arXiv subjects

Yanan Wang

Publications and source records attributed to Yanan Wang.

At least 19 recordsLinked to original sources

arXiv2026

Radio Emission with Dust: A Bow-Shock Interpretation for the TDE Candidate AT 2019avd

Forward shocks produced by interactions between a single ejected blob and the circumnuclear medium (CNM) around a supermassive black hole are widely invoked to explain radio emission from tidal disruption events (TDEs). However, recent observational evidence such as rapidly rising radio emission, double-peaked broadband spectral energy distributions, and declining shock energies at late times poses challenges to this picture. Alternative scenarios have therefore been proposed, including bow shocks arising from outflows colliding with dense clouds, the production of new ejecta at late times, and variations in microphysical parameters. To investigate the roles of these factors, we analyze the long-lived radio flare of the TDE candidate AT 2019avd, for which dust has been identified through mid-infrared observations, using a radio dataset spanning more than five years. We find that a bow-shock scenario can account for the majority of the observed radio emission, while a forward shock, or an additional bow shock involving different clouds, may begin to dominate at late times. However, a forward-shock scenario remains viable when allowing for evolving microphysical parameters and more complex CNM structures. We discuss the remaining tensions in both the outflow-CNM and outflow-cloud interaction scenarios.

astro-ph.HE↗

arXiv2026

Early Near-Infrared Excess and Rapid Disk-Corona Evolution in the Tidal Disruption Event 2024aepd

We present multi-wavelength observations of the tidal disruption event (TDE) 2024aepd, spanning primarily the first $\sim$300 days after discovery. The X-ray spectrum is initially dominated by a thermal disk component accompanied by a hard excess. From $\sim$178 days onward, the spectrum becomes power-law dominated and subsequently hardens, indicating the rapid emergence and strengthening of a hot corona. A prominent near-infrared (NIR) excess is detected as early as $\sim40$ days. Its nearly flat power-law spectrum strongly deviates from the Rayleigh-Jeans tail of the UV-optical blackbody. Although a conventional dust-echo origin cannot be completely ruled out, free-free emission from a reprocessing photospheric envelope provides a more plausible explanation. Moreover, the UV-optical-to-NIR break shifts to higher frequencies as the density-profile index remains nearly constant, implying evolving reprocessing conditions within a broadly unchanged density structure. Together with AT2019azh and TDE 2025abcr, TDE 2024aepd is the third TDE reported to exhibit an early-time NIR excess. A larger sample with early-time NIR coverage is needed to determine whether such excesses are common among TDEs.

astro-ph.HE↗

arXiv2026

GEAR-Seg: A Grounded Explainable Agent for Reasoning Segmentation and Data Engine

Reasoning segmentation requires localizing targets based on complex, implicit queries. Current end-to-end models typically entangle perception and deduction into an opaque black box, severely limiting interpretability and scalability. To address this, we propose GEAR-Seg (Grounded Explainable Agent for Reasoning Segmentation), an explicitly decoupled agent that shifts the paradigm by translating visual pixels into dense, attribute-rich text. By decoupling class-agnostic segmentation, semantic description, and Large Language Model (LLM) deduction, GEAR-Seg transforms implicit reasoning into an explicit, trackable logic chain. As a zero-shot inference framework, it achieves highly competitive performance across diverse reasoning and fine-grained referring segmentation benchmarks. Furthermore, GEAR-Seg inherently functions as a highly scalable data engine. Utilizing this engine, we construct GEAR-131K, a massive benchmark (over 38k images, 656k QA-mask pairs) introducing a multifaceted taxonomy tailored for complex real-world manipulation-oriented reasoning. Finally, distillation experiments demonstrate that lightweight models supervised exclusively by our automated pipeline closely match the upper-bound performance of costly human-annotated baselines.

cs.CV↗

arXiv2026

Multi-band power color-color diagrams of three black hole X-ray binaries observed with Insight-HXMT

Power color-color diagrams (PCCDs) provide a useful diagnostic tool for studying the evolution of outbursts in black hole X-ray binaries. In this paper, we present power color-color diagrams of three sources (MAXI J1348-630, MAXI J1820+070 and Swift J1727.8-1613) observed with Insight-HXMT in a wide energy range of 2-80 keV. We compared the hue regions defined by RXTE, which are associated with different spectral states, with the Insight-HXMT results. We find that, for hard and hard-intermediate states, the trajectories of Insight-HXMT in the power color-color diagrams are generally consistent with those of RXTE. In the soft and soft-intermediate states, weak variability generally prevents robust hue constraints. Nevertheless, a few points in MAXI J1348-630 deviate from the RXTE-defined regions, possibly because of averaged variable power spectra and the presence of a type-A QPO. The trajectories of MAXI J1348-630 and MAXI J1820+070 exhibited roughly consistent patterns over different energy bands, whereas Swift J1727.8-1613 was an exception during its very high state, caused by an additional low-frequency component in the low-energy band. We found that the very high state can be identified through the power color-color diagram, exhibiting a hue similar to that of the hard-intermediate state but not forming a loop pattern. We also investigated the relationship between hue and hardness and found that, although they are generally anti-correlated, they provide consistent timing for the spectral state transitions.

