Searcharxiv⌕ Search

arXiv subjects

Xin Zhang

Publications and source records attributed to Xin Zhang.

At least 217 records · Page 12Linked to original sources
arXiv2026

ZipVoice-Dialog: Non-Autoregressive Spoken Dialogue Generation with Flow Matching

Generating spoken dialogue is inherently more complex than monologue text-to-speech (TTS), as it demands both realistic turn-taking and the maintenance of distinct speaker timbres. While existing autoregressive (AR) models have made progress, they often suffer from high inference latency and stability issues. To overcome these limitations, we propose ZipVoice-Dialog, a non-autoregressive (NAR) zero-shot spoken dialogue generation model based on flow-matching. Observing that applying vanilla flow-matching to dialogue generation leads to poor speech intelligibility and turn-taking precision, we introduce two simple yet effective methods to adapt flow-matching architectures for dialogue generation: (1) a curriculum learning strategy to ensure robust speech-text alignment, and (2) speaker-turn embeddings to govern precise speaker turn-taking. Additionally, we introduce dedicated strategies to support stereo dialogue generation. Recognizing the lack of training datasets in this field, we curate and release OpenDialog, the first large-scale (6.8k hours) open-source spoken dialogue dataset derived from in-the-wild speech data. Moreover, for fair and rigorous evaluations, we established a benchmark to comprehensively evaluate dialogue generation models. Experiments demonstrate the effectiveness of the proposed methods and dataset, showing that ZipVoice-Dialog achieves superior performance in inference speed, intelligibility, speaker turn-taking accuracy, and speaker similarity. Our code, model checkpoints, and the OpenDialog dataset are publicly available at https://github.com/k2-fsa/ZipVoice.

eess.AS↗
arXiv2026

FedRio: Personalized Federated Social Bot Detection via Cooperative Reinforced Contrastive Adversarial Distillation

Social bot detection is critical to the stability and security of online social platforms. However, current state-of-the-art bot detection models are largely developed in isolation, overlooking the benefits of leveraging shared detection patterns across platforms to improve performance and promptly identify emerging bot variants. The heterogeneity of data distributions and model architectures further complicates the design of an effective cross-platform and cross-model detection framework. To address these challenges, we propose FedRio (Personalized Federated Social Bot Detection with Cooperative Reinforced Contrastive Adversarial Distillation framework. We first introduce an adaptive message-passing module as the graph neural network backbone for each client. To facilitate efficient knowledge sharing of global data distributions, we design a federated knowledge extraction mechanism based on generative adversarial networks. Additionally, we employ a multi-stage adversarial contrastive learning strategy to enforce feature space consistency among clients and reduce divergence between local and global models. Finally, we adopt adaptive server-side parameter aggregation and reinforcement learning-based client-side parameter control to better accommodate data heterogeneity in heterogeneous federated settings. Extensive experiments on two real-world social bot detection benchmarks demonstrate that FedRio consistently outperforms state-of-the-art federated learning baselines in detection accuracy, communication efficiency, and feature space consistency, while remaining competitive with published centralized results under substantially stronger privacy constraints.

cs.AI↗
arXiv2026

Revitalizing Black-Box Interpretability: Actionable Interpretability for LLMs via Proxy Models

Post-hoc explanations provide transparency and are essential for guiding model optimization, such as prompt engineering and data sanitation. However, applying model-agnostic techniques to Large Language Models (LLMs) is hindered by prohibitive computational costs, rendering these tools dormant for real-world applications. To revitalize model-agnostic interpretability, we propose a budget-friendly proxy framework that leverages efficient models to approximate the decision boundaries of expensive LLMs. We introduce a screen-and-apply mechanism to statistically verify local alignment before deployment. Our empirical evaluation confirms that proxy explanations achieve over 90% fidelity with only 11% of the oracle's cost. Building on this foundation, we demonstrate the actionable utility of our framework in prompt compression and poisoned example removal. Results show that reliable proxy explanations effectively guide optimization, transforming interpretability from a passive observation tool into a scalable primitive for LLM development. Additionally, we open-source code and datasets to facilitate future research.

