arXiv subjects

Yu Jiang

Publications and source records attributed to Yu Jiang.

At least 19 recordsLinked to original sources

arXiv2026

Interpreting Object-Dependent Concept Brittleness in Text-to-Image Diffusion Models

Although text-to-image diffusion models generally exhibit strong prompt-following ability, we identify a persistent and previously underexplored failure pattern in which a small subset of prompts differing only in the object consistently fails to realize the same target concept under identical generation settings. We term this phenomenon object-dependent concept brittleness. Such cases suggest systematic internal blind spots rather than random sampling noise. In this paper, we present an interpretability-oriented framework to audit and minimally correct these failures. Our key idea is to analyze denoising trajectories in a step-wise sparse autoencoder (SAE) space, where abstract style and attribute concepts become more separable than in the raw denoising representation. This sparse space enables us to compare successful and failed generations, identify concept dimensions whose evidence is missing, weakened, or temporally delayed, and construct class-level concept prototypes from reliable class-consistent samples. Based on this audit process, we introduce a lightweight inference-time correction strategy that interpolates denoising features toward the corresponding prototype in SAE space. Rather than serving as a task-specific retraining method, this intervention acts as a validation of the diagnosed concept deficiency. We evaluate the proposed framework on style and attribute failure cases across multiple diffusion backbones, with significant improvements in concept consistency, text fidelity, and repair success. Further analyses show that deeper denoising representations provide clearer concept structure, while early-stage intervention offers the strongest correction leverage. Code is available at https://github.com/Metecade/Object-Dependent-Concept-Brittleness.

cs.CV↗

arXiv2026

Investigations of MWISP Bubbles: Identification and Analysis of Enclosed Molecular Bubbles by Weight Fields

Molecular bubbles are widely used as tracers of stellar feedback; yet, their identification in spectral-line surveys remains challenging because both cavity morphology and kinematic structure must be assessed consistently in position--position--velocity (PPV) space. We present the Bubble-Weight Fields (BWFields) framework, a PPV-based method that for the first time enables the automated and objective identification and analysis of enclosed molecular bubbles directly from spectral-line data cubes. BWFields constructs a bubble-weight field, $W_{l,b,v}$, which encodes cumulative evidence for cavity interiors by aggregating topological signatures across multiple signal-to-noise tiers and velocity-integration scales. Contiguous cavity interiors are segmented as weight-clumps and associated with surrounding molecular gas, linking candidate bubbles to the structure of their host clouds. Shell morphology is characterized using radial intensity profiles and emission-defined intensity skeletons, which capture the shell geometry as traced by the observed emission. Bubble kinematics are quantified using azimuthally sampled position-velocity (PV) diagnostics, along with a turbulence-normalized expansion significance, which serves as a direct measure of the expansion-like velocity organisation. Applied to MWISP $^{13}$CO observations of the G17 region, BWFields identifies a population of bubble candidates with a broad range of morphologies and velocity structures in complex environments. BWFields establishes a scalable and physically interpretable framework for molecular-bubble studies in large surveys, enabling systematic investigations of stellar feedback in the Galactic interstellar medium.

astro-ph.GA↗

arXiv2026

Tool Specifications Matter: Uncovering and Mitigating Safety Risks in AI Agents

AI agents extend large language models (LLMs) with external tools, enabling them to perform complex tasks and translate model outputs into consequential real-world actions. Yet LLMs often become substantially less safe when deployed as agents, and the source of this degradation remains poorly understood. In this paper, we identify schema-formatted tool specifications as a primary source of agent safety degradation and show, through white-box representation analysis, that they weaken the model's internal refusal signals and contribute to unsafe tool execution. Building on this finding, we propose SafeKeep, an inference-time safeguard that decouples safety judgment from tool execution: it assesses requests using flattened textual tool specifications while retaining the original schema-formatted specifications for execution. Across two representative benchmarks and four LLMs, including both white-box and black-box models, SafeKeep increases the average refusal rate for harmful requests from 23.8% to 70.6% and reduces the average attack success rate under observation-level prompt injection from 25.6% to 2.5%. It also outperforms existing safeguards and preserves task-handling capability. We release the code and data at https://github.com/snowcatsmoking/SafeKeep .

