arXiv subjects

Siqi Wang

Publications and source records attributed to Siqi Wang.

At least 19 recordsLinked to original sources

arXiv2026

Sighted by Default: Addressing Implicit Vision Assumptions in Real-Time VLM Assistance for BLV Users

Vision-Language Model (VLM)-based assistance is reshaping independence for blind and low-vision (BLV) users, yet current tools fail in dynamic settings. While request-response architectures like BeMyAI impose prohibitive latencies, our formative study (N = 15) reveals that even real-time alternatives like Doubao fail due to a deeper structural problem: sighted-default bias -- the implicit assumption that users possess parallel visual access to their surroundings. This bias manifests as verbose, vision-centric narratives that overlook the serial nature of auditory perception, flooding the user's limited cognitive bandwidth with information that is neither timely nor actionable. To address this, we derive three design principles: Continuity, Conciseness, and Calibrated Honesty. We present VIA-Agent, which co-optimizes a specialized cognitive core with a low-latency Real-Time Communication (RTC) architecture for continuous bidirectional streaming. In a within-subjects evaluation (N = 9), VIA-Agent matched Doubao's success rate, significantly reducing mean task time by 21.4% (91.7s vs. 116.7s) and conversational turns from 5.9 to 4.3 while achieving higher trust.

cs.HC↗

arXiv2026

OpenLoopEvolve: A Verifiable Self-Evolution Framework for Loop Policies in Long-Horizon Complex Tasks

Long-horizon complex tasks require agents to repeatedly observe states, formulate plans, invoke tools, verify results, and recover from failures in continuously changing environments. However, such control experience often remains confined to a single context or a fixed prompt, and is difficult to accumulate and reuse across historical traces. This paper presents OpenLoopEvolve (OLE), a self-evolution framework centered on the Loop Policy. OLE represents an agent's observation, planning, memory, action, verification, recovery, stopping, and budget control as portable policy assets with versions and lineages, and provides online and offline evolution modes that can be selected according to practical needs: the online mode triggers candidate generation based on feedback from continuous operation, whereas the offline mode searches for candidate policies from archived traces and failure evidence. Both modes share an evolution mechanism consisting of autonomous proposals by a large language model, Champion--Challenger paired evaluation, and robust release. Policies released online are activated at a subsequent task boundary, monitored using subsequent feedback, and rolled back to their parent versions when degradation conditions are met. On the simulated business benchmark YC-Bench, both modes improve aggregate task performance, task success rate, and risk metrics relative to a fixed initial Loop Policy. The results indicate that treating the Loop Policy as a governable asset can support the accumulation, comparison, release, and reuse of control experience and improve agent performance on long-horizon complex tasks.

cs.AI↗

arXiv2026

Morphology Engineering of Mixed Ionic Electronic Conductors through Aqueous Phase Separation

The performance of organic electrochemical transistors (OECTs) is fundamentally governed by the interplay between ionic accessibility and electronic transport within organic mixed ionic-electronic conductors. Although increasing channel thickness enhances transconductance, it also prolongs ion transport, resulting in the well-known gain-speed trade-off. Here, we demonstrate that engineering the internal morphology of PEDOT:PSS:PEI films through pH-induced aqueous phase separation provides an effective route to mitigate this limitation. The resulting interconnected pore network promotes electrolyte penetration and increases the electrochemically addressable volume, while DMSO treatment and annealing enhance the continuity and ordering of the PEDOT-rich electronic phase. Consequently, porous OECT channels achieve a transconductance of 30 mS and a response time of 13 ms at an ultralow gate voltage of 0.05 V despite channel thicknesses exceeding 100 um. Comparison of films with comparable electronic conductivity but different pore architectures identifies morphology as the dominant factor governing device performance, supporting a transition from predominantly surface-limited modulation toward spatially distributed mixed ionic-electronic transport. Beyond demonstrating a scalable water-based fabrication strategy, this work establishes internal morphology as a design parameter that complements molecular structure and device geometry in organic mixed conductors, providing a general framework for the development of high-performance OECTs, soft bioelectronics, and future neuromorphic materials.