cs.LG↗
arXiv2026

HD-VGGT: High-Resolution Visual Geometry Transformer

High-resolution imagery is essential for accurate 3D reconstruction, as many geometric details only emerge at fine spatial scales. Recent feed-forward approaches, such as the Visual Geometry Grounded Transformer (VGGT), have demonstrated the ability to infer scene geometry from large collections of images in a single forward pass. However, scaling these models to high-resolution inputs remains challenging: the number of tokens in transformer architectures grows rapidly with both image resolution and the number of views, leading to prohibitive computational and memory costs. Moreover, we observe that visually ambiguous regions, such as repetitive patterns, weak textures, or specular surfaces, often produce unstable feature tokens that degrade geometric inference, especially at higher resolutions. We introduce HD-VGGT, a dual-branch architecture for efficient and robust high-resolution 3D reconstruction. A low-resolution branch predicts a coarse, globally consistent geometry, while a high-resolution branch refines details via a learned feature upsampling module. To handle unstable tokens, we propose Feature Modulation, which suppresses unreliable features early in the transformer. HD-VGGT leverages high-resolution images and supervision without full-resolution transformer costs, achieving state-of-the-art reconstruction quality.

cs.CV↗
arXiv2026

Nonmonotonic Evolution of the Superconducting Transition Temperature and Robust Multigap Extended s-wave + s-wave Pairing in Zn-Substituted FeSe Single Crystals

We report a systematic study of superconductivity on Fe1-xZnxSe single crystals synthesized over a broad Zn doping range (x = 0-0.023). High-quality single crystals across all compositions range exhibit superconducting transitions, while the transition temperature Tc shows a pronounced nonmonotonic dependence on Zn doping concentration, indicating that the underlying mechanism govering Tc its evolution cannot be explained solely by simple impurity pair breaking alone. Magnetization and transport measurements confirm the bulk behavior of superconductivity and reveal enhanced scattering effects with Zn doping. Low-temperature specific heat is consistently described by a two-gap scenario composed of an isotropic s-wave gap and an anisotropic extended s-wave gap, whereas single-gap and alternative pairing symmetries fail to describe the data. The nearly unchanged relative weights of the two gap components suggest the weak interband scattering induced by Zn substitution, thereby preserving multiband superconductivity. These results demonstrate the robustness of multigap superconductivity in FeSe and impose stringent constraints on candidate pairing mechanisms, highlighting the role of multiband electronic structure and anisotropic gap formation.

cond-mat.supr-con↗
arXiv2026

WASD: Locating Critical Neurons as Sufficient Conditions for Explaining and Controlling LLM Behavior

Precise behavioral control of large language models (LLMs) is critical for complex applications. However, existing methods often incur high training costs, lack natural language controllability, or compromise semantic coherence. To bridge this gap, we propose WASD (unWeaving Actionable Sufficient Directives), a novel framework that explains model behavior by identifying sufficient neural conditions for token generation. Our method represents candidate conditions as neuron-activation predicates and iteratively searches for a minimal set that guarantees the current output under input perturbations. Experiments on SST-2 and CounterFact with the Gemma-2-2B model demonstrate that our approach produces explanations that are more stable, accurate, and concise than conventional attribution graphs. Moreover, through a case study on controlling cross-lingual output generation, we validated the practical effectiveness of WASD in controlling model behavior.

cs.CL↗
arXiv2026

Mitigating Entangled Steering in Large Vision-Language Models for Hallucination Reduction

Large Vision-Language Models (LVLMs) have achieved remarkable success across cross-modal tasks but remain hindered by hallucinations, producing textual outputs inconsistent with visual content. Existing methods mitigate hallucinations but often alter generation behavior, resulting in shorter outputs and shifted token distributions, especially in latent space steering approaches. We identify that this issue stems from entangled steering signals, where suppressing hallucinations inadvertently disrupts the model's intrinsic generation behavior. To address this, we propose MESA, an effective plug-and-play framework that performs controlled and selective latent intervention for hallucination mitigation. Specifically, MESA targets hallucination-relevant responses while preserving the model's original token distribution, enabling effective hallucination reduction without compromising generation behavior. Extensive experiments across diverse generative and discriminative benchmarks demonstrate that MESA consistently reduces hallucinations while better preserving generation behavior, outperforming prior methods across multiple LVLM families.

