arXiv subjects

Xin Fu

Publications and source records attributed to Xin Fu.

At least 19 recordsLinked to original sources

arXiv2026

LLM eBay: A Simple and Truthful Marketplace for Fine-Tuned LLMs Trading

Personalized fine-tuning of Large Language Models (LLMs) enhances user experience but incurs substantial computational costs. Recent studies show that initializing training from a style-similar fine-tuned model reduces adaptation overhead, turning such checkpoints into reusable, tradable assets. A marketplace for these checkpoints, however, faces information asymmetry: sellers may misreport a checkpoint's style, while buyers cannot verify its usefulness before purchase, which invites adverse selection. We introduce LLM eBay, a simple single-round marketplace for trading fine-tuned LLMs. Sellers are paid per purchase at a posted price, and a truncated quadratic proper scoring rule adjusts each payment according to how well the seller's style report agrees with a style signal that the platform measures by generating text on hidden prompts. We show that the score elicits the verifier's expected signal, characterize exactly when truthful reporting is a seller best response under per-purchase demand, and bound the downside risk from noisy verification. In an end-to-end study with real LoRA checkpoints on two base models and three buyer tasks, LLM eBay reduces measured convergence steps by 8.5\% relative to retrieval over self-reported style (95\% CI [1.0\%, 15.9\%]). In model-based market simulations built on a measured style-to-cost relationship, it cuts style-claim error by 96.5\% (from 0.424 to 0.015), recovers 96.3\% of Oracle welfare, saves 72.4\% of the cost of fine-tuning from the base model, and keeps all providers profitable in the default setting.

eess.SY↗

arXiv2026

WoVR+: World Model as Reliable Simulator for Versatile VLA Policy Optimization with RL

Action-conditioned world models can reduce costly real-robot interaction in reinforcement learning for Vision-Language-Action models (VLA-RL), but hallucinations can mislead policy optimization. Extending them to richer VLA interfaces and long-horizon tasks raises challenges in spatiotemporal and cross-modal consistency, exploration-reliability balance in RL, and compounding rollout errors. We present WoVR+, extending WoVR's hallucination-aware WM-RL framework by redesigning the learned environment and imagined interaction protocol: A VLA-interface-aligned world model provides multi-view prediction and proprioceptive updates while mitigating hallucinations. Chunk-wise noise preserves flow-based policy exploration while limiting prediction-degrading action jitter. For long-horizon tasks, stage-local imagined training combines stage-entry resets, stage-specific rewards, local horizons, and controllable sampling to shorten generated prefixes and balance training across stages. WoVR+ improves success over SFT by 21.09% and 16.7% on RoboTwin and 6 real-robot tasks. Long-horizon phone-packaging success rises from 30% to 75%. Videos are available at https://youmo445.github.io/WoVR_Plus.github.io/.

cs.RO↗

arXiv2026

Mind Your Own Role: Reducing Role Overstepping for Effective and Efficient Multi-Agent Collaboration via Dynamic Role Activation Steering

Multi-Agent Systems (MAS) have shown strong capability in solving complex tasks by assigning different agents specialized roles and corresponding subtasks. However, agents may overstep their assigned responsibilities and perform subtasks belonging to other roles, which can disrupt workflow, reduce system efficiency, and degrade task performance. To address this issue, we investigate role-related patterns in the model's internal activation space and uncover two key observations: (i) role instructions induce consistent directional shifts in activation space, and (ii) role-adherent and role-overstepping trajectories exhibit distinct activation patterns. Based on these observations, we propose Dynamic Role Activation Steering (DRAS), a lightweight and effective inference-time activation intervention approach to reduce role overstepping in multi-agent collaboration. Specifically, DRAS constructs a role steering vector and a lightweight role-overstepping detector for each role. During LLM generation, the detector estimates the probability of role overstepping from the current activation state, and DRAS dynamically adjusts the steering strength accordingly to guide the agent toward its assigned responsibilities. Experimental results show that DRAS effectively reduces the number of role-overstepping trajectories by up to 86.6% compared with Vanilla MAS, thereby improving task quality score by 10.0%--11.1%, achieving a 1.50$\times$--1.61$\times$ task-completion speedup, and reducing token consumption by 26.8%--38.7%.

