Searcharxiv⌕ Search

arXiv subjects

Yuxin Zhang

Publications and source records attributed to Yuxin Zhang.

At least 73 records · Page 4Linked to original sources
arXiv2026

TAG-MoE: Task-Aware Gating for Unified Generative Mixture-of-Experts

Unified image generation and editing models suffer from severe task interference in dense diffusion transformers architectures, where a shared parameter space must compromise between conflicting objectives (e.g., local editing v.s. subject-driven generation). While the sparse Mixture-of-Experts (MoE) paradigm is a promising solution, its gating networks remain task-agnostic, operating based on local features, unaware of global task intent. This task-agnostic nature prevents meaningful specialization and fails to resolve the underlying task interference. In this paper, we propose a novel framework to inject semantic intent into MoE routing. We introduce a Hierarchical Task Semantic Annotation scheme to create structured task descriptors (e.g., scope, type, preservation). We then design Predictive Alignment Regularization to align internal routing decisions with the task's high-level semantics. This regularization evolves the gating network from a task-agnostic executor to a dispatch center. Our model effectively mitigates task interference, outperforming dense baselines in fidelity and quality, and our analysis shows that experts naturally develop clear and semantically correlated specializations.

cs.CV↗
arXiv2026

GPT4o-Receipt: A Dataset and Human Study for AI-Generated Document Forensics

Can humans detect AI-generated financial documents better than machines? We present GPT4o-Receipt, a benchmark of 1,235 receipt images pairing GPT-4o-generated receipts with authentic ones from established datasets, evaluated by five state-of-the-art multimodal LLMs and a 30-annotator crowdsourced perceptual study. Our findings reveal a striking paradox: humans are better at seeing AI artifacts, yet worse at detecting AI documents. Human annotators exhibit the largest visual discrimination gap of any evaluator, yet their binary detection F1 falls well below Claude Sonnet 4 and below Gemini 2.5 Flash. This paradox resolves once the mechanism is understood: the dominant forensic signals in AI-generated receipts are arithmetic errors -- invisible to visual inspection but systematically verifiable by LLMs. Humans cannot perceive that a subtotal is incorrect; LLMs verify it in milliseconds. Beyond the human--LLM comparison, our five-model evaluation reveals dramatic performance disparities and calibration differences that render simple accuracy metrics insufficient for detector selection. GPT4o-Receipt, the evaluation framework, and all results are released publicly to support future research in AI document forensics.

cs.AI↗
arXiv2026

From Sustainable Materials to User-Centered Sustainability: Material Experience in Art Healing

This study develops sustainable materials using hydrogel as the matrix and explores the transition from sustainable materials to user-centered sustainability, with a particular focus on achieving art healing through material experience. The findings reveal that "Aesthetic" property exert the greatest influence on art healing in the context of multimodal material experiences involving visual, tactile, and smell, followed by "Intrinsic" property, whereas "Physical" property have a comparatively limited effect. Furthermore, the study proposes a material experience framework that enables designers to systematically and holistically understanding material characteristics. It highlights the importance of considering users' psychological perceptions and emotional needs in the material design process.

cs.HC↗
arXiv2026

Out of the Memory Barrier: A Highly Memory Efficient Training System for LLMs with Million-Token Contexts

Training Large Language Models (LLMs) on long contexts is severely constrained by prohibitive GPU memory overhead, not training time. The primary culprits are the activations, whose memory footprints scale linearly with sequence length. We introduce OOMB, a highly memory-efficient training system that directly confronts this barrier. Our approach employs a chunk-recurrent training framework with on-the-fly activation recomputation, which maintains a constant activation memory footprint (O(1)) and shifts the primary bottleneck to the growing KV cache. To manage the KV cache, OOMB integrates a suite of synergistic optimizations: a paged memory manager for both the KV cache and its gradients to eliminate fragmentation, asynchronous CPU offloading to hide data transfer latency, and page-level sparse attention to reduce both computational complexity and communication overhead. The synergy of these techniques yields exceptional efficiency. Our empirical results show that for every additional 10K tokens of context, the end-to-end training memory overhead increases by a mere 10MB for Qwen2.5-7B. This allows training Qwen2.5-7B with a 4M-token context on a single H200 GPU, a feat that would otherwise require a large cluster using context parallelism. This work represents a substantial advance in resource efficiency for long-context LLM training. The source code is available at https://github.com/wenhaoli-xmu/OOMB.