cs.AI↗

arXiv2026

Asymmetric Floquet-Engineered Mode Coupling in Hybrid Magnonics

In hybrid magnonic systems, linear magnon--photon hybridization inherently produces symmetric, reciprocal interactions, precluding asymmetric mode coupling. Floquet driving can tailor mode coupling strengths but, with single-tone modulation, inevitably generates a symmetric interaction that preserves this reciprocity. Here we introduce dual-tone Floquet modulation to unlock a new degree of freedom in hybrid magnonic systems, where the relative phase $\theta$ of two commensurate drives continuously controls the asymmetry of the Floquet-engineered interaction, enabling asymmetric mode coupling absent in existing hybrid magnonic systems. We demonstrate this in a strongly coupled cavity magnonic device, where tuning $\theta$ reversibly switches single-sided Autler--Townes splitting between the two hybrid modes---a direct spectroscopic signature of phase-programmable asymmetric coupling. This approach opens a new path toward controllable nonreciprocal and topological functionalities in hybrid magnonic systems, with broad implications for advanced quantum and classical signal processing.

cond-mat.mes-hall↗

arXiv2026

Experimental Realization of Synthetic Magnonic Lattice via Floquet Engineering

Magnonic systems, which exploit spin-wave excitations in magnetic materials, offer a promising platform for coherent information processing due to their low dissipation, strong nonlinearities, and intrinsic nonreciprocity. However, scaling magnonic circuits remains challenging, particularly with low-loss insulators such as yttrium iron garnet (YIG), which are difficult to pattern. Here, we experimentally realize a synthetic dimension in a magnonic system by coupling multimode magnon resonances in the frequency domain using time-periodic Floquet modulation. This approach enables electronically tunable interactions between discrete modes within a single YIG device, forming a reconfigurable mode-space lattice that supports functionalities such as Bloch oscillation. Our results demonstrate that high-dimensional magnonic dynamics can be achieved without increasing device footprint, establishing synthetic dimensions as a scalable and programmable route for integrated magnonic technologies. This advancement positions magnonic systems as promising platforms for engineering emergent phenomena that are inaccessible at equilibrium.

cond-mat.mes-hall↗

arXiv2026

UXBench: Measuring the Actionability of LLM-Generated UX Critiques

Large language models (LLMs) are increasingly deployed as UX judges that inspect interfaces, diagnose usability problems, and propose repairs. Yet no controlled benchmark measures whether the resulting critiques are reliable and actionable across heterogeneous product surfaces. We introduce UXBench, a benchmark for evaluating LLMs as interaction-grounded UX judges. UXBench comprises local-first runnable web fixtures spanning ten product-surface families, paired with coverage-gated browser exploration that forces models to collect interaction evidence before reporting. Each judge model produces a structured UX report over seven rubric dimensions; report quality is measured by whether a fixed downstream repair agent can improve the interface based on the critique. We evaluate eight frontier models under both an automated repair-lift protocol and a blind human validation study. Results show that UX judging is neither saturated nor one dimensional: models differ meaningfully in report actionability, exhibit distinct rubric-level repair signatures, vary in fixture-level reliability, and trade leadership across surface categories

cs.SE↗

arXiv2026

HAFMat: Hybrid Priors Guided Adaptive Fusion for Single-Image Human Material Estimation

Physically based rendering (PBR) material estimation is a fundamental appearance decomposition task with broad applications in virtual content creation, relighting, and digital human rendering. However, estimating PBR materials from a single human image remains highly ill-posed, since illumination, geometry, and reflectance are heavily entangled in the observed appearance. To mitigate this ambiguity, we propose HAFMat, a hybrid-prior-guided framework for single-image human material estimation. Our method introduces guidance maps that encode complementary cues, including appearance, body geometry, structure, and prior material predictions from pre-trained models. A key observation is that these guidance cues are heterogeneous: some cues mainly provide texture-level constraints, while others convey higher-level semantic information. To exploit this property, we design a Multi-layer Adaptive Feature Fusion Mechanism, which adaptively fuses guidance features with decoder features at different stages. This design enables texture-dominant and semantic-dominant cues to guide material decoding at appropriate levels, leading to more accurate and physically plausible material estimation. Extensive experiments on both synthetic and real data demonstrate that our method achieves state-of-the-art performance in material estimation and downstream relighting.