cond-mat.mtrl-sci↗

arXiv2026

Vorch-Omni: Multi-Task Orchestration of Sight and Sound

Recent advances in generative video modeling have enabled diverse generation, reference-based synthesis, extension, and editing, but existing approaches often rely on fragmented task-specific models. A general model must distinguish heterogeneous target, source, and reference signals to determine what to generate, preserve, or use as guidance, while reducing interference among tasks. Joint audio-visual generation further increases this challenge by introducing diverse conditioning and output configurations across modalities. We present Vorch-Omni, a unified multi-task framework for audio-visual synthesis based on an arbitrary-condition-to-arbitrary-output formulation. It flexibly treats video and audio signals as either conditioning inputs or generation targets. Token-level conditioning masks and task identifiers distinguish targets, source content, and references, while position types separate temporal context from independent conditions. To capture semantic and structural information, Vorch-Omni employs complementary visual conditioning pathways: a vision-language model interprets sampled frames with text instructions, and a video VAE encodes conditions into latent tokens for direct guidance. We further build a distributed data pipeline to curate diverse temporally aligned audio-visual clips, generate structured captions and metadata, and balance heterogeneous task distributions. Built on a single flow-matching diffusion transformer without task-specific architectural changes, Vorch-Omni supports over 10 tasks, including text-to-video, text-to-audio-video, image- and reference-conditioned generation, temporal extension, audio-driven generation, video transformation, and audio-visual editing. This unified framework provides a scalable foundation for general-purpose audio-visual generation and manipulation.

cs.CV↗

arXiv2026

GA-VINO: A Geometry-Aware Variational Physics-informed Neural Operator for Mindlin-Reissner Plates

Plate and shell structures are widely used in engineering fields. Rapid response prediction for such structures under complex geometries, heterogeneous materials, and varying loads is important for engineering design, but conventional numerical methods usually require repeated modeling and solution when the physical configuration changes. To address this issue, this study proposes a geometry-aware variational physics-informed neural operator (GA-VINO) for Mindlin-Reissner plates. GA-VINO represents the plate geometry using boundary point clouds and incorporates a material encoder, a load encoder, and a scalar-parameter branch to handle spatially random material fields, spatially varying pressure loads, and sample-level uniform parameters. Through multi-branch point cloud encoding and cross-attention, GA-VINO fuses geometric, material, loading, and query point information, and predicts the transverse deflection and rotations at arbitrary query locations. Unlike conventional data-driven neural operators, GA-VINO requires no labeled solution data during training. Instead, it minimizes a variational physics-informed loss constructed from the discretized total potential energy of the Mindlin-Reissner plate. Compared with grid-based neural operators, GA-VINO directly processes irregular point clouds and allows different physical fields to be discretized on different point sets, avoiding forced interpolation onto a common grid. The method is validated on multiple examples involving different geometries, material fields, and load distributions. The results show that GA-VINO achieves promising accuracy in deflection, rotation, gradient-sensitive, and energy-based metrics, completes full-field inference for new samples within milliseconds, and exhibits promising cross-geometry generalization capability.

cs.AI↗

arXiv2026

SlimPer: Make Personalization Model Slim and Smart

Transformer-style architectures are increasingly adopted for industrial recommendation systems, yet they inherit a design premise misaligned with the task: generative models rely on per-token autoregressive prediction, which justifies maintaining large intermediate tensors that scale with sequence length. In contrast, recommendation systems produce a single set of relevance scores for each pair without token-level supervision. Leveraging this observation, we propose SlimPer, which reformulates personalized ranking as iterative refinement of a compact, unified knowledge base. At each layer, the model selectively queries raw multi-modal user-side tokens, computes explicit relevance matching scores, and refines the knowledge base, all in O(N) per-layer cost with a fixed-size intermediate representation. As a result, model depth is decoupled from user history length, enabling deeper relevance understanding without proportional growth in compute or memory; request-only optimization further trims memory by sharing a single copy of user-side tokens across all candidate items. SlimPer unifies sparse, dense, and sequence features within a single backbone and provides inherent interpretability through its attention mechanism. Deployed on Instagram Reels and Feed, SlimPer yields measurable improvements in user engagement while streamlining the overall system and enabling effective modeling of 10k+ fine-grained user history events.