astro-ph.HE↗

arXiv2026

Fairness Attacks on Recommender Systems

The unfairness of recommender systems has become a topic of concern due to its significant social and ethical implications. Although existing works have shown the effectiveness of attacks on the performance of recommender systems (e.g., promotion and demotion attack), the study of fairness attacks on recommender systems remains largely under-explored. To this end, we propose a novel structure-aware reinforcement learning-based fairness attack method designed to exacerbate the unfairness of target recommender systems. Specifically, we first employ a graph-based structure encoder to model the structural dependencies among the generated fake user-item interactions and the original user-item interactions. Then, we model the sequential dependency of the injected fake items using a recurrent neural network. Based on the learned structure-aware and sequence-aware representations of the fake user and item, the item selection policy attentively decides the next injected fake item. Since the target recommender system may employ fairness-aware training and leverage the user's sensitive attribute information, such as gender, we further designed a gender selection policy to decide the gender of the entire fake user profile. Both the item selection and gender selection policy are learned jointly in our proposed method. Finally, experimental results on four types of target recommendation models and two real-world datasets demonstrate the effectiveness of the proposed attack method in exacerbating the unfairness of recommender systems.

cs.IR↗

arXiv2026

OlfactProfile: Profile-Conditioned Odor Prediction from Audiovisual Content

Automated video-odor matching predicts scents aligned with audiovisual content for scent-enhanced media. Existing methods usually treat odor labels as determined only by scene content, but odor judgment also depends on individual olfactory profiles, including scent sensitivity, tolerance to unpleasant odors, and affective preference. Ignoring this observer context limits current systems' ability to predict scents that match perceived experience. We present OlfactProfile, a framework for profile-conditioned odor prediction from audiovisual content. Our results show that olfactory profiles are not beneficial by default: with matched feature backbones, naive profile concatenation and uniform profile modulation can degrade performance, while structured field-wise profile conditioning consistently improves prediction. Thus, the key challenge is not merely whether observer context is available, but how it is integrated into multimodal reasoning. To study this setting, we construct an audiovisual benchmark pairing temporally aligned odor annotations with annotator olfactory preference profiles. It contains 1,350 video clips, a 99-class scent vocabulary, and three semantic odor tracks: Foreground Odor, Background Odor, and Emotion Odor. We also propose OAR (Olfactory-Aware Routing), a multimodal fusion module that performs track-aware audiovisual routing with field-wise profile modulation, allowing profile dimensions to influence odor reasoning according to perceptual role. Experiments show that OlfactProfile outperforms supervised baselines and general-purpose multimodal large models, is competitive with odor experts in a small human comparison, and improves perceived scent fit in scent-enhanced applications without task-specific fine-tuning. Per-track analysis shows that gains are strongest for Background Odor and Emotion Odor, where observer-dependent judgment is most important.

cs.MM↗

arXiv2026

Multiwavelength Analysis of the Einstein Probe X-ray Transient EP240305a

We report multiwavelength observations of EP240305a, an uncatalogued X-ray transient detected by the Einstein Probe on March 5, 2024. The source exhibits distinct characteristics across the X-ray, optical, near-infrared, and radio bands. The soft X-ray observations show two significant flares lasting ~100-250 s, accompanied by rapid flux decay in a few days, and the optical and near-infrared data reveal a faint, candidate counterpart. In contrast, the radio observations expose a long-term spectral evolution from a self-absorbed to an optically thin state within two months, implying discrete jet ejection. We compare EP240305a with known classes of X-ray transients and find that it is unlikely to be associated with long-timescale transients such as jetted tidal disruption events or X-ray binaries. Its properties also disfavor a short-timescale stellar flare origin. Although the absence of optical spectroscopy prevents a redshift determination, the source exhibits properties similar to those of gamma-ray-dark gamma-ray burst-like transients, which may be associated with relativistic jets viewed off-axis or with choked jets. The discovery of EP240305a, along with other uncataloged transients detected by the Einstein Probe, underscores the scientific potential of highly sensitive X-ray survey telescopes and rapid-response multiwavelength follow-up observations in exploring the nature of atypical astronomical transients.