cs.CV↗

arXiv2026

Achieving Heisenberg limit under noisy conditions with quantum Zeno dynamics and dynamical decoupling

Quantum Zeno dynamics (QZD) and dynamical decoupling (DD) are useful tools that enable the effective suppression of noise in quantum systems. We consider the problem of when (i) noise can be suppressed and (ii) Heisenberg limit (HL) can be achieved in quantum metrology, and prove necessary and sufficient conditions for when QZD and DD are useful for achieving these two goals. We further show this condition can hold even when the Hamiltonian-not-in-Lindblad-span (HNLS) condition fails, allowing QZD/DD to recover Heisenberg scaling beyond what is certified by the usual HNLS-based quantum error correction (QEC) criterion. Finally, we demonstrate that the combination of both techniques can allow individually imperfect QZD and DD strategies to saturate HL.

quant-ph↗

arXiv2026

Speech Meets ELF: Audio Conditional Continuous-Target Diffusion for Speech Recognition and Translation

Speech-to-text (S2T) systems for recognition (ASR) and translation (S2TT) typically generate discrete text tokens. In contrast, continuous-target language modelling performs generation in a continuous space, yet its potential for S2T remains unexplored. To bridge this gap, we propose ELF-S2T, an audio-conditioned continuous-target generative model for S2T. Built upon the pre-trained Embedded Language Flows (ELF) backbone, ELF-S2T processes speech via a frozen Whisper encoder and a single linear projector, prepending the resulting audio condition to the noisy text latent for in-context, flow-matching denoising. To prevent the model from over-relying on its pre-trained text context, we introduce audio forcing during training, and further amplify the audio condition via classifier-free guidance at inference. Experiments on LibriSpeech and CoVoST2 show that ELF-S2T achieves competitive ASR and S2TT performance. Crucially, our error analysis reveals that, although ASR and S2TT errors look very different on the surface, both stem from the same underlying cause, a close distance confusion in the continuous latent space. This finding naturally aligns with the continuous representation generation paradigm, indicating a common semantic mapping process beneath recognition and translation. Our code and pretrained models are publicly available at https://github.com/Sslnon/ELF-S2T.

cs.SD↗

arXiv2026

On Terwilliger $\mathbb{F}$-algebras of factorial association schemes II

The Terwilliger algebras of association schemes over an arbitrary field $\mathbb{F}$ were called the Terwilliger $\mathbb{F}$-algebras of association schemes in [10]. In [7], He and Jiang studied the Terwilliger $\mathbb{F}$-algebras of factorial association schemes. In this paper, we continue studying the Terwilliger $\mathbb{F}$-algebras of factorial association schemes. We get all block idempotents of the Terwilliger $\mathbb{F}$-algebras of factorial association schemes. We get the $\mathbb{F}$-dimensions, the centers, the Jacobson radicals of the block algebras of the Terwilliger $\mathbb{F}$-algebras of factorial association schemes.

math.CO↗

arXiv2026

CODEFUSE-DEBENCH: An Empirical Study on Readability, Recompilability, and Functionality

Binary decompilation aims to recover binaries into high-level source code, but existing evaluations mainly rely on syntactic similarity or single-axis readability metrics, which fail to capture practical reusability. We propose a reusability-driven evaluation paradigm that measures decompiler quality along three orthogonal dimensions: readability, recompilability, and functionality. We present DEBENCH, the first automated framework for multidimensional decompilation evaluation. DEBENCH contains 240 atomic test functions, organized into 8 source files and compiled into 640 binaries. It combines LLM-as-judge readability scoring with URAF (18 sub-dimensions), iterative compile-and-repair under a fixed 50-iteration budget, and Frida-based differential dynamic tracing at the program, function, and instruction levels. We evaluate five mainstream decompilers and three repair LLMs. Our study reveals four findings. First, the reusability cliff is steep: the best decompiler-LLM pair reaches 22.3% Exact+Partial program-level behavioral overlap but only 1.2% exact stdout match, nearly 50 points below recompilability. Second, settings that maximize readability do not maximize functionality: -O3 yields the lowest readability but the highest functionality, and Clang gives lower readability than GCC but 2.6x higher functionality. Third, cross-decompiler variation at the functional level is 20x, far larger than the 1.6x cross-LLM variation, showing that progress depends more on decompiler engines than larger repair models. Fourth, failures fall into three categories: syntactic noise, type-system collapse (about 19% of repair errors), and irreversible upstream losses such as ARM64 relocation idioms and C++ ABI features.