cs.MA↗

arXiv2026

Fast and Accurate: Improving Personalized LLM Performance by Semantic Concentration Inspired Workload Split between RAG and Fine-Tuning

With ever increasing personalized large language model (LLM) services, the next question to answer is how to improve the quality of services/users' experiences, which is indicated by personalized LLM generation accuracy/quality and inference latency. Although state-of-the-art designs, e.g., parameter-efficient fine-tuning (PEFT), retrieval-augmented generation (RAG) or Full PEFT+RAG, can help, they largely ignore the heterogeneous semantic concentrations of users' historical data and cannot fully unleash the potential of personal data for personalized LLM performance enhancement. In this paper, we propose PEFT Union RAG (PUR), a novel ``fast and accurate'' approach that exploits the semantic concentration of users' personal historical data to split workload between PEFT and RAG, aiming to improve personalized LLM performance in terms of better generation accuracy/quality and lower inference latency. Briefly, PUR partitions user's personal historical data and assigns semantic concentrated ``core'' information to PEFT for offline parametric fine-tuning, while retaining dispersed information for RAG. During the online LLM inference, PUR directly generates responses for ``core'' queries covered by the PEFT region and invokes RAG only for sporadic ``peripheral'' queries. Experimental results show that, by appropriately splitting workload between PEFT and RAG, PUR consistently achieves the best generation accuracy/quality for various users across multiple tasks, while reducing inference latency by 14.4%--46.6% compared with the best performed method.

cs.LG↗

arXiv2026

Fundamental groups of Calabi-Yau spaces

We study the fundamental groups of a normal projective Calabi-Yau variety $X$ with klt singularities and of its regular locus $X_{reg}$. We prove that both groups are almost abelian, and that their ranks are determined by the irregularities of finite etale and quasi-etale covers, respectively. Our approach is differential geometric, and relies on recent advances in the geometric theory of complex Monge-Ampere equations.

math.DG↗

arXiv2026

Homotopy rigidity for quasitoric manifolds over a product of simplices

We consider a variation of the Cohomological Rigidity Problem: whether two quasitoric manifolds whose cohomology rings are isomorphic are homotopy equivalent. We show that if two quasitoric manifolds have isomorphic cohomology rings and the associated polytope of one of them is a product of simplices, then they are homotopy equivalent after localizing away from an explicit list of primes.

math.AT↗

arXiv2026

Rigid and Nonrigid Isotropic Nef Classes on K3 Surfaces

Let $X$ be a fixed complex K3 surface and let $0\neα\in\overline{Kah_X}$ be a $(1,1)$-class on the boundary of the Kähler cone with $α^2=0$. We build a novel connection between two seemingly unrelated, intensively studied problems: the rigidity of a closed positive current in the class $α$ and the collapse geometry of a sequence of Ricci-flat Kähler metrics in the Kähler class $α+tκ$ as $t\to 0^+$. More precisely, we prove that $α$ contains a unique closed positive $(1,1)$-current if and only if the Ricci-flat metrics in class $α+tκ$ collapse to a point in the Gromov--Hausdorff sense. Then we exhibit the first example of a nonrigid irrational nef class $α$ on a Kummer surface, which answers a question of Filip--Tosatti and Sibony--Soldatenkov--Verbitsky. We also construct many new examples of rigid nef classes on hyperkähler manifolds, which extend the work of Filip--Tosatti and Sibony--Soldatenkov--Verbitsky.