astro-ph.HE↗

arXiv2026

Long thermonuclear burst driven thermal-viscous instability of accretion disk: triggering an outburst-like X-ray flare

We report on NICER and MAXI observations of a long-duration thermonuclear X-ray burst and a subsequent outburst-like X-ray flare from the neutron star low-mass X-ray binary MAXI J0911--655. Prior to the burst, the source was in a persistent low/hard state with a power-law-dominated spectrum ($\Gamma \sim 1.7$) and a mass accretion rate of $\sim 1\%$ of the Eddington limit. The long burst, detected by MAXI on 2020 May 22 (MJD 58991.7101), was rapidly followed up by NICER. From time-resolved spectroscopy of the cooling tail, we estimate an exponential decay time of $\approx43$ minutes, the ignition column depth of $\approx0.1\times 10^{12}~{\rm g ~cm^{-2}}$, the burst fluence of $\approx 1.1\times 10^{-4}~{\rm erg~cm^{-2}}$, and the total energy release of $\approx1.2\times10^{42}$ erg. Approximately one day after the burst onset, the 0.5-10 keV light curve unexpectedly re-brightened, initiating an outburst-like flare. During the peak of this flare, the persistent power-law flux increased from its pre-burst level of $\sim0.27\times10^{-9}~{\rm erg~cm^{-2}~s^{-1}}$ to $1.4\times10^{-9}~{\rm erg~cm^{-2}~s^{-1}}$. This flux enhancement was accompanied by significant spectral softening, with the photon index increasing to $\Gamma \sim 2.2$. Subsequently, the flux decayed and the source returned to its baseline low/hard state. The observed timescales and energetics suggest that intense irradiation from the long burst amplified the ongoing thermal-viscous accretion process. This heating drove an inside-out heating front that temporarily enhanced the mass accretion rate, providing compelling observational evidence of a thermonuclear burst directly modulating the accretion dynamics of its surrounding disk.

astro-ph.HE↗

arXiv2026

Why Specialist Models Still Matter: A Heterogeneous Multi-Agent Paradigm for Medical Artificial Intelligence

The impressive performance of generalist large language models (LLMs) such as GPT and Claude in healthcare raises a critical question: will domain-specific medical specialist models become obsolete? We argue that the future of medical artificial intelligence (AI) lies not in building monolithic medical foundation models, nor in replacing human expertise, but in orchestrating collaboration among generalist LLMs, domain-specific specialist models, and clinicians. We propose HetMedAgent, a heterogeneous medical multi-agent framework that enables conflict-aware evidence fusion, uncertainty-based clinician intervention triggering, and adaptive threshold calibration. Experiments on three real-world clinical decision-making tasks demonstrate that the synergy between generalist LLMs and domain-specific specialist models significantly outperforms using either type of model alone, validating the irreplaceable value of specialist models in modality-specific analysis. HetMedAgent represents a shift from building medical LLMs or foundation models to multi-agent collaboration, achieving a balance between general reasoning capabilities and domain-specific precision.

cs.AI↗

arXiv2026

Qwen-Image-2.0 Technical Report

We present Qwen-Image-2.0, an omni-capable image generation foundation model that unifies high-fidelity generation and precise image editing within a single framework. Despite recent progress, existing models still struggle with ultra-long text rendering, multilingual typography, high-resolution photorealism, robust instruction following, and efficient deployment, especially in text-rich and compositionally complex scenarios. Qwen-Image-2.0 addresses these challenges by coupling Qwen3-VL as the condition encoder with a Multimodal Diffusion Transformer for joint condition-target modeling, supported by large-scale data curation and a customized multi-stage training pipeline. This enables strong multimodal understanding while preserving flexible generation and editing capabilities. The model supports instructions of up to 1K tokens for generating text-rich content such as slides, posters, infographics, and comics, while significantly improving multilingual text fidelity and typography. It also enhances photorealistic generation with richer details, more realistic textures, and coherent lighting, and follows complex prompts more reliably across diverse styles. Extensive human evaluations show that Qwen-Image-2.0 substantially outperforms previous Qwen-Image models in both generation and editing, marking a step toward more general, reliable, and practical image generation foundation models.

cs.CV↗

arXiv2026

DAPE: Dynamic Non-uniform Alignment and Progressive Detail Enhancement Techniques for Improving the Performance of Efficient Visual Language Models