cs.CL↗
arXiv2026

SongEcho: Towards Cover Song Generation via Instance-Adaptive Element-wise Linear Modulation

Cover songs constitute a vital aspect of musical culture, preserving the core melody of an original composition while reinterpreting it to infuse novel emotional depth and thematic emphasis. Although prior research has explored the reinterpretation of instrumental music through melody-conditioned text-to-music models, the task of cover song generation remains largely unaddressed. In this work, we reformulate our cover song generation as a conditional generation, which simultaneously generates new vocals and accompaniment conditioned on the original vocal melody and text prompts. To this end, we present SongEcho, which leverages Instance-Adaptive Element-wise Linear Modulation (IA-EiLM), a framework that incorporates controllable generation by improving both conditioning injection mechanism and conditional representation. To enhance the conditioning injection mechanism, we extend Feature-wise Linear Modulation (FiLM) to an Element-wise Linear Modulation (EiLM), to facilitate precise temporal alignment in melody control. For conditional representations, we propose Instance-Adaptive Condition Refinement (IACR), which refines conditioning features by interacting with the hidden states of the generative model, yielding instance-adaptive conditioning. Additionally, to address the scarcity of large-scale, open-source full-song datasets, we construct Suno70k, a high-quality AI song dataset enriched with comprehensive annotations. Experimental results across multiple datasets demonstrate that our approach generates superior cover songs compared to existing methods, while requiring fewer than 30% of the trainable parameters. The code, dataset, and demos are available at https://github.com/lsfhuihuiff/SongEcho_ICLR2026.

cs.SD↗
arXiv2026

LAURA: Enhancing Code Review Generation with Context-Enriched Retrieval-Augmented LLM

Code review is critical for ensuring software quality and maintainability. With the rapid growth in software scale and complexity, code review has become a bottleneck in the development process because of its time-consuming and knowledge-intensive nature and the shortage of experienced developers willing to review code. Several approaches have been proposed for automatically generating code reviews based on retrieval, neural machine translation, pre-trained models, or large language models (LLMs). These approaches mainly leverage historical code changes and review comments. However, a large amount of crucial information for code review, such as the context of code changes and prior review knowledge, has been overlooked. This paper proposes an LLM-based review knowledge-augmented, context-aware framework for code review generation, named LAURA. The framework integrates review exemplar retrieval, context augmentation, and systematic guidance to enhance the performance of ChatGPT-4o and DeepSeek v3 in generating code review comments. Besides, given the extensive low-quality reviews in existing datasets, we also constructed a high-quality dataset. Experimental results show that for both models, LAURA generates review comments that are either completely correct or at least helpful to developers in 42.2% and 40.4% of cases, respectively, significantly outperforming SOTA baselines. Furthermore, our ablation studies demonstrate that all components of LAURA contribute positively to improving comment quality.

cs.SE↗
arXiv2026

Out of the box age estimation through facial imagery: A Comprehensive Benchmark of Vision-Language Models vs. out-of-the-box Traditional Architectures

Facial age estimation plays a critical role in content moderation, age verification, and deepfake detection. However, no prior benchmark has systematically compared modern vision-language models (VLMs) with specialized age estimation architectures. We present the first large-scale cross-paradigm benchmark, evaluating 34 models - 22 specialized architectures with publicly available pretrained weights and 12 general-purpose VLMs - across eight standard datasets (UTKFace, IMDB-WIKI, MORPH, AFAD, CACD, FG-NET, APPA-REAL, and AgeDB), totaling 1,100 test images per model. Our key finding is striking: zero-shot VLMs significantly outperform most specialized models, achieving an average mean absolute error (MAE) of 5.65 years compared to 9.88 years for non-LLM models. The best-performing VLM (Gemini 3 Flash Preview, MAE 4.32) surpasses the strongest non-LLM model (MiVOLO, MAE 5.10) by 15%. MiVOLO - unique in combining face and body features using Vision Transformers - is the only specialized model that remains competitive with VLMs. We further analyze age verification at the 18-year threshold and find that most non-LLM models exhibit false adult rates between 39% and 100% for minors, whereas VLMs reduce this to 16%-29%. Additionally, coarse age binning (8-9 classes) consistently increases MAE beyond 13 years. Stratified analysis across 14 age groups reveals that all models struggle most at extreme ages (under 5 and over 65). Overall, these findings challenge the assumption that task-specific architectures are necessary for high-performance age estimation and suggest that future work should focus on distilling VLM capabilities into efficient specialized models.

