arXiv subjects

Yitong Wang

Publications and source records attributed to Yitong Wang.

At least 19 recordsLinked to original sources

arXiv2026

A Massively Parallel Three-Grid Preconditioner for the High-Frequency Helmholtz Equation

Accurate simulation of three-dimensional time-harmonic wave propagation over many wavelengths requires control of phase error and efficient solution of large indefinite systems. We develop a three-grid solver based on the compact 27-point interpolated optimized finite-difference (IOFD) discretization. Its wavenumber-dependent stencil supports a fine-grid resolution of six points per shortest wavelength and an unshifted physical correction on the $2h$ grid at only three points per shortest wavelength. The method retains unshifted IOFD operators on the $h$ and $2h$ grids, while a complex-shifted $2h$--$4h$ auxiliary cycle preconditions a factorization-free iterative approximation of the coarse inverse. Restricting the shift to this auxiliary cycle preserves the propagative character of the coarse correction. Comparison with the outgoing Green function confirms phase and relative-amplitude accuracy on a sequence of meshes up to $6144^3$, with the largest problem spanning approximately 1024 wavelengths per coordinate. The same fixed solver configuration retains robust convergence across smooth, discontinuous, high-contrast, and geophysical velocity models and exhibits scalable parallel performance. In particular, a problem spanning approximately 340 wavelengths in each coordinate direction is solved in 18.1 seconds on just 64 NVIDIA A100 GPUs.

math.NA↗

arXiv2026

Personalization Meets Safety:Mechanisms,Risks,and Mitigations in Personalized LLMs

Large Language Models (LLMs) have enabled increasingly personalized interactions by adapting to users' preferences, contexts, and long-term histories. However, the mechanisms that enable personalization also expand the safety landscape in ways not systematically addressed by existing literature. Existing reviews typically focus either on personalization or safety, leaving their intersection largely unexplored. We present the first comprehensive, safety-aware review of personalized LLMs. We organize personalization along three dimensions-user representation, personalization paradigm, and evaluation-and introduce a unified taxonomy of safety risks. At the representation level, we analyze risks arising from diverse user representations. Across mainstream personalization paradigms, we delineate vulnerabilities inherent to prompting, retrieval augmentation, parameter fine-tuning, reinforcement learning, Mixture-of-Experts (MoE), pruning, agent frameworks, and multimodal personalization, and synthesize mitigation strategies across the model lifecycle. Beyond these fine-grained risks, we characterize paradigm-agnostic safety risks arising from personalized adaptation. We further summarize personalized datasets and evaluation methodologies. Through a case study of OpenClaw, we analyze deployment trends in personalized agent ecosystems. Our analysis reveals three structural inadequacies in existing research: safety is evaluated as user-invariant rather than relational, personalization techniques are analyzed in isolation rather than in composition, and evaluation frameworks cannot capture emergent long-term risks. By jointly examining personalized representations, personalization paradigms, safety risks, defenses, and evaluation methods, we provide a unified framework for developing safe personalized LLMs and highlight key directions for future research.