cs.IR↗

arXiv2026

Conflict-Based Lazy Search for Fast Multi-Manipulator Planning

Employing multiple manipulators can boost efficiency and accomplish tasks that a single manipulator cannot do. However, real-time planning for multiple manipulators in a cluttered workspace still poses significant challenges for planning algorithms. This article proposes a new planning algorithm called Conflict-Based Lazy Search (CBLS) for multimanipulator planning. CBLS is built on Conflict-Based Search (CBS), an efficient multiagent pathfinding (MAPF) algorithm that has shown an order of magnitude speedup over previous approaches [1], [2]. CBS addresses MAPF by solving many single-agent pathfinding (SAPF) problems. Thus, its planning time directly depends on the efficiency of the SAPF algorithm adopted. Our CBLS algorithm enhances CBS with precomputation and lazy search. First, a lazily evaluated graph with controlled sparsity is precomputed for a single manipulator. Second, we propose the Lazy Edged-based A* (LEA*) for efficient SAPF. Since edge evaluation is the computational bottleneck of manipulator planning, LEA* uses lazy search and an edge queue to reduce the number of edge evaluations. We show that LEA* is optimally vertex efficient and has improved edge efficiency compared to A*. We apply the proposed CBLS to multi-manipulator planning problems and show its superior performance by comparing it with CBS and a sampling-based algorithm, namely, RRT-Connect.

cs.RO↗

arXiv2026

Shigatse Astronomical Site Testing. I. Cloud-cover Climatology and Selected Local Meteorological Conditions

As the first paper in a Shigatse astronomical site-testing series, we present a multi-source assessment of cloud cover and selected local meteorological conditions at the Shigatse 40 m site on the southern Tibetan Plateau. The study combines CALIPSO-GOCCP active-lidar climatology, ISCCP HXG passive-satellite cloud fields, conventional total-cloud-amount observations from the Shigatse Meteorological Station, and on-site Weather Station measurements. Together, these records characterize Shigatse as a southern-plateau monsoon-transition cloud regime: the active-lidar climatology gives a moderate-to-low annual cloud fraction, and the cloudier months are concentrated in the June--September monsoon interval. In GOCCP, the annual mean cloud fraction is 42.1%, while the October--May low-cloud season has a mean cloud fraction of 26.3%, compared with 73.7% during the June--September monsoon interval. ISCCP gives higher absolute cloud fractions but supports the same seasonal phase and local spatial placement. The aligned 1988--2013 meteorological-station record gives a total-cloud-amount <=40% fraction of 80.7% during October--May, rising to 90.7% in the November--January core, and decreasing to 39.9% during June--September. The 2024--2025 Weather Station archive further shows high fractions of valid samples satisfying the adopted meteorological criteria during the low-cloud months: 92.6% for the October--May night-time proxy and 94.6% for the corresponding 24 h samples. These results identify Shigatse as a measured lower-latitude southern-plateau cloud-cover reference within China's site-testing network, with a well-defined October--May low-cloud observing period and a Shigatse--Ali low-cloud corridor for subsequent regional site testing.

astro-ph.IM↗

arXiv2026

Active-Passive Federated Learning for Vertically Partitioned Multi-view Data

Vertical federated learning is a natural and elegant approach to integrate multi-view data vertically partitioned across devices (clients) while preserving their privacies. Apart from the model training, existing methods requires the collaboration of all clients in the model inference. However, the model inference is probably maintained for service in a long time, while the collaboration, especially when the clients belong to different organizations, is unpredictable in real-world scenarios, such as concellation of contract, network unavailablity, etc., resulting in the failure of them. To address this issue, we, at the first attempt, propose a flexible Active-Passive Federated learning (APFed) framework. Specifically, the active client is the initiator of a learning task and responsible to build the complete model, while the passive clients only serve as assistants. Once the model built, the active client can make inference independently. In addition, we instance the APFed framework into two classification methods with employing the reconstruction loss and the contrastive loss on passive clients, respectively. Meanwhile, the two methods are tested in a set of experiments and achieves desired results, validating their effectiveness.