cs.CV↗
arXiv2026

PASK: Toward Intent-Aware Proactive Agents with Long-Term Memory

Proactivity is a core expectation for AGI. Prior work remains largely confined to laboratory settings, leaving a clear gap in real-world proactive agent: depth, complexity, ambiguity, precision and real-time constraints. We study this setting, where useful intervention requires inferring latent needs from ongoing context and grounding actions in evolving user memory under latency and long-horizon constraints. We first propose DD-MM-PAS (Demand Detection, Memory Modeling, Proactive Agent System) as a general paradigm for streaming proactive AI agent. We instantiate this paradigm in Pask, with streaming IntentFlow model for DD, a hybrid memory (workspace, user, global) for long-term MM, PAS infra framework and introduce how these components form a closed loop. We also introduce LatentNeeds-Bench, a real-world benchmark built from user-consented data and refined through thousands of rounds of human editing. Experiments show that IntentFlow matches leading Gemini3-Flash models under latency constraints, while identifying deeper user intent.

cs.AI↗
arXiv2026

Context-Value-Action Architecture for Value-Driven Large Language Model Agents

Large Language Models (LLMs) have shown promise in simulating human behavior, yet existing agents often exhibit behavioral rigidity, a flaw frequently masked by the self-referential bias of current "LLM-as-a-judge" evaluations. By evaluating against empirical ground truth, we reveal a counter-intuitive phenomenon: increasing the intensity of prompt-driven reasoning does not enhance fidelity but rather exacerbates value polarization, collapsing population diversity. To address this, we propose the Context-Value-Action (CVA) architecture, grounded in the Stimulus-Organism-Response (S-O-R) model and Schwartz's Theory of Basic Human Values. Unlike methods relying on self-verification, CVA decouples action generation from cognitive reasoning via a novel Value Verifier trained on authentic human data to explicitly model dynamic value activation. Experiments on CVABench, which comprises over 1.1 million real-world interaction traces, demonstrate that CVA significantly outperforms baselines. Our approach effectively mitigates polarization while offering superior behavioral fidelity and interpretability.

cs.AI↗
arXiv2026

Structural Regularities of Cinema SDR-to-HDR Mapping in a Controlled Mastering Workflow: A Pixel-wise Case Study on ASC StEM2

We present an empirical case study of cinema SDR-to-HDR mapping using ASC StEM2, a rare common-source dataset containing EXR scene-referred images and matched SDR/HDR cinema release masters from the same ACES-based mastering workflow. Based on pixel-wise statistics over all 18,580 frames of the test film, we construct a three-domain comparison involving EXR source data, SDR release masters, and HDR release masters to characterize their luminance and color structural relationships within this controlled workflow. In the luminance dimension, SDR and HDR masters exhibit a highly stable global monotonic correspondence, with geometric structure remaining largely consistent overall; sparse and structured deviations appear in self-luminous highlights and specific material regions. In the color dimension, the two masters remain largely consistent in hue, with saturation exhibiting a redistribution pattern of shadow suppression, midtone expansion, and highlight convergence. Using EXR as a scene-referred anchor, we further define a pixel-level decision map that operationally separates EXR-closer recovery regions from content-adaptive adjustment regions. Under this operational definition, 82.4% of sampled image regions are classified as EXR-closer recovery, while the remainder require localized adaptive adjustment. Rather than claiming a universal law for all cinema mastering pipelines, the study provides an interpretable quantitative baseline for structure-aware SDR-to-HDR analysis and for designing learning-based models under shared-source mastering conditions.