math.DG↗

arXiv2026

Band-Edge Homogenization and the Sound-Soft Limit of Finite Bubbly Crystals

We establish a quantitative sound-soft scattering limit for a finite bubbly crystal near the upper edge of the first Bloch band of the corresponding infinite crystal. The inclusions form a dense periodic array of period $\varepsilon$ in a bounded Lipschitz domain, and their density contrast is $δ=\varepsilon^2$, with fixed positive wave speeds. Under coordinate-reflection symmetry, we prove that the normalized capacitance symbol has a unique nondegenerate maximum on the Brillouin torus. Its Hessian defines a positive Dirichlet elliptic operator governing the limiting spectral detunings. A mean-constrained variational formulation allows us to compare the actual finite-array capacitance matrix with the truncated infinite-lattice operator and to prove exponential localization of their difference near the sample boundary. We obtain an $O(\varepsilon)$ norm-resolvent approximation and connect it to the full acoustic scattering problem, retaining the second-order Bloch correction required by the frequency scaling. For rescaled detunings outside the effective Dirichlet spectrum, the exterior $L^2$ and far-field discrepancies are $O(\varepsilon)$, while the interior $L^2$ field is $O(\varepsilon^{1/2})$. Numerical experiments illustrate the far-field convergence and the effective spectral modes. Explicit one-dimensional calculations describe finer Fabry--Pérot transmission windows and show that a fixed frequency strictly inside the first band need not have a unique scattering limit.

math.AP↗

arXiv2026

A Novel Semantic Manifold Alignment Attack against Embedding-to-Embedding Obfuscation in Privacy-Preserving LLMs

With the widespread applications of large language models (LLMs), privacy-preserving inference has become increasingly essential for sensitive queries. To balance privacy and utility, a series of lightweight obfuscation approaches has recently been proposed, where users locally transform plaintext embeddings into the fixed ciphertext ones. While such Embedding-to-Embedding Obfuscation (E2EO) schemes demonstrate considerable resilience against traditional token frequency and embedding inversion attacks, the core mechanism behind remains to be the large-scale one-to-one substitution, which provides no cryptographic guarantees. In this paper, we propose Proxy Manifold Alignment (PMA), a novel attack against E2EO in privacy-preserving LLMs. Our key observation is that E2EO schemes keep the original semantic structure, so that the obfuscated vector stream can be regarded as an unknown tokenizer-language whose symbols are the vectors themselves. Therefore, the proposed ciphertext to plaintext reconstruction attack can be formulated as a translation task from the unknown tokenizer-language to plaintext. Specifically, by only accessing the obfuscated vector stream, the target tokenizer and a public corpus, the PMA attack first employs Word2Vec to model the co-occurrence patterns within the obfuscated stream and the public corpus independently, and constructs two proxy vector embeddings. Then, the attack aligns the underlying manifolds of these two embeddings based on structural similarity. Finally, it maps the obfuscated vectors back to plaintext. Experimental results demonstrate that PMA consistently achieves higher plaintext recovery than other state-of-the-art attack methods.

cs.CR↗

arXiv2026

On a conjecture of Demailly-Peternell-Schneider: the Kahler case

Let $f:(X,Δ)\rightarrow Y$ be a surjective holomorphic map between two normal Kahler varieties, where (X,Δ) is a log canonical pair, -(KX+Δ) is nef and Y is Q-Gorenstein. In Fu-Guo-Song-Wang [18], it is proved that -KY is pseudo-effective if f is a projective morphism. In this paper, prove that -KY is pseudo-effective without assuming that f is projective

math.AG↗

arXiv2026

On the Datar-Mete-Song minimal slope conjecture

We prove a conjecture of Datar-Mete-Song \cite{DMS} characterizing $J$-slope semi-stability by the minimal $J$-slope. More precisely, for a semi-stable pair of Kähler classes $(α,β)$ on a compact Kähler manifold $X$, every big and nef birational test class has slope at least the topological $J$-slope, whereas an unstable pair admits a test class with strictly smaller slope. We also introduce the $J$-null locus of a semi-stable pair and prove that it is an analytic subset of $X$ if $X$ is a compact Kähler surface or a compact toric Kähler manifold. In the toric invariant case, we show that Murakami's \cite{Murakami} weak solution to the $J$-equation is smooth and Kähler on the dense big torus $(\mathbb{C}^*)^n$ of $X$.