cs.AI↗

arXiv2026

Lens: Rethinking Training Efficiency for Foundational Text-to-Image Models

We introduce Lens, a 3.8B-parameter T2I model that achieves performance competitive with, and in several cases surpassing, state-of-the-art models with more than 6B parameters across various benchmarks, while requiring significantly less training compute. For example, Lens requires only about 19.3% of the training compute used by Z-Image. The training efficiency of Lens stems from two key strategies beyond its compact model size. First, we maximize data information density per training batch by (i) training on Lens-800M, a dataset of 800M densely captioned image-text pairs whose captions are generated by GPT-4.1 and contain approximately 109 words on average, providing richer semantic supervision than conventional short captions, and (ii) constructing each batch from images with multiple resolutions and diverse aspect ratios, thereby enlarging the effective visual coverage of each optimization step. Second, we improve convergence speed through careful architectural choices, including adopting a semantic VAE that provides better latent representations and employing a strong language encoder that accelerates optimization while enabling multilingual generalization from English-only training data. After pre-training, we apply RL with taxonomy-driven prompts (Lens-RL-8K) and structured reward rubrics to suppress artifacts and improve visual quality, a reasoner module with training-free system prompt search to better align user requests with the model, and distillation-based acceleration for 4-step inference. Through efficient training and systematic optimization, Lens generalizes to arbitrary aspect ratios from 1:2 to 2:1 and resolutions up to 1440^2, and supports prompts in several commonly used languages. Thanks to its compact size, Lens generates a 1024^2 image in 3.15 seconds on a single NVIDIA H100 GPU, while its distilled turbo version performs 4-step generation in 0.84 seconds.

cs.CV↗

arXiv2026

DreamSR: Towards Ultra-High-Resolution Image Super-Resolution via a Receptive-Field Enhanced Diffusion Transformer

Large-scale pre-trained diffusion models have been extensively adopted for real-world image Super-Resolution because of their powerful generative priors through textual guidance. However, when super-resolving high-resolution images with patch-wise inference strategy, most existing diffusion-based SR methods tend to suffer from over-generation, due to the misalignment between the global prompt from LR image and the incomplete semantic information of local patches during each inference step. On the other hand, most existing methods also failed to generate detailed texture in local patches due to the overemphasis on global generation capabilities in network designs and training strategies. To address this issue, we present DreamSR, a novel SR model that suppresses local over-generation and improves fine-detail synthesis, thereby achieving visually faithful results with ultra-high-quality details. Specifically, we propose a dual-branch MM-ControlNet, where the ControlNet generates local textual feature with patch-level prompts while the pre-trained DiT provides global textual feature with global prompts, thereby mitigating over-generation and ensuring semantic consistency across patches. We also design a comprehensive training strategy with stage-specific data processing pipelines and a Receptive-Field Enhancement strategy, enhancing the model's capability to capture patch information and effectively restore local textures. Extensive experiments demonstrate that DreamSR outperforms state-of-the-art methods, providing high-quality SR results. Code and model are available at https://github.com/jerrydong0219/DreamSR.

cs.CV↗

arXiv2026

DreamShot: Personalized Storyboard Synthesis with Video Diffusion Prior

Storyboard synthesis plays a crucial role in visual storytelling, aiming to generate coherent shot sequences that visually narrate cinematic events with consistent characters, scenes, and transitions. However, existing approaches are mostly adapted from text-to-image diffusion models, which struggle to maintain long-range temporal coherence, consistent character identities, and narrative flow across multiple shots. In this paper, we introduce DreamShot, a video generative model based storyboard framework that fully exploits powerful video diffusion priors for controllable multi-shot synthesis. DreamShot supports both Text-to-Shot and Reference-to-Shot generation, as well as story continuation conditioned on previous frames, enabling flexible and context-aware storyboard generation. By leveraging the spatial-temporal consistency inherent in video generative models, DreamShot produces visually and semantically coherent sequences with improved narrative fidelity and character continuity. Furthermore, DreamShot incorporates a multi-reference role conditioning module that accepts multiple character reference images and enforces identity alignment via a Role-Attention Consistency Loss, explicitly constraining attention between reference and generated roles. Extensive experiments demonstrate that DreamShot achieves superior scene coherence, role consistency, and generation efficiency compared to state-of-the-art text-to-image storyboard models, establishing a new direction toward controllable video model-driven visual storytelling.

cs.CV↗

arXiv2026

On the $K$-extensions between Serre weights for unramified $\mathrm{GL}_3$

Let $p\geq5$ be a prime number. Let $L$ be a finite unramified extension of $\mathbb{Q}_p$ with ring of integers $\mathcal{O}_L$ and residue field $\mathbb{F}_q$. Given two Serre weights for $\mathrm{GL}_3(\mathbb{F}_q)$, we prove that in most cases the extensions between them for $\mathrm{GL}_3(\mathcal{O}_L)$ modulo the center coincide with their $\mathrm{GL}_3(\mathbb{F}_q)$-extensions.