eess.IV↗
arXiv2026

Bootstrap percolation of extension hypergraphs

For $k$-graphs $F$ and $H_0$ the $F$-bootstrap percolation process (or $F$-process) starting with $H_0$ is a sequence $(H_i)_{i\geq0}$ of $k$-graphs such that $H_{i+1}$ is obtained from $H_i$ by adding all those $e\in V(H_0)^{(k)}\setminus E(H_i)$ as edges that complete a new copy of $F$. The running time of this $F$-process, denoted by $M_F(H_0)$, is the smallest $i$ with $H_i=H_{i+1}$. Bollobás proposed the problem of determining the maximum running time for $n\in\mathbb{N}$, i.e., $$M_F(n)=\max_{\vert V(H_0)\vert=n}M_F(H_0)\,.$$ Recently, Noel and Ranganathan initiated the study of this quantity for $k$-graphs. In this work, we determine the asymptotics of $M_F(n)$ for a large class of $k$-graphs. Given a graph $G=(V,E)$, the $k$-extension of $G$ is a $k$-graph $F^{(k)}(G)$ obtained from $G$ by enlarging each edge with a $(k-2)$-set of new vertices. We show that for every graph $G$ on $t$ vertices and every $k\geq 3$, $M_{F^{(k)}(G)}(n)\leq C_{k,t}$ for some constant $C_{k,t}$ depending only on $t$ and $k$.

math.CO↗
arXiv2026

Topological defect induced phase separation in a holographic system

We investigate the coupled dynamics of symmetry breaking and phase separation during quenches across the critical point in a first-order phase transition. Based on the Einstein-Maxwell-scalar theory, we construct a holographic superfluid model with $\mathbb{Z}_2$ symmetry. By introducing higher-order nonlinear terms $λΨ^4$ and $τΨ^6$ into the scalar field potential, we realize a rich phase structure, which enables us to study the coupling effects between symmetry breaking and phase separation. Furthermore, by preparing initial conditions with well-defined spatial partitions, we discover a new triggering mechanism for the invasion phenomenon, namely that kinks serve as triggering sites for the phase separation process. This study reveals a novel coupling mechanism between topological defects and phase separation, enriches our understanding of nonequilibrium structure formation in strongly coupled systems.

hep-th↗
arXiv2026

Stability analysis and double critical phenomenon in the Einstein-Maxwell-scalar theory

We investigate the dynamical stability and phase transition behavior in a holographic superfluid model incorporating higher-order self-interaction terms $λ|ψ|^4$, $τ|ψ|^6$, and a non-minimal coupling $h(ψ)=e^{α|ψ|^2}$. Thermodynamic and dynamical stability analyzes show that the thermodynamic stability and dynamical stability of the system are consistent. Phase diagram analysis reveals rich critical and supercritical phenomena. For fixed $λ<0$ and $α$, increasing $τ$ shrinks the first-order phase transition region to a critical point and then enters the supercritical region. When varying $α$, the system can exhibit no critical point and, most notably, a double critical phenomenon in which, as $α$ increases, the system first enters the supercritical region and then re-enters the first-order phase transition region. This double critical phenomenon driven by a single parameter is reported for the first time in holographic superfluid models, revealing a complex nonmonotonic coupling effect between the non-minimal coupling and higher-order interaction terms.

gr-qc↗
arXiv2026

Between proper and square colorings of planar graphs with maximum degree at most four

An $i$-independent set is a vertex set whose pairwise distance is at least $i+1$. A proper (square) $k$-coloring of a graph $G$ is a partition of its vertex set into $k$ independent ($2$-independent) sets. A packing $(1^{j}, 2^k)$-coloring of a graph $G$ is a partition of $V(G)$ into $j$ independent sets and $k$ $2$-independent sets. It can be viewed as intermediate colorings between proper and square coloring. Wegner conjectured in 1977 that every planar graph with maximum degree at most four is square $9$-colorable. Bousquet, Deschamps, de Meyer, and Pierron proved an upper bound of $12$, which is the current best result toward the conjecture of Wegner. In this paper, we prove two analogue results that every planar graph with maximum degree at most four is packing $(1,2^{10})$-colorable and packing $(1^2,2^7)$-colorable.

math.CO↗
arXiv2026

Simultaneous determination of Hubble constant and cosmic baryon density: Forecasts for the synergy between FRBs and emerging probes