cs.CV↗
arXiv2026

The Effect of Design Thinking on Creative & Innovation Processes: An Empirical Study Across Different Design Experience Levels

This study employs linear regression and structural equation modeling to explore how Thinking Skills, Design Thinking, Creative Self-Efficacy (CSE), and Collective Creative Efficacy (CCE) drive Design Creativity & Innovation, and analyzes the structural stability of the model across different levels of experience. Path analysis results indicate that the four Design Thinking Skills, Problem-driven Design (beta = 0.198, p < 0.01), Information-driven Design (beta = 0.241, p < 0.001), Solution-driven Design (beta = 0.227, p < 0.001), and Knowledge-driven Design (beta = 0.263, p < 0.001) all significantly and positively influence Design Thinking. Furthermore, Design Thinking has a significant positive predictive effect on Design Creativity & Innovation (beta = 0.286, p < 0.001). Mediation analysis confirms three significant mediation paths: the CSE mediation path (beta = 0.128, p < 0.001), the CCE mediation path (beta = 0.073, p < 0.01), and the "CSE to CCE" chain mediation path (beta = 0.025, p < 0.01). Multi-group comparison results reveal significant differences between the student and professional groups under the full equivalence model. After relaxing specific constraints, there were no significant differences between the nested models of the baseline model, partial measurement invariance, structural weight invariance, and structural covariance invariance. These findings elucidate the multi-dimensional pathways of Design Creativity & Innovation, providing a robust empirical basis for optimizing differentiated pedagogical models and professional practice guidelines.

cs.HC↗
arXiv2026

Test-Time Iterative Error Correction for Efficient Diffusion Models

With the growing demand for high-quality image generation on resource-constrained devices, efficient diffusion models have received increasing attention. However, such models suffer from approximation errors introduced by efficiency techniques, which significantly degrade generation quality. Once deployed, these errors are difficult to correct, as modifying the model is typically infeasible in deployment environments. Through an analysis of error propagation across diffusion timesteps, we reveal that these approximation errors can accumulate exponentially, severely impairing output quality. Motivated by this insight, we propose Iterative Error Correction (IEC), a novel test-time method that mitigates inference-time errors by iteratively refining the model's output. IEC is theoretically proven to reduce error propagation from exponential to linear growth, without requiring any retraining or architectural changes. IEC can seamlessly integrate into the inference process of existing diffusion models, enabling a flexible trade-off between performance and efficiency. Extensive experiments show that IEC consistently improves generation quality across various datasets, efficiency techniques, and model architectures, establishing it as a practical and generalizable solution for test-time enhancement of efficient diffusion models. The code is available in https://github.com/zysxmu/IEC.

cs.LG↗
arXiv2026

Benchmarking Autonomous Vehicles: A Driver Foundation Model Framework

Autonomous vehicles (AVs) are poised to revolutionize global transportation systems. However, its widespread acceptance and market penetration remain significantly below expectations. This gap is primarily driven by persistent challenges in safety, comfort, commuting efficiency and energy economy when compared to the performance of experienced human drivers. We hypothesize that these challenges can be addressed through the development of a driver foundation model (DFM). Accordingly, we propose a framework for establishing DFMs to comprehensively benchmark AVs. Specifically, we describe a large-scale dataset collection strategy for training a DFM, discuss the core functionalities such a model should possess, and explore potential technical solutions to realize these functionalities. We further present the utility of the DFM across the operational spectrum, from defining human-centric safety envelopes to establishing benchmarks for energy economy. Overall, We aim to formalize the DFM concept and introduce a new paradigm for the systematic specification, verification and validation of AVs.

