Searcharxiv⌕ Search

arXiv subjects

Jie Xiao

Publications and source records attributed to Jie Xiao.

At least 37 records · Page 2Linked to original sources
arXiv2026

MERA: Model Evolution and Routing with Skill Adaptation for Agentic Systems at Scale

LLM agents execute heterogeneous sequences of model calls within a single task: some invocations require careful reasoning, while others are structured steps such as formatting or tool-argument construction. Prior routing methods exploit this asymmetry by assigning easy invocations to a cheaper small model and difficult ones to a large model. Such policies reduce inference cost, but they leave the small model's capability unchanged, so attainable savings remain bounded by the work the student can already solve. MERA instead improves the small model itself, using a single model invocation as the unit of adaptation. In each cycle, MERA replays failed student invocations to obtain execution-verified teacher demonstrations, distills recurring procedures into an iteratively updated SkillBook, and fine-tunes a student LoRA adapter via supervised learning and optional GRPO. Routing serves as supporting machinery for deployment: the improved student is served behind a cost-calibrated router with verifier-backed fallback, and a candidate SkillBook, adapter, or router is admitted only when joint replay preserves task quality. Empirically, four-cycle adaptation raises Qwen2.5-Coder-1.5B from 28.7% to 49.7% pass on held-out HumanEval+MBPP. Under verifier-backed fallback, the deployed policy retains 88.3% pass at 60.8% of always-Luna cost. On TAU-2, a fine-tuned Qwen3.5-2B improves from 14/35 to 18/35 and matches an unadapted 4B model. These results indicate that verifier-backed multi-cycle adaptation can increase small-model capability, rather than only routing around a fixed student.

cs.LG↗
arXiv2026

A Counterexample to the Liu--Lou--Zhu $\mathcal Q_p$--Carleson Embedding Conjecture

In this paper, we disprove a conjecture of Liu, Lou, and Zhu concerning Carleson embeddings of $\mathcal Q_p$ spaces into tent spaces for $0<p<1$. More precisely, we construct a finite positive $p$-Carleson measure $μ$ on the unit disc $\mathbb D$ such that the canonical embedding $$ \operatorname{id}:\mathcal Q_p \longrightarrow \mathcal T_{p,2}^2(μ) $$ is not bounded. The main ingredient is a new family of $\mathcal Q_p$ test functions that encodes Cantor-type structures on the unit circle $\mathbb T$ into the analytic behavior of functions in $\mathcal Q_p$. This construction is inspired by ideas developed in a recent work of the first and third authors on composition operators on $\mathcal Q_p$ spaces.

math.CV↗
arXiv2026

On the holomorphic differential operator $\frac{d}{dz}: Q_K\to L^q(WdA)$

In this paper, we obtain non-testing characterizations, in terms of dyadic capacity gauges, of the boundedness and compactness of the differentiation operator $$ \frac{d}{dz}:Q_K\longrightarrow L^q(W\,dA), \qquad 0<q<\infty. $$ We also characterize the limiting case as $q\to0^+$, formulated in terms of a logarithmic geometric mean, while the endpoint $q=\infty$ is treated separately using a standard testing argument. These results greatly extend the previous work on ${\mathcal Q}_p$-spaces to the general setting of $Q_K$-spaces. As applications, we characterize composition operators and Volterra-type integral operators between different $Q_K$-spaces. In particular, the off-diagonal characterization established here, together with the previously established diagonal case, completely resolves Zhao's 2009 open question on composition operators between ${\mathcal Q}_p$-spaces.

math.CV↗
arXiv2026

On the growth of Bloch functions

We prove that there exist two Bloch functions $f_1$ and $f_2$ on $\mathbb D$ such that $$ |f_1(z)|+|f_2(z)| \geq \left(\log\frac{1}{1-|z|}\right)^{1/2}, \qquad z\in\mathbb D, $$ thereby resolving an open problem posed in 2008 by Girela, Peláez, Pérez-González and Rättyä. Our proof is based on a new Szegő-type recursion involving $\mathbb C^2$-valued polynomials and their reciprocal polynomials.

math.CV↗
arXiv2026

Proximity-Induced Skyrmion Stabilization at the Cu2OSeO3/Bi2Se3 Interface

We investigate proximity-induced magnetic interactions at the interface between the topological insulator Bi2Se3 and the chiral magnetic insulator Cu2OSeO3, with particular focus on the low temperature skyrmion phase. Broadband ferromagnetic resonance spectroscopy reveals enhanced stability of noncollinear spin textures in the Cu2OSeO3/Bi2Se3 heterostructure compared with bare Cu2OSeO3. In addition to an extra resonance mode in the tilted conical phase that is absent in bare Cu2OSeO3, field cycling resolves two counterclockwise skyrmion resonance branches separated by approximately 238 MHz, consistent with the coexistence of a bulk skyrmion lattice and an interfacial skyrmion phase stabilized by proximity-induced exchange coupling and enhanced interfacial Dzyaloshinskii-Moriya interactions. The finite frequency separation indicates that the two skyrmion phases occupy distinct magnetic energy landscapes while retaining similar resonance character. Resonant elastic x-ray scattering measurements further confirm that the interfacial skyrmion phase spans a broader magnetic-field range than the bulk phase, demonstrating enhanced stability and ordering of topological spin textures at the interface. These findings establish interface engineering as a promising route for extending the stability regime of skyrmion and tilted-conical phases in topological-magnetic heterostructures.