cs.SE↗

arXiv2026

Towards Reliable Fetal Ultrasound Interpretation with Multi-Agent Collaboration

Automated fetal ultrasound interpretation requires a workflow from visual perception, including plane recognition and anatomical segmentation, to clinical understanding, including biometric measurement and diagnostic reporting. However, the prevailing "one-task, one-model" paradigm limits systematic integration of evidence across this multi-step process. Although multimodal large language models (MLLMs) show promising visual understanding, their limited domain-specific grounding and hallucination risks restrict reliability in fetal ultrasound analysis. To address these limitations, we propose FetUSAgents, a tool-augmented multi-agent system for comprehensive fetal ultrasound interpretation, supporting visual question answering (VQA), report generation, image captioning, and video summarization. FetUSAgents coordinates task-specific visual tools through collaborative LLM agents and decomposes clinical queries into subtasks that progress from anatomical recognition to quantitative measurement. We further introduce Dual-Path Evidence Arbitration (DPEA), which integrates LLM-based deliberative reasoning with structured computational evidence from specialized visual tools. A retrieval-enhanced evidence bank consolidates intermediate findings to support traceable and clinically grounded conclusions. In addition, we construct FetUS-VQA, a dedicated VQA benchmark for fetal ultrasound, comprising 1,892 images and 3,205 question-answer pairs across 10 clinical tasks. Extensive out-of-distribution experiments show that FetUSAgents outperforms general and medical MLLMs, exceeding the strongest baseline by more than 25 percent in VQA accuracy. These results suggest a scalable route toward evidence-driven clinical assistants for prenatal imaging. Code is available.

cs.CV↗

arXiv2026

ViViD-5K: Vineyard vision dataset for field-based berry detection and segmentation and grape cluster closure estimation

Cluster closure, defined as the progressive filling of gaps between the berries in a grape bunch, is a key trait in vineyard management, impacting disease risk. However, traditional visual scoring methods are labor-intensive, subjective, and lack temporal resolution. Existing datasets rarely support fine-grained berry-level analysis, limiting the development of robust deep learning models. In this work, we present ViViD-5k, a large-scale in-field Vineyard Vision Dataset containing 5,000 images with dense annotations, including over 648,000 berry centroids and cluster segmentation masks spanning 13 grape varieties. Building on this dataset, we introduce GrapeSAM, a two-stage visual pipeline that combines point-based berry localization with prompt-based segmentation using Segment Anything, followed by transformer-based cluster segmentation. The pipeline enables automated, in-field estimation of cluster closure with minimal supervision. Quantitative results demonstrate strong segmentation and counting accuracy across diverse conditions, while visualizations confirm robustness on both in-domain and out-of-domain samples. This work provides a scalable and objective alternative to manual compactness scoring and supports high-throughput grape phenotyping with enhanced spatial detail.

cs.CV↗

arXiv2026

Separate First, Fuse Later: Mitigating Cross-Modal Interference in Audio-Visual LLMs Reasoning with Modality-Specific Chain-of-Thought

Audio and vision provide complementary evidence for audio-visual question answering, yet current audio-visual large language models may suffer from cross-modal interference: information from one modality misguides the interpretation of another, thereby inducing hallucinations. We attribute this issue to uncontrolled cross-modal interactions during intermediate reasoning. To mitigate this, we propose Separate First, Fuse Later (SFFL), an audio-visual reasoning framework designed to reduce cross-modal interference. SFFL enforces modality-specific chain-of-thought reasoning, producing separate audio and visual reasoning traces and integrating evidence for answering. We construct modality-preference labels via a data pipeline under different modality input settings. We use these labels as an auxiliary reward in reinforcement learning to encourage a instance-dependent preference for modality cues when answering. We further introduce a modality-specific reasoning mechanism that preserves modality isolation during the separated reasoning stage while enabling full access to cross-modal information at the evidence fusion stage. Experiments demonstrate consistent improvements in both accuracy and robustness, yielding an average relative gain of 5.16\% on general AVQA benchmarks and 11.17\% on a cross-modal hallucination benchmark.

