reasoning RL and embodied RL tasks demonstrate that RLinf consistently outperforms state-of-the-art systems, achieving $1.07\times-2.43\times$ speedup in end-to-end training throughput.

cs.LG↗

arXiv2025

IGR J17091-3624: Newly Formed Periodic Dips and Multiwavelength Activity During the 2025 Outburst

The black hole low-mass X-ray binary (LMXB) candidate IGR J17091-3624 experienced a hard-state-only outburst in 2025. In this paper, we show that IXPE detected a series of intermittent X-ray dips, spanning a total interval of ~1 day. Subsequent observations with NICER, EP, NuSTAR, and Swift reveal that these dips recur with a period of 2.83$\pm$0.07 days and are accompanied by an increase in spectral hardness. This is the first time such quasi-periodic dipping behavior has been observed in this target since discovery. Our spectral analysis shows that the dips can be explained by obscuration from an ionized absorber characterized by an ionization parameter of $logξ$ ~1-3 erg cm s$^{-1}$ and an equivalent hydrogen column density of $N^{\rm zxipcf}_{\rm H}$~(1-30)$\times10^{22}$ cm$^{-2}$. The periodic reappearance of the absorber is likely caused by obscuring material located in the outer accretion disk, modulated by the binary orbital period. If confirmed, this period would suggest that the donor star in IGR J17091-3624 has deviated from the standard main-sequence evolutionary path and is likely a (partially) stripped giant. In the optical band, no significant periodicity or correlation with the X-ray dips was detected, whereas the radio counterpart exhibited a flat to steep spectrum, in contrast to the inverted spectrum typically observed during the hard state of LMXBs.

astro-ph.HE↗

arXiv2025

SN 2024aecx: a fast-evolving Type IIb supernova with a prominent shock-cooling peak

SN 2024aecx is a nearby ($\sim$11 Mpc) Type IIb SN discovered within $\sim$1 d after explosion. In this paper we report high-cadence photometric (typically 0.5$\sim$1 day) and spectroscopic follow-up observations, conducted from as early as 0.27 d post discovery out to the nebular phase at 158.4 d. We analyze the environment of SN 2024aecx and derive a new distance (11.3$\pm$1.1 Mpc), metallicity and host extinction. The light curve exhibits a hot and luminous shock-cooling peak at the first few days, followed by a main peak with very rapid post-maximum decline. The earliest spectra are blue and featureless, while from 2.3 d after discovery prominent P-Cygni profiles emerge. At nebular phase, the emission lines exhibit asymmetric and double-peaked profiles, indicating asphericity and/or early dust formation in the ejecta. Nebular spectral modelling indicates a blueshifted O-rich clump moving toward observer, and the $[\text{OI}]/[\text{CaII}]$ line ratio suggests an intermediate-mass progenitor. We simulated the progenitor and explosion using a two-component model of shock cooling and radioactive $^{56}$Ni heating; our model favors an extended, low-mass H-rich envelope with $M_{\mathrm{e}} = 0.04\pm{0.01} M_{\odot}$ and a low ejecta mass of$M_{\mathrm{ej}} = 1.55^{+0.18}_{-0.14} M_{\odot}$. And the nebular-phase spectra and light-curve modelling both suggest that it most likely originated from an intermediate-mass binary progenitor system. The comprehensive monitoring of SN 2024aecx, coupled with the detailed characterization of its local environment, establishes it as a benchmark event for probing the progenitors and explosion mechanisms of Type IIb SNe.

astro-ph.SR↗

arXiv2025

SN 2025coe: A Multiple-Peaked Calcium-Strong Transient from A White-Dwarf Progenitor