cond-mat.mtrl-sci↗
arXiv2026

Lusztig sheaves and integrable highest weight modules in the symmetrizable case

This paper continues the work of \cite{fang2023lusztigsheavesintegrablehighest} and \cite{fang2023lusztigsheavestensorproducts}. For a symmetrizable generalized Cartan matrix $C$ and the corresponding quantum group $\mathbf{U}$, we consider an associated quiver $Q$ equipped with an admissible automorphism $a$. We construct a category $\widetilde{\mathcal{Q}/\mathcal{N}}$ obtained from localizations of Lusztig sheaves for the corresponding framed and $2$-framed quivers with automorphism. The Grothendieck groups of these categories realize the integrable highest weight module $L(λ)$ and the tensor product $L(λ_1)\otimes L(λ_2)$ of integrable highest weight $\mathbf{U}$-modules. After quotienting by traceless objects, Lusztig sheaves yield the signed canonical bases of $L(λ)$ and $L(λ_1)\otimes L(λ_2)$. As applications, we recover symmetrizable crystal structures on Nakajima quiver varieties, Nakajima tensor product varieties, and Lusztig nilpotent varieties of preprojective algebras.

math.RT↗
arXiv2026

Geometric realization of affine bases: the Kronecker quiver case

In this paper, we study the transition matrix between the PBW basis and the canonical basis for the negative part of the quantized enveloping algebra of the Kronecker quiver from a geometric viewpoint. Building on Lusztig's geometric construction of the canonical basis, we construct sheaf-complex realizations of PBW basis elements by means of flag sheaf complexes over the strata $X(α,m)$ of representation varieties. Our first goal is to give a geometric description of the simple constituents appearing in the restrictions of these flag sheaf complexes to the strata $X(α,m)$. This allows us to compare the PBW-type sheaf complexes with the simple perverse sheaves $IC(X(α),L_χ)$ arising in Lusztig's construction. Using this description together with a purity result for the relevant $\mathbb{F}_q$-structures, we obtain another proof that the elements defined by Lusztig's perverse sheaves indeed form a basis of the composition algebra.Our second goal is to make the transition coefficients between the PBW basis and the canonical basis geometrically explicit. More precisely, we show that these coefficients are governed by the multiplicities of local systems in the restrictions of intersection cohomology complexes to smaller strata. As a consequence, the transition matrix from the canonical basis to the PBW basis is upper triangular with diagonal entries equal to $1$, and its coefficients admit a direct geometric interpretation. In particular, in the Kronecker quiver case we recover the triangularity of the transition matrix and obtain positivity properties of the corresponding coefficient polynomials.

math.QA↗
arXiv2026

The iterated geometric Green's formula

Fang, Lan, and Xiao established the geometric Green's formula as a categorical isomorphism for arbitrary semisimple complexes. In this short note, we generalize their work to multi-step compositions. Specifically, we establish the iterated geometric Green's formulas for the composition of an $(n-1)$-fold restriction and an induction, as well as its dual.

math.RT↗
arXiv2026

T2S: A Rehearsal-Based Approach for Extraction-Resistant Model Watermarking

Model watermarking safeguards AI model intellectual property by embedding distinctive knowledge that induces unique behavioral signatures. The primary technical challenge lies in ensuring watermark robustness against various post-processing attacks on the watermarked model. Model extraction attacks emerge as the most severe threat, where adversaries exploit prediction outputs to train surrogate models that illegally replicate the original model's functionality. In this work, we propose a rehearsal-based watermark embedding framework to enhance the robustness of model watermarks against model extraction attacks. By simulating the extraction process, our method leverages the loss of a \textit{simulated stolen model} on a trigger set as a training signal to fine-tune the watermark knowledge within the target model. This fine-tuning step encourages the watermark to be embedded in a way that boosts transferability, thereby increasing its chances of persisting and remaining detectable in stolen models. Comprehensive experiments conducted under diverse settings demonstrate that the proposed method significantly improves the robustness of model watermarks against both model extraction and subsequent watermark removal attacks.

cs.CR↗
arXiv2026

Mitosis Detection in the Wild: Multi-Tumor and Context-Aware Generalization in the MIDOG 2025 Challenge