cs.AI↗

arXiv2026

Evaluating the Expressive Appropriateness of Speech in Rich Contexts

Evaluating expressive speech remains challenging, as existing methods mainly assess emotional intensity and overlook whether a speech sample is expressively appropriate for its contextual setting. This limitation hinders reliable evaluation of speech systems used in narrative-driven and interactive applications, such as audiobooks and conversational agents. We introduce CEAEval, a Context-rich framework for Evaluating Expressive Appropriateness in speech, which assesses whether a speech sample expressively aligns with the underlying communicative intent implied by its discourse-level narrative context. To support this task, we construct CEAEval-D, the first context-rich speech dataset with real human performances in Mandarin conversational speech, providing narrative descriptions together with fifteen dimensions of human annotations covering expressive attributes and expressive appropriateness. We further develop CEAEval-M, a model that integrates knowledge distillation, planner-based multi-model collaboration, adaptive audio attention bias, and reinforcement learning to perform context-rich expressive appropriateness evaluation. Experiments on a human-annotated test set demonstrate that CEAEval-M substantially outperforms existing speech evaluation and analysis systems.

eess.AS↗

arXiv2026

Conditonal Lipschitz stability for the Inverse Problem of the 2D Navier-Stokes System in a Bounded Domain

This paper concerns an inverse problem for the initial boundary value problem of the two-dimensional Navier-Stokes system defined in a bounded simply connected domain with slip, vorticity boundary conditions, and a global vorticity invariant constraint. We establish conditional Lipschitz stability and a local recovery for this inverse problem, where the velocity field and space-independent boundary vorticity are locally recovered from the given initial velocity field and the global vorticity invariant. Our analysis is based on well-posedness estimates and energy methods for the vorticity transport equation.

math.AP↗

arXiv2026

Rethinking Semantic Collaborative Integration: Why Alignment Is Not Enough

Large language models (LLMs) have become an important semantic infrastructure for modern recommender systems. A prevailing paradigm integrates LLM-derived semantic embeddings with collaborative representations via representation alignment, implicitly assuming that the two views encode a shared latent entity and that stronger alignment yields better results. We formalize this assumption as the global low-complexity alignment hypothesis and argue that it is stronger than necessary and often structurally mismatched with real-world recommendation settings. We propose a complementary perspective in which semantic and collaborative representations are treated as partially shared yet fundamentally heterogeneous views, each containing both shared and view-specific factors. Under this shared-plus-private latent structure, enforcing global geometric alignment may distort local structure, suppress view-specific signals, and reduce informational diversity. To support this perspective, we develop complementarity-aware diagnostics that quantify overlap, unique-hit contribution, and theoretical fusion upper bounds. Empirical analyses on sparse recommendation benchmarks reveal low item-level agreement between semantic and collaborative views and substantial oracle fusion gains, indicating strong complementarity. Furthermore, controlled alignment probes show that low-capacity mappings capture only shared components and fail to recover full collaborative geometry, especially under distribution shift. These findings suggest that alignment should not be treated as the default integration principle. We advocate a shift from alignment-centric modeling to complementarity fusion-centric, complementarity-aware design, where shared factors are selectively integrated while private signals are preserved. This reframing provides a principled foundation for the next generation of LLM-enhanced recommender systems.

cs.IR↗

arXiv2026

Uniqueness of dynamic elastography for isotropic standard linear solid viscoelastic media

Dynamic elastography is a widely used, safe, convenient, and cost-effective method to aid in medical diagnosis. It visualizes the wave field propagating through living tissues and quantitatively determines the wave propagation speed from the acquired data, thereby enabling the extraction of the viscoelastic properties of in vivo tissues. Notably, this identification process relies on the mathematical modeling of the viscoelastic characteristics of living tissues. When living tissues are simply modeled as isotropic elastic media, J. McLaughlin and J. Yoon established the uniqueness of the identification in \cite{MY} by reasoning that they called the ``shrink and spread argument". Given the realistic viscoelastic nature of biological tissues, generalizing their results by adopting viscoelastic models is of great significance. In this paper, using their reasoning, we prove the uniqueness of identification for two typical viscoelastic media: the isotropic extended Maxwell model and the isotropic extended standard linear solid model. More precisely, we demonstrate that the shear wave speed within a region of interest $\Omega$ can be uniquely determined from a single measurement of the wave field in $\Omega$.

math.AP↗