math.NT↗

arXiv2026

VisualAD: Language-Free Zero-Shot Anomaly Detection via Vision Transformer

Zero-shot anomaly detection (ZSAD) requires detecting and localizing anomalies without access to target-class anomaly samples. Mainstream methods rely on vision-language models (VLMs) such as CLIP: they build hand-crafted or learned prompt sets for normal and abnormal semantics, then compute image-text similarities for open-set discrimination. While effective, this paradigm depends on a text encoder and cross-modal alignment, which can lead to training instability and parameter redundancy. This work revisits the necessity of the text branch in ZSAD and presents VisualAD, a purely visual framework built on Vision Transformers. We introduce two learnable tokens within a frozen backbone to directly encode normality and abnormality. Through multi-layer self-attention, these tokens interact with patch tokens, gradually acquiring high-level notions of normality and anomaly while guiding patches to highlight anomaly-related cues. Additionally, we incorporate a Spatial-Aware Cross-Attention (SCA) module and a lightweight Self-Alignment Function (SAF): SCA injects fine-grained spatial information into the tokens, and SAF recalibrates patch features before anomaly scoring. VisualAD achieves state-of-the-art performance on 13 zero-shot anomaly detection benchmarks spanning industrial and medical domains, and adapts seamlessly to pretrained vision backbones such as the CLIP image encoder and DINOv2. Code: https://github.com/7HHHHH/VisualAD

cs.CV↗

arXiv2026

ReTracing: An Archaeological Approach Through Body, Machine, and Generative Systems

We present ReTracing, a multi-agent embodied performance art that adopts an archaeological approach to examine how artificial intelligence shapes, constrains, and produces bodily movement. Drawing from science-fiction novels, the project extracts sentences that describe human-machine interaction. We use large language models (LLMs) to generate paired prompts "what to do" and "what not to do" for each excerpt. A diffusion-based text-to-video model transforms these prompts into choreographic guides for a human performer and motor commands for a quadruped robot. Both agents enact the actions on a mirrored floor, captured by multi-camera motion tracking and reconstructed into 3D point clouds and motion trails, forming a digital archive of motion traces. Through this process, ReTracing serves as a novel approach to reveal how generative systems encode socio-cultural biases through choreographed movements. Through an immersive interplay of AI, human, and robot, ReTracing confronts a critical question of our time: What does it mean to be human among AIs that also move, think, and leave traces behind?

cs.CV↗

arXiv2026

Generative AI in Action: Field Experimental Evidence from Alibaba's Customer Service Operations

In collaboration with Alibaba, we study how a generative AI assistant affects service performance in e-commerce after-sales operations. In a large-scale field experiment, human agents providing digital chat support were randomly assigned access to a gen AI assistant. The assistant drafts issue diagnoses and solution proposals in the opening stage only; agents can adopt, modify, or disregard them. Because of this discretion, we estimate the effects of both gen AI access and usage. On average, gen AI improves service speed and subjective service quality, measured by customer ratings, but has no significant effect on objective service quality, measured by customer retrials. These gains come from more than automation. Gen AI reshapes agent-customer interactions: treated agents respond faster and take a more proactive role, while customers provide less input; both patterns persist into later chat stages. These average effects, however, mask heterogeneity across agents. Lower-performing agents benefit the most, indicating that gen AI can narrow performance gaps. Top-performing agents experience declines in both subjective and objective service quality. This decline is consistent with workflow disruption: among top performers, gen AI use increases shift-away time, response delays, and immediate customer retrials, suggesting weakened service continuity in the focal chat. Successful gen AI deployment therefore requires careful performance evaluation and rollout tailored to agent skill.