Automated mitosis detection is a well-established task in computational pathology. While previous benchmarks focused on scanner-induced domain shift, clinical "real-world" application requires models to be robust across the vast variance to be expected in the histological landscape. The MItosis DOmain Generalization (MIDOG) 2025 challenge was designed to evaluate algorithmic performance across unprecedented biological and contextual diversity. We curated a test dataset of 365 cases, encompassing 12 distinct human, canine and feline tumor types, digitized across multiple scanning platforms. Moving beyond hand-selected hotspots, the challenge required detection also in random tissue areas (representative of the whole slide detection situation) and challenging areas (areas rich in hard negatives). In the second track, we introduced the classification of atypical mitotic figures (AMFs). There were 18 teams submitting to the detection track, with F1 scores ranging up to 0.740. In the AMF detection track, we had 21 submissions with balanced accuracy values up to 0.908. Our analysis reveals that while most models perform reliably in traditional hotspots, significant performance degradation occurs in challenging ROIs, where false positive rates tripled. Furthermore, performance varied significantly across the 12 tumor types, highlighting "blind spots" in current state-of-the-art architectures when encountering rare or highly pleomorphic malignancies. Moreover, we evaluated the effectiveness of ensembling and found a mean increases of 1.5 and 1.3 percentage points in F1 score and balanced accuracy, respectively. In contrast, TTA showed no relevant improvement. MIDOG 2025 demonstrates that "in the wild" mitosis detection remains a significant hurdle. The transition from hotspot-only evaluation to a multi-contextual framework provides a more realistic proxy for clinical reliability.

cs.CV↗
arXiv2026

Robust and Generalizable Safety Steering for Text-to-Image Diffusion Transformers

Diffusion Transformers have become a powerful backbone for text-to-image generation, but their layered and cross-modal generation process makes safety control fundamentally different from prompt-level filtering or output-level detection. Harmful semantics may be weakly expressed in text representations, progressively bound to visual latents, and finally entangled with rendering dynamics. As a result, safety steering at a fixed layer can be unstable, and a steering mechanism learned from known risks may not transfer reliably to a shifted target risk domain. We propose SafeDIG, a safety steering framework that formulates DiT safety adaptation as position-aware sparse feature transfer. SafeDIG first constructs Sparse Autoencoders over functionally distinct DiT intervention positions and uses robustness-aware pre-training routing to prioritize intervention sites that are expected to remain stable under source-target risk shift. It then separates transferable safety features from domain-specific activation geometry by freezing the SAE encoder as a reusable sparse safety dictionary and adapting only the decoder to the target-domain activation manifold. During inference, SafeDIG combines Blend and Repel operations to steer unsafe activations toward transferred safety manifolds or away from harmful sparse directions. Experiments on FLUX.1 Dev and Stable Diffusion 3.5 Large show that SafeDIG consistently reduces target-domain and overall unsafe generation rates while preserving source-domain safety and image quality.

cs.AI↗
arXiv2026

Hunyuan3D 2.0: Scaling Diffusion Models for High Resolution Textured 3D Assets Generation

We present Hunyuan3D 2.0, an advanced large-scale 3D synthesis system for generating high-resolution textured 3D assets. This system includes two foundation components: a large-scale shape generation model -- Hunyuan3D-DiT, and a large-scale texture synthesis model -- Hunyuan3D-Paint. The shape generative model, built on a scalable flow-based diffusion transformer, aims to create geometry that properly aligns with a given condition image, laying a solid foundation for downstream applications. The texture synthesis model, benefiting from strong geometric and diffusion priors, produces high-resolution and vibrant texture maps for either generated or hand-crafted meshes. Furthermore, we build Hunyuan3D-Studio -- a versatile, user-friendly production platform that simplifies the re-creation process of 3D assets. It allows both professional and amateur users to manipulate or even animate their meshes efficiently. We systematically evaluate our models, showing that Hunyuan3D 2.0 outperforms previous state-of-the-art models, including the open-source models and closed-source models in geometry details, condition alignment, texture quality, and etc. Hunyuan3D 2.0 is publicly released in order to fill the gaps in the open-source 3D community for large-scale foundation generative models. The code and pre-trained weights of our models are available at: https://github.com/Tencent/Hunyuan3D-2

cs.CV↗
arXiv2026

Heat dispersion laws in smooth compact manifolds

Given a Lipschitz conductor $K$ in the smooth compact Riemannian $2\le n$-manifold $(M,g)$, such a half generic heat dispersion law $$ {\rm H^d}_{p,\varPhi,\varPsi}(K,M)=2^{-1} {\rm H^d}_{Δ_p,\varPhi,\varPsi}(K,M) $$ is not only newly-established via Theorem 1.1 but also deeply-explored through not only Proposition 3.1 (a comparison law for the generic heat dispersion) but also Proposition 3.2 (a recycling law for the quasilinear Laplace-Robin eigenvalue).