In recent years, pre-trained visual-linguistic models have demonstrated tremendous potential, becoming a crucial foundational framework for numerous downstream tasks. However, the information density between text and images is not uniformly distributed. Existing methods often overlook the inherent and dynamic differences in information density and semantic scope between text tags and image blocks. These common uniform alignment strategies result in coarse-grained cross-modal interactions and loss of fine semantic details. Moreover, pursuing finer alignment typically requires substantial computational overhead, limiting practical model deployment. To address this challenge, this paper proposes a novel framework for dynamic cross-modal alignment with continuous detail introduction. First, we design a dynamically adaptive cross-modal matching mechanism that uses a learnable matching function to dynamically assign varying numbers and sizes of image tags to text tags of the same size but different information density, enabling more precise attention interaction. Second, we develop a continuous detail introduction module to progressively incorporate high-resolution visual feature enhancement into the alignment process. Extensive experiments across multiple benchmarks demonstrate significant improvements in the accuracy of various downstream tasks while reducing computational overhead.

cs.CV↗

arXiv2026

Deformed and undeformed localized wave solutions for the two-component (2+1)-dimensional Fokas-Lenells equation

In this paper, we focus on the two-component (2+1)-dimensional Fokas-Lenells equation, which models the propagation of ultrashort optical pulses in nonlinear media with multi-mode interactions and multi-dimensional effects. Firstly, we construct the determinant form of the generalized Darboux transformation (DT). Secondly, we obtain deformed solitons, deformed positons on the zero background and deformed breathers, deformed Y-shaped breathers on the nonzero backgrounds by DT method. Finally, the undeformed solutions including higher-order rogue wave solutions and breather-rogue wave solutions are derived by generalized DT. This work enriches the solution family associated with the equation, but also illustrates the efficiency of DT method in multi-dimensional and multi-component systems.

math-ph↗

arXiv2026

Align Generative Artificial Intelligence with Human Preferences: A Novel Large Language Model Fine-Tuning Method for Online Review Management

Online reviews have played a pivotal role in consumers' decision-making processes. Existing research has highlighted the significant impact of managerial review responses on customer relationship management and firm performance. However, a large portion of online reviews remains unaddressed due to the considerable human labor required to respond to the rapid growth of online reviews. While generative AI has achieved remarkable success in a range of tasks, they are general-purpose models and may not align well with domain-specific human preferences. To tailor these general generative AI models to domain-specific applications, finetuning is commonly employed. Nevertheless, several challenges persist in finetuning with domain-specific data, including hallucinations, difficulty in representing domain-specific human preferences, and over conservatism in offline policy optimization. To address these challenges, we propose a novel preference finetuning method to align an LLM with domain-specific human preferences for generating online review responses. Specifically, we first identify the source of hallucination and propose an effective context augmentation approach to mitigate the LLM hallucination. To represent human preferences, we propose a novel theory-driven preference finetuning approach that automatically constructs human preference pairs in the online review domain. Additionally, we propose a curriculum learning approach to further enhance preference finetuning. To overcome the challenge of over conservatism in existing offline preference finetuning method, we propose a novel density estimation-based support constraint method to relax the conservatism, and we mathematically prove its superior theoretical guarantees. Extensive evaluations substantiate the superiority of our proposed preference finetuning method.

cs.AI↗

arXiv2026

LILAC: Language-Conditioned Object-Centric Optical Flow for Open-Loop Trajectory Generation

We address language-conditioned robotic manipulation using flow-based trajectory generation, which enables training on human and web videos of object manipulation and requires only minimal embodiment-specific data. This task is challenging, as object trajectory generation from pre-manipulation images and natural language instructions requires appropriate instruction-flow alignment. To tackle this challenge, we propose the flow-based Language Instruction-guided open-Loop ACtion generator (LILAC). This flow-based Vision-Language-Action model (VLA) generates object-centric 2D optical flow from an RGB image and a natural language instruction, and converts the flow into a 6-DoF manipulator trajectory. LILAC incorporates two key components: Semantic Alignment Loss, which strengthens language conditioning to generate instruction-aligned optical flow, and Prompt-Conditioned Cross-Modal Adapter, which aligns learned visual prompts with image and text features to provide rich cues for flow generation. Experimentally, our method outperformed existing approaches in generated flow quality across multiple benchmarks. Furthermore, in physical object manipulation experiments using free-form instructions, LILAC demonstrated a superior task success rate compared to existing methods. The project page is available at https://lilac-75srg.kinsta.page/.