Two of the most pressing challenges in cosmology are the persistent discrepancy in measurements of the Hubble constant, referred to as the Hubble tension, and the deficit of baryons in the local Universe, known as the missing baryon problem. Fast radio bursts (FRBs) provide a unique probe of both the Hubble constant $H_0$ and the cosmic baryon density $Ω_{\rm b}$. However, constraints from FRBs alone suffer from a severe $H_0$-$Ω_{\rm b}$ degeneracy that prevents them from resolving either problem. We show that this degeneracy can be broken by combining FRBs with other emerging probes whose degeneracy directions differ in the $H_0$-$Ω_{\rm b}$ plane. Specifically, we quantify three multi-messenger approaches: FRBs paired with gravitational wave (GW) standard sirens, strong gravitational lensing (SGL) time delays, and 21 cm intensity mapping (IM) surveys. The combinations FRB+GW, FRB+SGL, and FRB+21 cm IM each deliver simultaneous constraints on $H_0$ and $Ω_{\rm b}$ better than (1%, 1.5%) in the $Λ$CDM model, and when dynamical dark energy is introduced, the constraining precision degrades gracefully as model complexity increases. Furthermore, within a model-independent framework, both FRB+GW and FRB+SGL constrain $H_0$ and $Ω_{\rm b}$ to precisions better than (1.5%, 3%). These precision levels are based on nominal observational expectations and would improve significantly under optimistic observational scenarios.

astro-ph.CO↗
arXiv2026

Dark radiation from Kerr primordial black holes: the role of superradiance

Light primordial black holes (PBHs) that fully evaporate before Big Bang Nucleosynthesis (BBN) produce dark radiation (DR) via Hawking radiation of gravitons, contributing to the effective number of relativistic species $ΔN_{\rm eff}$. If the particle spectrum contains a beyond-the-Standard-Model (BSM) boson with Compton wavelength comparable to the black hole (BH) gravitational radius, superradiant instability extracts angular momentum from the BH into a bosonic cloud, whose gravitational wave (GW) emission contributes an additional source of DR. By simultaneously evolving the BH mass and spin, superradiant mode occupation numbers, comoving entropy and cosmological energy densities in an expanding early-universe background, we find that superradiance generically suppresses $ΔN_{\rm eff}$: by extracting angular momentum before Hawking radiation can convert it into gravitons, superradiance starves the dominant dark-radiation channel. The GWs emitted by the superradiant cloud can partially compensate this loss, but only when the superradiant and BH evaporation timescales are comparable; otherwise the cloud GWs are emitted too early and diluted by cosmological expansion. The results imply that existing $ΔN_{\rm eff}$ bounds on PBH mass and spin derived without superradiance must be revisited if BSM bosons are present in the particle spectrum.

astro-ph.CO↗
arXiv2026

Interior structure of black holes with nonlinear terms

We investigate the oscillation of the Kasner exponent $p_t$ near critical point of the hairy black holes dual to holographic superfluid and reveal a clear inverse periodicity $f(T_c/(T_c-T))$ in a large region below the critical temperature. We first introduce the fourth-power term with a coefficient $λ$ to adjust the oscillatory behavior of the Kasner exponent $p_t$ near the critical point. Importantly, we show that the nonlinear coefficient $λ$ provides accurate control of this periodicity: a positive $λ$ stretches the region, while a negative $λ$ compresses it. By contrast, the influence of another coefficient $τ$ is more concentrated in regions away from the critical point. This work provides a new perspective for understanding the complex dynamical structure inside black holes and extends the actively control from the fourth- and sixth-power term into the black hole interior region.

gr-qc↗
arXiv2026

Scaling Limits for Exponential Hedging in Trinomial Models

We study scaled trinomial models converging to the Black--Scholes model, and analyze exponential certainty-equivalent prices for path-dependent European options. As the number of trading dates $n$ tends to infinity and the risk aversion is scaled as $nl$ for a fixed constant $l>0$, we derive a nontrivial scaling limit. Our analysis is purely probabilistic. Using a duality argument for the certainty equivalent, together with martingale and weak-convergence techniques, we show that the limiting problem takes the form of a volatility control problem with a specific penalty. For European options with Markovian payoffs, we analyze the optimal control problem and show that the corresponding delta-hedging strategy is asymptotically optimal for the primal problem.

q-fin.MF↗