cs.RO↗
arXiv2026

DeepPrep: An LLM-Powered Agentic System for Autonomous Data Preparation

Data preparation, which aims to transform heterogeneous and noisy raw tables into analysis-ready data, remains a major bottleneck in data science. Recent approaches leverage large language models (LLMs) to automate data preparation from natural language specifications. However, existing LLM-powered methods either make decisions without grounding in intermediate execution results, or rely on linear interaction processes that offer limited support for revising earlier decisions. To address these limitations, we propose DeepPrep, an LLM-powered agentic system for autonomous data preparation. DeepPrep constructs data preparation pipelines through iterative, execution-grounded interaction with an environment that materializes intermediate table states and returns runtime feedback. To overcome the limitations of linear interaction, DeepPrep organizes pipeline construction with tree-based agentic reasoning, enabling structured exploration and non-local revision based on execution feedback. To enable effective learning of such behaviors, we propose a progressive agentic training framework, together with data synthesis that supplies diverse and complex ADP tasks. Extensive experiments show that DeepPrep achieves data preparation accuracy comparable to strong closed-source models (e.g., GPT-5) while incurring 15x lower inference cost, while establishing state-of-the-art performance among open-source baselines and generalizing effectively across diverse datasets.

cs.DB↗
arXiv2026

SIDeR: Semantic Identity Decoupling for Unrestricted Face Privacy

With the deep integration of facial recognition into online banking, identity verification, and other networked services, achieving effective decoupling of identity information from visual representations during image storage and transmission has become a critical challenge for privacy protection. To address this issue, we propose SIDeR, a Semantic decoupling-driven framework for unrestricted face privacy protection. SIDeR decomposes a facial image into a machine-recognizable identity feature vector and a visually perceptible semantic appearance component. By leveraging semantic-guided recomposition in the latent space of a diffusion model, it generates visually anonymous adversarial faces while maintaining machine-level identity consistency. The framework incorporates momentum-driven unrestricted perturbation optimization and a semantic-visual balancing factor to synthesize multiple visually diverse, highly natural adversarial samples. Furthermore, for authorized access, the protected image can be restored to its original form when the correct password is provided. Extensive experiments on the CelebA-HQ and FFHQ datasets demonstrate that SIDeR achieves a 99% attack success rate in black-box scenarios and outperforms baseline methods by 41.28% in PSNR-based restoration quality.

cs.CV↗
arXiv2026

KTV: Keyframes and Key Tokens Selection for Efficient Training-Free Video LLMs

Training-free video understanding leverages the strong image comprehension capabilities of pre-trained vision language models (VLMs) by treating a video as a sequence of static frames, thus obviating the need for costly video-specific training. However, this paradigm often suffers from severe visual redundancy and high computational overhead, especially when processing long videos. Crucially, existing keyframe selection strategies, especially those based on CLIP similarity, are prone to biases and may inadvertently overlook critical frames, resulting in suboptimal video comprehension. To address these significant challenges, we propose \textbf{KTV}, a novel two-stage framework for efficient and effective training-free video understanding. In the first stage, KTV performs question-agnostic keyframe selection by clustering frame-level visual features, yielding a compact, diverse, and representative subset of frames that mitigates temporal redundancy. In the second stage, KTV applies key visual token selection, pruning redundant or less informative tokens from each selected keyframe based on token importance and redundancy, which significantly reduces the number of tokens fed into the LLM. Extensive experiments on the Multiple-Choice VideoQA task demonstrate that KTV outperforms state-of-the-art training-free baselines while using significantly fewer visual tokens, \emph{e.g.}, only 504 visual tokens for a 60-min video with 10800 frames, achieving $44.8\%$ accuracy on the MLVU-Test benchmark. In particular, KTV also exceeds several training-based approaches on certain benchmarks.

cs.CV↗
arXiv2026

FSVideo: Fast Speed Video Diffusion Model in a Highly-Compressed Latent Space

We introduce FSVideo, a fast speed transformer-based image-to-video (I2V) diffusion framework. We build our framework on the following key components: 1.) a new video autoencoder with highly-compressed latent space ($64\times64\times4$ spatial-temporal downsampling ratio), achieving competitive reconstruction quality; 2.) a diffusion transformer (DIT) architecture with a new layer memory design to enhance inter-layer information flow and context reuse within DIT, and 3.) a multi-resolution generation strategy via a few-step DIT upsampler to increase video fidelity. Our final model, which contains a 14B DIT base model and a 14B DIT upsampler, achieves competitive performance against other popular open-source models, while being an order of magnitude faster. We discuss our model design as well as training strategies in this report.

cs.CV↗
arXiv2026

Design Perspective on Materials Experience: A CiteSpace-Based Bibliometric and Visual Analysis of Interdisciplinary Research