cs.LG↗

arXiv2026

VIABLE: A Visually Impaired Assistance Benchmark for VLM-as-a-Judge Evaluation

AI-based Visually Impaired Assistance (VIA) remains challenging, largely due to the high cost of human evaluation. The VLM-as-a-Judge paradigm may offer a promising alternative, although it has mostly been studied in general domains. We therefore ask whether such judges can be trusted for VIA tasks. To investigate this question, we introduce VIABLE (Visually Impaired Assistance Benchmark for VLM-as-a-Judge Evaluation), the first benchmark for VLM-as-a-Judge evaluation in VIA. VIABLE contains over 300K judgment samples across three scenarios and introduces an Effectiveness--Impartiality--Stability framework with a 12-mode failure taxonomy. Based on VIABLE, our systematic study of seven judges across different model scales shows that existing models are largely unreliable across all evaluation axes. The strongest judge, GPT-5.4, achieves only 52.6% single-failure diagnostic accuracy, yet exhibits the highest self-preference rate at 94.2%; while open-source judges are strongly biased and adversarially fragile. To address these issues, we propose VIA-Judge-Agent, a model-agnostic inference-time harness that augments judges with visual evidence extraction and a taxonomy-guided workflow. It enables positive improvements in diagnostic accuracy and downstream VIA responses more preferred by BLV users. Data and code are available at: https://github.com/YiyiyiZhao/VIABLE

cs.CL↗

arXiv2026

ERNIE-Image Technical Report

We introduce ERNIE-Image, an open-source text-to-image generation model built upon an 8B single-stream DiT architecture. ERNIE-Image aims to bridge the gap between current open-source models and leading closed-source systems through more effective mining of large-scale pre-training data and improved supervision quality throughout training. During pre-training, we adopt a bottom-up data construction pipeline that combines fine-grained image categorization, rich caption annotation, aesthetic assessment, and hierarchical sampling. This strategy reduces data noise while preserving long-tail concepts and detailed real-world knowledge, providing a stronger foundation for complex generation tasks. In the post-training stage, we use a top-down data construction pipeline for high-demand scenarios, diversify prompt annotations to better match real user inputs, and apply a stabilized DPO strategy to align the model with human aesthetic preferences. We further train ERNIE-Image-Turbo for efficient 8-NFE generation and propose MT-DMD to mitigate capability drift during distillation. To make the model easier to use in practical scenarios, we equip it with a lightweight Prompt Enhancer that expands concise user intents into structured visual descriptions. In addition, we develop ERNIE-Image-Aes, an industrial-grade aesthetic model, together with ERNIE-Image-Aes-1K, a human-annotated benchmark for realistic aesthetic evaluation. Extensive qualitative and quantitative experiments show that ERNIE-Image achieves leading performance among open-source models and approaches top-tier commercial models in instruction following, text rendering, and aesthetic quality. We release the trained models and aesthetic resources to facilitate further academic research and technical progress in the AIGC community.

cs.CV↗

arXiv2026

How Far Has AI Come in Liver Fibrosis Staging? A Large-Scale Real-World Dataset and Benchmark

Despite years of methodological progress, how far AI has come in liver fibrosis staging has never been systematically evaluated under the heterogeneous, multi-center conditions that define clinical practice. To address this gap, we introduce LiFS, a large-scale dataset and benchmark derived from the MICCAI 2025 CARE-Liver challenge, comprising 610 patients across multiple centers and scanners with multi-sequence MRI. To the best of our knowledge, LiFS is the first benchmark providing complete gadoxetic acid-enhanced sequences with histopathology-confirmed annotations from diverse real-world scanners. Through systematic evaluation of 9 independently developed methods selected from 96 registered teams against in-cohort radiologist reference results, our findings address how far current AI has progressed toward clinical-level liver fibrosis staging from three complementary perspectives. First, against radiologists, the best AI methods were broadly comparable to the senior radiologist and significantly exceeded the junior radiologist in selected settings, while median AI performance generally approached junior-radiologist levels. Second, from a data perspective, cross-center heterogeneity, label imbalance, and contrast-enhanced sequence variability emerge as the dominant challenges for AI methods. Third, from a technical perspective, methodological design choices, including spatial registration, input dimensionality, multi-modal fusion strategy, and backbone architecture, appear to modulate cross-center robustness, although no single choice alone closes the gap. Overall, LiFS provides a rigorous real-world benchmark for positioning the current state of AI in liver fibrosis staging and for enabling future research on the key challenges that limit clinically reliable deployment.

cs.CV↗

arXiv2026

RealRep: Generalized SDR-to-HDR Conversion via Attribute-Disentangled Representation Learning