math.DG↗
arXiv2026

Wan-Image: Pushing the Boundaries of Generative Visual Intelligence

We present Wan-Image, a unified visual generation system explicitly engineered to paradigm-shift image generation models from casual synthesizers into professional-grade productivity tools. While contemporary diffusion models excel at aesthetic generation, they frequently encounter critical bottlenecks in rigorous design workflows that demand absolute controllability, complex typography rendering, and strict identity preservation. To address these challenges, Wan-Image features a natively unified multi-modal architecture by synergizing the cognitive capabilities of large language models with the high-fidelity pixel synthesis of diffusion transformers, which seamlessly translates highly nuanced user intents into precise visual outputs. It is fundamentally powered by large-scale multi-modal data scaling, a systematic fine-grained annotation engine, and curated reinforcement learning data to surpass basic instruction following and unlock expert-level professional capabilities. These include ultra-long complex text rendering, hyper-diverse portrait generation, palette-guided generation, multi-subject identity preservation, coherent sequential visual generation, precise multi-modal interactive editing, native alpha-channel generation, and high-efficiency 4K synthesis. Across diverse human evaluations, Wan-Image exceeds Seedream 5.0 Lite and GPT Image 1.5 in overall performance, reaching parity with Nano Banana Pro in challenging tasks. Ultimately, Wan-Image revolutionizes visual content creation across e-commerce, entertainment, education, and personal productivity, redefining the boundaries of professional visual synthesis.

cs.CV↗
arXiv2026

IR-Flow: Bridging Discriminative and Generative Image Restoration via Rectified Flow

In image restoration, single-step discriminative mappings often lack fine details via expectation learning, whereas generative paradigms suffer from inefficient multi-step sampling and noise-residual coupling. To address this dilemma, we propose IR-Flow, a novel image restoration method based on Rectified Flow that serves as a unified framework bridging the gap between discriminative and generative paradigms. Specifically, we first construct multilevel data distribution flows, which expand the ability of models to learn from and adapt to various levels of degradation. Subsequently, cumulative velocity fields are proposed to learn transport trajectories across varying degradation levels, guiding intermediate states toward the clean target, while a multi-step consistency constraint is presented to enforce trajectory coherence and boost few-step restoration performance. We show that directly establishing a linear transport flow between degraded and clean image domains not only enables fast inference but also improves adaptability to out-of-distribution degradations. Extensive evaluations on deraining, denoising and raindrop removal tasks demonstrate that IR-Flow achieves competitive quantitative results with only a few sampling steps, offering an efficient and flexible framework that maintains an excellent distortion-perception balance. Our code is available at https://github.com/fanzh03/IR-Flow.

cs.CV↗
arXiv2026

Notes on Chevalley Groups and Root Category III: the Region of Total Positivity

In [4], we use the root categories to realize Chevalley groups. Lusztig's theory of total positivity for reductive groups can be naturally applied to Chevalley groups. In this paper, we explicitly determine regions of $\mathbb{R}_{>0}^t$ for describing the size of monoids of totally positive elements, with respect to the root subgroups corresponding to the indecomposable objects in the root category.

math.RT↗
arXiv2026

HY-World 2.0: A Multi-Modal World Model for Reconstructing, Generating, and Simulating 3D Worlds

We introduce HY-World 2.0, a multi-modal world model framework that advances our prior project HY-World 1.0. HY-World 2.0 accommodates diverse input modalities, including text prompts, single-view images, multi-view images, and videos, and produces 3D world representations. With text or single-view image inputs, the model performs world generation, synthesizing high-fidelity, navigable 3D Gaussian Splatting (3DGS) scenes. This is achieved through a four-stage method: a) Panorama Generation with HY-Pano 2.0, b) Trajectory Planning with WorldNav, c) World Expansion with WorldStereo 2.0, and d) World Composition with WorldMirror 2.0. Specifically, we introduce key innovations to enhance panorama fidelity, enable 3D scene understanding and planning, and upgrade WorldStereo, our keyframe-based view generation model with consistent memory. We also upgrade WorldMirror, a feed-forward model for universal 3D prediction, by refining model architecture and learning strategy, enabling world reconstruction from multi-view images or videos. Also, we introduce WorldLens, a high-performance 3DGS rendering platform featuring a flexible engine-agnostic architecture, automatic IBL lighting, efficient collision detection, and training-rendering co-design, enabling interactive exploration of 3D worlds with character support. Extensive experiments demonstrate that HY-World 2.0 achieves state-of-the-art performance on several benchmarks among open-source approaches, delivering results comparable to the closed-source model Marble. We release all model weights, code, and technical details to facilitate reproducibility and support further research on 3D world models.

cs.CV↗