cs.HC↗

arXiv2026

NextFlow: Unified Sequential Modeling Activates Multimodal Understanding and Generation

We present NextFlow, a unified decoder-only autoregressive transformer trained on 6 trillion interleaved text-image discrete tokens. By leveraging a unified vision representation within a unified autoregressive architecture, NextFlow natively activates multimodal understanding and generation capabilities, unlocking abilities of image editing, interleaved content and video generation. Motivated by the distinct nature of modalities - where text is strictly sequential and images are inherently hierarchical - we retain next-token prediction for text but adopt next-scale prediction for visual generation. This departs from traditional raster-scan methods, enabling the generation of 1024x1024 images in just 5 seconds - orders of magnitude faster than comparable AR models. We address the instabilities of multi-scale generation through a robust training recipe. Furthermore, we introduce a prefix-tuning strategy for reinforcement learning. Experiments demonstrate that NextFlow achieves state-of-the-art performance among unified models and rivals specialized diffusion baselines in visual quality.

cs.CV↗

arXiv2025

DreamOmni3: Scribble-based Editing and Generation

Recently unified generation and editing models have achieved remarkable success with their impressive performance. These models rely mainly on text prompts for instruction-based editing and generation, but language often fails to capture users intended edit locations and fine-grained visual details. To this end, we propose two tasks: scribble-based editing and generation, that enables more flexible creation on graphical user interface (GUI) combining user textual, images, and freehand sketches. We introduce DreamOmni3, tackling two challenges: data creation and framework design. Our data synthesis pipeline includes two parts: scribble-based editing and generation. For scribble-based editing, we define four tasks: scribble and instruction-based editing, scribble and multimodal instruction-based editing, image fusion, and doodle editing. Based on DreamOmni2 dataset, we extract editable regions and overlay hand-drawn boxes, circles, doodles or cropped image to construct training data. For scribble-based generation, we define three tasks: scribble and instruction-based generation, scribble and multimodal instruction-based generation, and doodle generation, following similar data creation pipelines. For the framework, instead of using binary masks, which struggle with complex edits involving multiple scribbles, images, and instructions, we propose a joint input scheme that feeds both the original and scribbled source images into the model, using different colors to distinguish regions and simplify processing. By applying the same index and position encodings to both images, the model can precisely localize scribbled regions while maintaining accurate editing. Finally, we establish comprehensive benchmarks for these tasks to promote further research. Experimental results demonstrate that DreamOmni3 achieves outstanding performance, and models and code will be publicly released.

cs.CV↗

arXiv2025

MMSRARec: Summarization and Retrieval Augumented Sequential Recommendation Based on Multimodal Large Language Model

Recent advancements in Multimodal Large Language Models (MLLMs) have demonstrated significant potential in recommendation systems. However, the effective application of MLLMs to multimodal sequential recommendation remains unexplored: A) Existing methods primarily leverage the multimodal semantic understanding capabilities of pre-trained MLLMs to generate item embeddings or semantic IDs, thereby enhancing traditional recommendation models. These approaches generate item representations that exhibit limited interpretability, and pose challenges when transferring to language model-based recommendation systems. B) Other approaches convert user behavior sequence into image-text pairs and perform recommendation through multiple MLLM inference, incurring prohibitive computational and time costs. C) Current MLLM-based recommendation systems generally neglect the integration of collaborative signals. To address these limitations while balancing recommendation performance, interpretability, and computational cost, this paper proposes MultiModal Summarization-and-Retrieval-Augmented Sequential Recommendation. Specifically, we first employ MLLM to summarize items into concise keywords and fine-tune the model using rewards that incorporate summary length, information loss, and reconstruction difficulty, thereby enabling adaptive adjustment of the summarization policy. Inspired by retrieval-augmented generation, we then transform collaborative signals into corresponding keywords and integrate them as supplementary context. Finally, we apply supervised fine-tuning with multi-task learning to align the MLLM with the multimodal sequential recommendation. Extensive evaluations on common recommendation datasets demonstrate the effectiveness of MMSRARec, showcasing its capability to efficiently and interpretably understand user behavior histories and item information for accurate recommendations.