High-Dynamic-Range Wide-Color-Gamut (HDR-WCG) technology is becoming increasingly widespread, driving a growing need for converting Standard Dynamic Range (SDR) content to HDR. Existing methods primarily rely on fixed tone mapping operators, which struggle to handle the diverse appearances and degradations commonly present in real-world SDR content. To address this limitation, we propose a generalized SDR-to-HDR framework that enhances robustness by learning attribute-disentangled representations. Central to our approach is Realistic Attribute-Disentangled Representation Learning (RealRep), which explicitly disentangles luminance and chrominance components to capture intrinsic content variations across different SDR distributions. Furthermore, we design a Luma-/Chroma-aware negative exemplar generation strategy that constructs degradation-sensitive contrastive pairs, effectively modeling tone discrepancies across SDR styles. Building on these attribute-level priors, we introduce the Degradation-Domain Aware Controlled Mapping Network (DDACMNet), a lightweight, two-stage framework that performs adaptive hierarchical mapping guided by a control-aware normalization mechanism. DDACMNet dynamically modulates the mapping process via degradation-conditioned features, enabling robust adaptation across diverse degradation domains. Extensive experiments demonstrate that RealRep consistently outperforms state-of-the-art methods in both generalization and perceptually faithful HDR color gamut reconstruction.

cs.CV↗

arXiv2026

Electronic origin of delicate antiferromagnetism in Fe$_{x}$NbS$_2$

Among the family of intercalated transition-metal dichalcogenides (TMDs), Fe$_{x}$NbS$_2$ is found to possess unique current-induced resistive switching behaviors, tunable antiferromagnetic states, and a commensurate charge order, all of which are tied to a critical Fe doping of $x_c$ = 1/3. However, the electronic origin of such extreme stoichiometry sensitivities remains unclear. Combining angle-resolved photoemission spectroscopy (ARPES) with density functional theory (DFT) calculations, we identify and characterize a dramatic eV-scale electronic restructuring that occurs across the $x_c$. Moment-carrying Fe 3$d_{z^2}$ electrons manifest as narrow bands within 200 meV of the Fermi level, distinct from other transition metal intercalated TMD magnets. These states strongly hybridize with itinerant electrons in TMD layer, rapidly lose coherence above $x_c$ due to correlation-driven effects. This sudden quasiparticle decoherence collapses the Fe-Nb hybridization, which explicitly suppresses the out-of-plane effective Fe-Fe exchange interaction, driving the transformation of the magnetic ground state from an antiferromagnetic stripe phase to a zigzag phase. These observations resemble the exceptional electronic and magnetic sensitivity of strongly correlated systems, and demonstrate that quantifying orbital-specific hybridization via ARPES offers an alternative pathway to evaluate effective magnetic exchange in metallic magnets, complementing inelastic neutron and resonant x-ray scattering probes.

cond-mat.str-el↗

arXiv2026

A Unified Conditional Flow for Motion Generation, Editing, and Intra-Structural Retargeting

Text-driven motion editing and intra-structural retargeting, where skeletons share topology but may differ in bone lengths and rest pose, are traditionally handled by fragmented pipelines with incompatible inputs and representations: editing relies on specialized generative steering, while retargeting is deferred to geometric post-processing. We present a unified conditional-flow framework that casts generation, semantic editing, and intra-structural retargeting as condition-modulated transport within one text- and skeleton-conditioned rectified-flow model. Under this formulation, editing changes the semantic condition while preserving skeletal structure, whereas retargeting changes the skeletal condition while preserving motion semantics. This makes FlowEdit-style transport a unified inference rule for motion manipulation rather than a task-specific editor. To instantiate this for articulated 3D motion, we develop a text- and skeleton-conditioned rectified-flow transformer. The model uses per-joint tokenization and explicit joint self-attention to capture spatial kinematic dependencies. We further inject text conditions at both joint and frame levels, while residual multi-condition classifier-free guidance balances text adherence and skeletal conformity. Experiments on SnapMoGen and a multi-character Mixamo subset show that one trained model supports text-to-motion generation, zero-shot editing, and zero-shot intra-structural retargeting without task-specific fine-tuning. This unified framework replaces separate pipelines with a single conditional motion transport model while keeping the same-topology retargeting scope explicit.