math.AG↗

arXiv2026

Lightweight Chunk Selection for Mobile Retrieval-Augmented Generation

RAG improves the factual grounding of LLM by incorporating external knowledge, but deploying RAG on mobile and edge devices remains challenging because retrieved context increases computation and memory. A direct way to reduce this cost is to retain only one retrieved chunk before generation, but the top-ranked retrieved chunk is not always the most evidence-supporting one, since retrieval similarity does not necessarily imply evidential sufficiency. Existing context-reduction methods can improve context quality, but often require additional LLMs or compressors that are costly under a strict mobile budget. In this paper, we study lightweight RAG chunk selection as an evidence-alignment problem. Our selector combines three complementary feature sources: question hidden states that represent LLM-side query intent, MoE routing-derived expert signals that capture the generator's internal routing structure, and retrieved chunk embeddings that preserve candidate-side evidence geometry. A compact multilayer perceptron maps these features to an evidence prototype in the chunk embedding space, and the candidate most aligned with this prototype is selected by cosine similarity. For stricter deployment budgets, we further introduce an optional task-aware feature selection strategy to reduce the selector input dimension. To support supervised evaluation, we construct semantic chunk-correctness labels based on evidence sufficiency rather than answer-string containment. Experiments show that the proposed selector consistently improves rank-1 evidence selection over mobile-applicable baselines by an average of 2.5%. These results suggest that using LLM-side query representations and MoE routing information and aligning them with retrieval-side candidate embedding is an effective and parameter-efficient strategy for mobile-applicable RAG chunk selection.

cs.LG↗

arXiv2026

MVMD: A Multi-View Approach for Enhanced Mirror Detection

In 3D reconstruction, mirrors introduce significant challenges by creating distorted and fragmented spaces, resulting in inaccurate and unreliable 3D models. As 3D reconstruction typically relies on multi-view images to capture different perspectives of a scene, detecting and labeling mirrors in multi-view images before reconstruction can effectively address this issue. However, existing methods focus solely on single-image detection, overlooking the rich information provided by multi-view setups. To overcome this limitation, we propose MVMD, a novel Multi-View Mirror Detection method, along with the first database specifically designed for mirror detection in multi-view scenes. The design of MVMD is grounded in the inherent associations between objects seen from different views and those reflected inside and outside of mirrors. These relationships are learned through cross- and self-attention mechanisms. MVMD consists of three key blocks: the Inter-Views Block tracks the shifts of objects within mirrors caused by changes in viewpoint; the Intra-View Block detects object reflections inside mirrors; and the Refinement Block sharpens mirror boundaries and enhances detected details. Experimental results show that our method improves accuracy by up to 2.6% and IoU by up to 11.1%, compared to single-image mirror detection techniques. This substantial improvement makes MVMD particularly effective for computer vision tasks, especially in enhancing the accuracy of 3D reconstruction in mirror-dense environments.

cs.CV↗

arXiv2026

Minnaert resonances and higher-order acoustic modes for bubbles in a viscous fluid with surface tension

The aim of this paper is to account for viscosity, surface tension, and interactions between micro-bubbles in approximating their resonant behavior in the ultrasonic regime. Original asymptotic formulas for the resonance frequencies are derived in terms of the difference in the acoustic impedance at the interface between the gas and the fluid. Both low-frequency resonances (Minnaert resonances) and higher-frequency resonances, i.e., beyond the subwavelength regime, are considered. We also provide a resonant characterization for a system of several micro-bubbles.

math.AP↗

arXiv2026

DMT: Demographic Conditioning, Morphology-Enhanced Transformer for Cuffless Blood Pressure Estimation from PPG Signals