cs.IR↗

arXiv2025

CustomX: Unified Character, Action, and Scene Customization in Video World Models

Recent advances in world models have greatly enhanced interactive environment simulation. Existing methods mainly fall into two categories: (1) static world generation models, which construct 3D environments without active agents, and (2) controllable-entity models, which allow a single entity to perform limited actions in an otherwise uncontrollable environment. In this work, we introduce CustomX, leveraging the realism and structural grounding of static world generation while extending controllable-entity models to support user-specified characters capable of performing open-ended actions. Users can provide a 3DGS scene and a character, then use natural language to direct the character to perform diverse behaviors, ranging from basic locomotion to object-centric interactions, while freely exploring the environment. CustomX synthesizes temporally coherent video clips that preserve visual fidelity with the provided scene and character, formulated as a conditional autoregressive video generation problem. Built upon a pre-trained video generator, our training strategy significantly enhances motion dynamics while maintaining generalization across actions and characters. Our evaluation covers a broad range of aspects, including visual quality, character consistency, action controllability, and long-horizon coherence.

cs.CV↗

arXiv2025

Why Did Apple Fall: Evaluating Curiosity in Large Language Models

Curiosity serves as a pivotal conduit for human beings to discover and learn new knowledge. Recent advancements of large language models (LLMs) in natural language processing have sparked discussions regarding whether these models possess capability of curiosity-driven learning akin to humans. In this paper, starting from the human curiosity assessment questionnaire Five-Dimensional Curiosity scale Revised (5DCR), we design a comprehensive evaluation framework that covers dimensions such as Information Seeking, Thrill Seeking, and Social Curiosity to assess the extent of curiosity exhibited by LLMs. The results demonstrate that LLMs exhibit a stronger thirst for knowledge than humans but still tend to make conservative choices when faced with uncertain environments. We further investigated the relationship between curiosity and thinking of LLMs, confirming that curious behaviors can enhance the model's reasoning and active learning abilities. These findings suggest that LLMs have the potential to exhibit curiosity similar to that of humans, providing experimental support for the future development of learning capabilities and innovative research in LLMs.

cs.CL↗

arXiv2025

DreamOmni2: Multimodal Instruction-based Editing and Generation

Recent advancements in instruction-based image editing and subject-driven generation have garnered significant attention, yet both tasks still face limitations in meeting practical user needs. Instruction-based editing relies solely on language instructions, which often fail to capture specific editing details, making reference images necessary. Meanwhile, subject-driven generation is limited to combining concrete objects or people, overlooking broader, abstract concepts. To address these challenges, we propose two novel tasks: multimodal instruction-based editing and generation. These tasks support both text and image instructions and extend the scope to include both concrete and abstract concepts, greatly enhancing their practical applications. We introduce DreamOmni2, tackling two primary challenges: data creation and model framework design. Our data synthesis pipeline consists of three steps: (1) using a feature mixing method to create extraction data for both abstract and concrete concepts, (2) generating multimodal instruction-based editing training data using the editing and extraction models, and (3) further applying the extraction model to create training data for multimodal instruction-based editing. For the framework, to handle multi-image input, we propose an index encoding and position encoding shift scheme, which helps the model distinguish images and avoid pixel confusion. Additionally, we introduce joint training with the VLM and our generation/editing model to better process complex instructions. In addition, we have proposed comprehensive benchmarks for these two new tasks to drive their development. Experiments show that DreamOmni2 has achieved impressive results. Models and codes will be released.

cs.CV↗

arXiv2025

Quantum Optics Nature of the Elementary Excitations in Few-Layer WSe2 Semiconductors