cs.RO↗

arXiv2026

Logics-Parsing-Omni Technical Report

Addressing the challenges of fragmented task definitions and the heterogeneity of unstructured data in multimodal parsing, this paper proposes the Omni Parsing framework. This framework establishes a Unified Taxonomy covering documents, images, and audio-visual streams, introducing a progressive parsing paradigm that bridges perception and cognition. Specifically, the framework integrates three hierarchical levels: 1) Holistic Detection, which achieves precise spatial-temporal grounding of objects or events to establish a geometric baseline for perception; 2) Fine-grained Recognition, which performs symbolization (e.g., OCR/ASR) and attribute extraction on localized objects to complete structured entity parsing; and 3) Multi-level Interpreting, which constructs a reasoning chain from local semantics to global logic. A pivotal advantage of this framework is its evidence anchoring mechanism, which enforces a strict alignment between high-level semantic descriptions and low-level facts. This enables ``evidence-based'' logical induction, transforming unstructured signals into standardized knowledge that is locatable, enumerable, and traceable. Building on this foundation, we constructed a standardized dataset and released the Logics-Parsing-Omni model, which successfully converts complex audio-visual signals into machine-readable structured knowledge. Experiments demonstrate that fine-grained perception and high-level cognition are synergistic, effectively enhancing model reliability. Furthermore, to quantitatively evaluate these capabilities, we introduce OmniParsingBench. Code, models and the benchmark are released at https://github.com/alibaba/Logics-Parsing/tree/master/Logics-Parsing-Omni.

cs.AI↗

arXiv2026

ReviveMoE: Fast Recovery for Hardware Failures in Large-Scale MoE LLM Inference Deployments

As LLM deployments scale over more hardware, the probability of a single failure in a system increases significantly, and cloud operators must consider robust countermeasures to handle these inevitable failures. A common recovery approach is to simply restart the LLM serving instance; however, this is costly in model-as-a-service (MaaS) inference settings, where reloading model weights and recompiling computation graphs can introduce significant delays to incoming requests. We propose ReviveMoE, a method for rapid failure recovery in large-scale LLM deployments without restarting the serving instance. ReviveMoE is designed to support both the traditional LLM architecture, which collocates MoE and attention on the same hardware, and the disaggregated architectures, which separate MoE from attention. Integrated into Huawei Cloud's MaaS, ReviveMoE is built on top of Huawei's xDeepServe serving platform and the XCCL communications library.

cs.DC↗

arXiv2026

Ca-MCF: Category-level Multi-label Causal Feature selection

Multi-label causal feature selection has attracted extensive attention in recent years. However, current methods primarily operate at the label level, treating each label variable as a monolithic entity and overlooking the fine-grained causal mechanisms unique to individual categories. To address this, we propose a Category-level Multi-label Causal Feature selection method named Ca-MCF. Ca-MCF utilizes label category flattening to decompose label variables into specific category nodes, enabling precise modeling of causal structures within the label space. Furthermore, we introduce an explanatory competition-based category-aware recovery mechanism that leverages the proposed Specific Category-Specific Mutual Information (SCSMI) and Distinct Category-Specific Mutual Information (DCSMI) to salvage causal features obscured by label correlations. The method also incorporates structural symmetry checks and cross-dimensional redundancy removal to ensure the robustness and compactness of the identified Markov Blankets. Extensive experiments across seven real-world datasets demonstrate that Ca-MCF significantly outperforms state-of-the-art benchmarks, achieving superior predictive accuracy with reduced feature dimensionality.

cs.LG↗

arXiv2026

Training and Benchmarking Code Generation for Physics-Inspired Animations

Large language models (LLMs) have been widely studied in areas such as mathematical reasoning, complex coding, and scientific problem solving. However, their ability to generate executable code that visually depicts physical scenarios and their qualitative dynamics remains underexplored. We propose SimuScene, the first systematic study that trains and evaluates LLMs on code generation for physics-inspired animations across 52 concepts spanning five physics domains. We build an automated data collection pipeline with human verification to ensure data quality. The resulting dataset contains 7,659 scenarios, including a 334-example human-verified test set. We evaluate 10 contemporary LLMs and find that even the strongest model achieves only a 21.5\% Avg@8 accuracy, demonstrating the difficulty of generating animations that are both executable and visually aligned with physical scenario descriptions. Finally, we introduce a reinforcement learning pipeline that uses visual rewards from code-generated videos to train text-only LLMs, with a vision-language model evaluating videos through verification questions. Experiments show that training with our data and video-based rewards improves LLM performance on physics-inspired animation generation.

cs.LG↗