Blood pressure (BP) is a key marker for cardiovascular risk assessment and therapeutic decision-making, and Photoplethysmography (PPG) enables low-cost, wearable-friendly cuffless BP estimation. However, even with recent progress, many PPG-based models are trained with BP regression alone and may rely on amplitude-dominated shortcuts. In addition, demographic covariates that systematically modulate vascular compliance are often incorporated only via late fusion, limiting subject-specific representation learning. We propose a Transformer-based network for cuffless BP estimation from PPG signal, leveraging self-attention to capture long-range dependencies across multiple cardiac cycles. To account for subject-specific vascular differences, the model is conditioned on demographics via FiLM-style feature modulation applied through the attention and feed-forward sublayers of Transformer blocks. In addition, we add an auxiliary morphology head to guide the model to attend to BP-relevant waveform morphology associated with arterial stiffness and wave reflection. Under calibration-based evaluation protocols on the large-scale PulseDB dataset, the proposed method achieves MAE of 4.56 mmHg for systolic BP and 2.62 mmHg for diastolic BP, reducing errors by 47% and 50% compared with prior demographic-enhanced PPG baselines. The resulting lightweight, single-sensor model supports scalable and clinically grounded cuffless BP estimation in calibration-enabled deployment settings.

eess.SP↗

arXiv2026

The integral cohomology rings of four-dimensional toric orbifolds

Let $X(P,λ)$ be a 4-dimensional toric orbifold associated to a polygon $P$ and a characteristic function $λ$. Assuming that $X(P,λ)$ is locally smooth over a vertex of $P$, we determine the integral cohomology ring $H^*(X(P,λ);\Z)$ by constructing an explicit basis and expressing the cup products of the basis elements in terms of $P$ and $λ$.

math.AT↗

arXiv2026

Nationwide EHR-Based Chronic Rhinosinusitis Prediction Using Demographic-Stratified Models

Chronic rhinosinusitis (CRS) is a common heterogeneous inflammatory disorder that causes substantial morbidity and healthcare costs. CRS is difficult to identify early from routine encounters, as symptom presentations overlap with common conditions such as allergic rhinitis, and heterogeneous phenotypes further obscure risk patterns. Prior predictive studies often rely on single-institutional cohorts , which reduce population-level generalizability. To overcome this, we leveraged nationwide longitudinal EHR data from the \textit{All of Us} Research Program to predict CRS diagnosis using two years of pre-diagnostic history. To address extreme feature sparsity and dimensionality in coded EHR data, we implemented a hybrid feature-selection pipeline that combines prevalence-based statistical screening with model-based importance ranking, compressing approximately 110,000 candidate codes into 100 interpretable features. To capture demographic heterogeneity, we trained demographic stratified models across six adult sex and life-stage subgroups with subgroup-specific hyperparameter tuning. Our framework achieved an overall AUC of 0.8461, improving discrimination by 0.0168 over the best baseline. These results demonstrate that routinely collected EHR data may support population-representative CRS risk stratification and inform earlier triage and referral prioritization in primary care.

cs.LG↗

arXiv2026

Observability and Semiclassical Control for Schrödinger Equations on Non-compact Hyperbolic Surfaces

We study the observability of the Schrödinger equation on $X$, a non-compact covering space of a compact hyperbolic surface $M$. Using a generalized Bloch theory, functions on $X$ are identified as sections of flat Hilbert bundles over $M$. We develop a semiclassical analysis framework for such bundles and generalize the result of semiclassical control estimates in [Dyatlov and Jin, Acta Math., 220 (2018), pp. 297-339] to all flat Hilbert bundles over $M$, with uniform constants with respect to the choice of bundle. Furthermore, when the Riemannian cover $X \to M$ is a normal cover with a virtually Abelian deck transformation group $Γ$, we combine the uniform semiclassical control estimates on flat Hilbert bundles with the generalized Bloch theory to derive observability from any $Γ$-periodic open subsets of $X$. We also discuss applications of the uniform semiclassical control estimates in spectral geometry.

math.AP↗