A fully quantized description of a two-level system resonantly coupled with an electromagnetic field (light) is among the central topics of quantum electrodynamics, which is theorized by the quantum Rabi model. It is also a fundamental issue of light-matter interactions. The rapid development of two-dimensional (2D) transition-metal dichalcogenide (TMDC) atomically-thin semiconductors brings excellent great chance to test and demonstrate some basic predictions by quantum optics theory in textbook, i.e., by the quantum Rabi model. Such test and demonstration are of both scientific and technological significance because of the quick emergence of the second quantum revolution. In this Letter, we show a quantum optics demonstration of the variable-temperature optical responses of the elementary excitations in few-layer WSe2 flakes. It is unraveled that the variable-temperature reflectance and fluorescence patterns of the elementary excitations (i.e., the band-edge excitons) of monolayer, bilayer and hBN capped WSe2 match well with the predictions by the quantum Rabi model under the rotating wave approximation. Decoherence times, Rabi frequencies, and transition matrix elements of the elementary excitations in these few-layer WSe2 flakes are found to be all negatively correlated with temperature, and to show dependence on layer number and capping layer. These findings may provide a novel perspective for comprehending the fundamental quantum physical properties of two-dimensional materials.

physics.optics↗

arXiv2025

Real, Fake, or Manipulated? Detecting Machine-Influenced Text

Large Language Model (LLMs) can be used to write or modify documents, presenting a challenge for understanding the intent behind their use. For example, benign uses may involve using LLM on a human-written document to improve its grammar or to translate it into another language. However, a document entirely produced by a LLM may be more likely to be used to spread misinformation than simple translation (\eg, from use by malicious actors or simply by hallucinating). Prior works in Machine Generated Text (MGT) detection mostly focus on simply identifying whether a document was human or machine written, ignoring these fine-grained uses. In this paper, we introduce a HiErarchical, length-RObust machine-influenced text detector (HERO), which learns to separate text samples of varying lengths from four primary types: human-written, machine-generated, machine-polished, and machine-translated. HERO accomplishes this by combining predictions from length-specialist models that have been trained with Subcategory Guidance. Specifically, for categories that are easily confused (\eg, different source languages), our Subcategory Guidance module encourages separation of the fine-grained categories, boosting performance. Extensive experiments across five LLMs and six domains demonstrate the benefits of our HERO, outperforming the state-of-the-art by 2.5-3 mAP on average.

cs.CL↗

arXiv2025

Odd-Parity Selection in Parity-Forbidden Electronic Transitions Revealed by Mn4+ Fluorescence Spectroscopy

Mn4+-doped fluoride phosphors represent a significant class of narrow band red-emitting materials, whose luminescent properties are profoundly influenced by electron-phonon coupling. However, the parity-forbidden nature of these electronic transition systems is incompatible with the conventional Condon approximation, which is widely adopted in the classic theories such as the Huang-Rhys theory, a framework established on the assumption of parity-allowed electric dipole transitions. This results in a critical knowledge gap regarding the principles governing the phonon sidebands of parity-forbidden electronic transitions. This study experimentally reveals a pronounced parity-dependent intensity distribution in the phonon sidebands of these systems: significantly suppressed even-order sidebands and normally observed odd-order sidebands. To elucidate the phenomenon, we extend the Huang-Rhys theory to parity-forbidden systems by incorporating the Herzberg-Teller approximation into the treatment of the transition matrix elements. The improved theory successfully uncovers the physical mechanism behind the strong suppression of the even-order sidebands in the parity-forbidden systems, in which the Huang-Rhys factor is derived as S=((2I_3)/(9I_1 ))^(1/2). This work not only reveals new findings regarding the phonon sidebands of the parity-forbidden electronic transition systems, but also establishes an improved theoretical framework for understanding the electron-phonon coupling mechanisms of color centers in solids.

cond-mat.mtrl-sci↗