Based on a bibliometric analysis of literature from 2005 to 2024, this study reveals that material experience is undergoing a profound transformation characterized by evolving material definitions, methodological advances, and increasing interdisciplinary integration. Material types now extend beyond traditional substances to encompass virtual and biological media, underscoring a growing emphasis on perception and interaction. Methodologically, the field has transitioned from subjective descriptions to data-driven, quantifiable models focused on objective sensory analysis and multisensory integration to enhance immersion. Key drivers, including human-machine perception convergence, material-driven interface interactions, and the embedding of intelligent interactive functions, propel the discipline toward an experience-centered paradigm reflecting a deep convergence of design, science, and technology. At the national/regional level, the United States, China, Japan, Germany, and the Netherlands lead in contributions, while France, the United Kingdom, and Romania demonstrate significant interdisciplinary progress. At the institutional level, Delft University of Technology, Justus Liebig University Giessen, and the Centre National de la Recherche Scientifique show significant advantages. In particular, the Material-Driven Design theory has established a foundational impact on the discipline, while, regarding general research trends, scholars from the United States, the Netherlands, and Germany maintain the highest academic visibility. Overall, material experience research is at a critical juncture, its future development will depend on progress in material innovation, technological integration, and perceptual quantification, as well as the establishment of socio-cultural values, all of which must be effectively unified through design to address complex evolving needs.

cs.HC↗
arXiv2025

ObjectAdd: Adding Objects into Image via a Training-Free Diffusion Modification Fashion

We introduce ObjectAdd, a training-free diffusion modification method to add user-expected objects into user-specified area. The motive of ObjectAdd stems from: first, describing everything in one prompt can be difficult, and second, users often need to add objects into the generated image. To accommodate with real world, our ObjectAdd maintains accurate image consistency after adding objects with technical innovations in: (1) embedding-level concatenation to ensure correct text embedding coalesce; (2) object-driven layout control with latent and attention injection to ensure objects accessing user-specified area; (3) prompted image inpainting in an attention refocusing & object expansion fashion to ensure rest of the image stays the same. With a text-prompted image, our ObjectAdd allows users to specify a box and an object, and achieves: (1) adding object inside the box area; (2) exact content outside the box area; (3) flawless fusion between the two areas

cs.CV↗
arXiv2025

Step-Audio-R1 Technical Report

Recent advances in reasoning models have demonstrated remarkable success in text and vision domains through extended chain-of-thought deliberation. However, a perplexing phenomenon persists in audio language models: they consistently perform better with minimal or no reasoning, raising a fundamental question - can audio intelligence truly benefit from deliberate thinking? We introduce Step-Audio-R1, the first audio reasoning model that successfully unlocks reasoning capabilities in the audio domain. Through our proposed Modality-Grounded Reasoning Distillation (MGRD) framework, Step-Audio-R1 learns to generate audio-relevant reasoning chains that genuinely ground themselves in acoustic features rather than hallucinating disconnected deliberations. Our model exhibits strong audio reasoning capabilities, surpassing Gemini 2.5 Pro and achieving performance comparable to the state-of-the-art Gemini 3 Pro across comprehensive audio understanding and reasoning benchmarks spanning speech, environmental sounds, and music. These results demonstrate that reasoning is a transferable capability across modalities when appropriately anchored, transforming extended deliberation from a liability into a powerful asset for audio intelligence. By establishing the first successful audio reasoning model, Step-Audio-R1 opens new pathways toward building truly multimodal reasoning systems that think deeply across all sensory modalities.

cs.AI↗
arXiv2025

Step-Audio-EditX Technical Report

We present Step-Audio-EditX, the first open-source LLM-based audio model excelling at expressive and iterative audio editing encompassing emotion, speaking style, and paralinguistics alongside robust zero-shot text-to-speech (TTS) capabilities. Our core innovation lies in leveraging only large-margin synthetic data, which circumvents the need for embedding-based priors or auxiliary modules. This large-margin learning approach enables both iterative control and high expressivity across voices, and represents a fundamental pivot from the conventional focus on representation-level disentanglement. Evaluation results demonstrate that Step-Audio-EditX surpasses both MiniMax-2.6-hd and Doubao-Seed-TTS-2.0 in emotion editing and other fine-grained control tasks.

cs.CL↗