cs.GR↗

arXiv2026

A Generalist Model Including Evolved Star Mass and Age

Determining precise stellar ages and masses for evolved giants is crucial for Galactic archaeology but challenged by spectral degeneracies. Gaia's low-resolution XP spectra offer a unique opportunity to infer these parameters on a massive scale using data-driven methods. We extend a transformer-based astronomical foundation model to evolved stars, establishing a unified framework to simultaneously predict atmospheric parameters ($T_{\mathrm{eff}}$, $\log g$, $[\mathrm{M}/\mathrm{H}]$) and evolutionary labels (mass, age) with physical consistency. Treating spectra as token sequences, we integrated mass and age into the model's vocabulary. The model is trained on Gaia XP spectra cross-matched with the APOGEE DR17 DistMass catalog. Our generative approach enables flexible input handling, including spectral inpainting and parameter-to-spectrum generation. On an independent test set, the model achieves a prediction scatter of $σ\approx 0.114 \, M_{\odot}$ for mass and $σ\approx 1.334$ Gyr for age. Beyond numerical accuracy, it successfully reproduces the giant branch's mass-luminosity relation and autonomously disentangles interstellar extinction from intrinsic temperature variations without explicit physical priors. It also robustly recovers missing spectral data and estimates reliable uncertainties. Validating that foundation models can internalize stellar physics from data, this physically-aware, probabilistic framework offers a powerful tool for unraveling Milky Way history using large-scale spectroscopic surveys.

astro-ph.SR↗

arXiv2026

xLLM Technical Report

We introduce xLLM, an intelligent and efficient Large Language Model (LLM) inference framework designed for high-performance, large-scale enterprise-grade serving, with deep optimizations for diverse AI accelerators. To address these challenges, xLLM builds a novel decoupled service-engine architecture. At the service layer, xLLM-Service features an intelligent scheduling module that efficiently processes multimodal requests and co-locates online and offline tasks through unified elastic scheduling to maximize cluster utilization. This module also relies on a workload-adaptive dynamic Prefill-Decode (PD) disaggregation policy and a novel Encode-Prefill-Decode (EPD) disaggregation policy designed for multimodal inputs. Furthermore, it incorporates a distributed architecture to provide global KV Cache management and robust fault-tolerant capabilities for high availability. At the engine layer, xLLM-Engine co-optimizes system and algorithm designs to fully saturate computing resources. This is achieved through comprehensive multi-layer execution pipeline optimizations, an adaptive graph mode and an xTensor memory management. xLLM-Engine also further integrates algorithmic enhancements such as optimized speculative decoding and dynamic EPLB, collectively serving to substantially boost throughput and inference efficiency. Extensive evaluations demonstrate that xLLM delivers significantly superior performance and resource efficiency. Under identical TPOT constraints, xLLM achieves throughput up to 1.7x that of MindIE and 2.2x that of vLLM-Ascend with Qwen-series models, while maintaining an average throughput of 1.7x that of MindIE with Deepseek-series models. xLLM framework is publicly available at https://github.com/jd-opensource/xllm and https://github.com/jd-opensource/xllm-service.

cs.DC↗

arXiv2026

PhysFormer: A Physics-Embedded Generative Model for Physically Self-Consistent Spectral Synthesis

In scientific and engineering domains, modeling high-dimensional complex systems governed by partial differential equations (PDEs) remains challenging in terms of physical consistency and numerical stability. However, existing approaches, such as physics-informed neural networks (PINNs), typically rely on known physical fields or coefficients and enforce physical constraints via external loss functions, which can lead to training instability and make it difficult to handle high-dimensional or unobservable scenarios. To this end, we propose PhysFormer, a generative modeling framework that is self-consistent at both the data and physical levels. PhysFormer leverages a low-dimensional, physically interpretable latent space to learn key physical quantities directly from data without requiring known high-dimensional physical field parameters, and embeds the physical process of radiative flux generation within the network to ensure the physical consistency of the generated spectra. In high-dimensional, degenerate inversion tasks, PhysFormer constrains generation within physical limits and enhances spectral fidelity and inversion stability under varying signal-to-noise ratios (SNRs). More broadly, this approach shifts the physical processes from external loss functions into the generative mechanism itself, providing a physically consistent generative modeling paradigm for complex systems involving unknown or unobservable physical quantities.

astro-ph.IM↗