SN 2025coe is a calcium-strong transient located at an extremely large projected offset $\sim$39.3 kpc from the center of its host, the nearby early-type galaxy NGC 3277 at a distance of $\sim$25.5 Mpc. In this paper, we present multi-band photometric and spectroscopic observations spanning $\sim$100 days post-discovery. Its multi-band light curves display {multiple} distinct peaks: (1) an initial peak at $t \approx 1.6$ days attributed to shock cooling emission, (2) a secondary peak of $M_{R, \, peak} \approx$ $-$15.8 mag at $t \approx 10.2$ days powered by radioactive decay, and (3) a {possible} late-time bump at $t \approx 42.8$ days likely caused by ejecta-circumstellar material/clump interaction. Spectral evolution of SN 2025coe reveals a fast transition to the nebular phase within 2 months, where it exhibits an exceptionally high [Ca II]/[O I] ratio larger than 6. Modeling of the bolometric light curve suggests an ejecta mass of $M_{\rm ej} = 0.29^{+0.14}_{-0.15} \, M_{\odot}$, a $^{56}$Ni mass of $M_{\rm ^{56}Ni} = 2.4^{+0.06}_{-0.05} \times 10^{-2} M_{\odot}$, and a progenitor envelope with mass $M_e = 1.4^{+6.9}_{-1.2} \times 10^{-3} \, M_{\odot}$ and radius $R_e = 13.5^{+64.1}_{-11.1} \, R_{\odot}$. The tidal disruption of a hybrid HeCO white dwarf (WD) by a low-mass CO WD provides a natural explanation for the low ejecta mass, the small fraction of $^{56}$Ni, and the presence of an extended, low-mass envelope.

astro-ph.HE↗

arXiv2025

STAlloc: Enhancing Memory Efficiency in Large-Scale Model Training with Spatio-Temporal Planning

The rapid scaling of large language models (LLMs) has significantly increased GPU memory pressure, which is further aggravated by training optimization techniques such as virtual pipeline and recomputation that disrupt tensor lifespans and introduce considerable memory fragmentation. Such fragmentation stems from the use of online GPU memory allocators in popular deep learning frameworks like PyTorch, which disregard tensor lifespans. As a result, this inefficiency can waste as much as 43% of memory and trigger out-of-memory errors, undermining the effectiveness of optimization methods. To address this, we introduce STAlloc, a GPU memory allocator for deep learning frameworks that reduces fragmentation by exploiting the spatial and temporal regularity in memory allocation behaviors of training workloads. STAlloc introduces a novel paradigm that combines offline planning with online allocation. The offline planning leverages spatio-temporal regularities to generate a near-optimal allocation plan, while the online allocation handles complex and dynamic models such as Mixture-of-Experts (MoE). Built as a pluggable PyTorch memory allocator, STAlloc reduces fragmentation ratio on average by 85.1% (up to 100%) across both dense and MoE models, with negligible overhead. This enables more efficient, high-throughput training configurations and improves throughput performance by up to 32.5%.

cs.LG↗

arXiv2025

DarkMind: Latent Chain-of-Thought Backdoor in Customized LLMs

With the rapid rise of personalized AI, customized large language models (LLMs) equipped with Chain of Thought (COT) reasoning now power millions of AI agents. However, their complex reasoning processes introduce new and largely unexplored security vulnerabilities. We present DarkMind, a novel latent reasoning level backdoor attack that targets customized LLMs by manipulating internal COT steps without altering user queries. Unlike prior prompt based attacks, DarkMind activates covertly within the reasoning chain via latent triggers, enabling adversarial behaviors without modifying input prompts or requiring access to model parameters. To achieve stealth and reliability, we propose dual trigger types instant and retrospective and integrate them within a unified embedding template that governs trigger dependent activation, employ a stealth optimization algorithm to minimize semantic drift, and introduce an automated conversation starter for covert activation across domains. Comprehensive experiments on eight reasoning datasets spanning arithmetic, commonsense, and symbolic domains, using five LLMs, demonstrate that DarkMind consistently achieves high attack success rates. We further investigate defense strategies to mitigate these risks and reveal that reasoning level backdoors represent a significant yet underexplored threat, underscoring the need for robust, reasoning aware security mechanisms.

cs.CR↗