Search arXiv⌕ Search

arXiv subjects

Yi Zhou

Publications and source records attributed to Yi Zhou.

At least 55 records · Page 3Linked to original sources
arXiv2026

Global Existence of classical solutions to 3D nonlinear Klein-Gordon equations with low-regularity initial data

This paper studies global existence for the Cauchy problem of nonlinear Klein-Gordon equations in three space dimensions, strictly within the regularity regime of classical local existence. We prove it via higher-order and lower-order energy estimates. The proof relies on two key ingredients. The first is due to the work of Georgiev and Popivanov, which reduces quadratic nonlinearities to cubic terms plus ghost-energy-controllable terms, securing lower-order estimates. The second is a sharp Klainerman-Sobolev-type inequality without the scaling operator, established herein, which yields enough time decay for derivatives up to second order; integration by parts then controls derivative loss terms in the higher-order estimates.

math.AP↗
arXiv2026

Localize-Then-Decide Guarantees for LLM Judgments

Large language models (LLMs) are increasingly used as evaluators to assess output quality and preference alignment, yet providing reliable guarantees of agreement with human judgments remains challenging. Recent work introduces confidence-thresholding methods that provide such guarantees for pairwise comparisons, relying on the assumption that higher estimated confidence implies lower disagreement risk with humans. However, this assumption can break down when the number of candidate responses increases, since distributing probability mass across many alternatives can distort confidence estimates. To address this issue, we propose a Localize-Then-Decide framework. First, conformal prediction localizes a small shortlist that contains the human-preferred response with high probability. Then, a calibrated confidence-based rule selectively chooses a single response from this shortlist or abstains. This design restores the monotonic relationship between confidence and disagreement risk and enables high-probability agreement guarantees. Experiments with multiple candidate sizes across several datasets and judge LLMs demonstrate that our framework consistently achieves higher guarantee success rates and substantially higher coverage than single-stage baselines.

cs.CL↗
arXiv2026

Unlocking Multimodal Protein Language Models at Inference Time

Multimodal protein language models (pLMs) learn joint protein sequence-structure distributions, and their generation performance should also depend critically on inference-time sampling strategies. Yet prior work has focused more on model training than on how inference-time strategies behave. In this paper, we establish a three-stage investigation framework to empirically study the inference design space of multimodal pLMs across three representative pLMs and four fundamental tasks. We evaluate vanilla sampling, task-specific classifier-free guidance, and reward-guided beam search on multimodal pLMs, corresponding to controls over sampling distributions, per-step logits, and parallel trajectories. Throughout the complementary advancements centered on exploration-exploitation trade-off, we (1) reveal the suboptimality of default inference protocols and identify task-oriented sampling preferences; (2) observe substantial quantitative gains across tasks, consistently boosting the upper bound performance of multimodal pLMs without updating model parameters; (3) derive conclusions about base models that differ from prior consensus.

cs.CE↗
arXiv2026

Unit-to-Plant Stability Shaping of Multi-Electrolyzer ReP2H Plants via Interface Design and Dispatch

Alkaline water electrolysis (AWE) units supplied by insulated gate bipolar transistor rectifiers (IGBT-Rs) may experience oscil-lations caused by coupling between rectifier control and electro-lyzer (ELZ) dynamics. Because this risk varies with unit loading and power allocation, production-oriented dispatch may place a multi-ELZ renewable power-to-hydrogen (ReP2H) plant near or exceed its stability boundary. This paper proposes a stability-oriented framework for control design and plant production dis-patch. A three-port admittance model links the ac port, dc link, and electrolysis stack. Unit-level dc-port analysis quantifies the effects of loading, temperature, Buck bandwidth, and dc-link capacitance, while plant-level aggregation evaluates how unit commitment and power allocation affect stability. Results show that higher loading reduces stability, whereas larger dc-link ca-pacitance and higher Buck bandwidth improve it. Under the same plant loading, different power allocations result in different plant-level stability margins, with balanced allocation generally providing a larger margin than concentrated allocation. The plant-level model thus distinguishes the stability margins of ad-missible schedules. Hardware-in-the-loop (HIL) tests validate these trends and the proposed redistribution rule. The resulting operating regions and dispatch rules can be used to screen unit commitment and power allocation decisions in plant production scheduling.

math.OC↗
arXiv2026

Density Spectra of Topological Groups

We study density spectra associated with dense and closed subgroups of topological groups. For a topological group $G$, let $\dd^*(G)$ denote the set of densities of its dense subgroups. We prove that every compact group satisfies $\dd^*(G)=[d(G),w(G)]$. We also investigate the density spectrum $\cd(G)$ of infinite closed subgroups. In $\mathbf{ZFC}$, we construct a separable countably compact Boolean group containing a closed non-separable subgroup, answering a question of Leiderman, Morris, and Tkachenko. We further show that a free pro-$p$ group of uncountable rank has no non-trivial metrizable closed normal subgroup, giving a negative answer to a problem of Hernández, Hofmann, and Morris. Finally, we study when $\cd(G)$ is an interval. Among other results, we show that under Shelah's Strong Hypothesis ($\mathbf{SSH}$), the closed density spectrum of every infinite $ω$-bounded group of countable tightness is an interval of cardinals.

math.GR↗
arXiv2026

OrthoSkillVLA: Continual Skill Learning via Gradient-Informed Skill Subspace Adaptation

Pretrained Vision-Language-Action models provide a strong foundation for robot learning, but sequentially adapting them to diverse skills can perturb the representations and velocity mappings used by previous skills, leading to catastrophic forgetting. Architecture-based approaches improve retention by isolating skills but lead to increased inference footprint. Recent subspace-constrained methods restrict parameter updates in an orthogonal subspace to minimize interference but impose a unified constraint on the entire model. We analyze the distinct roles of internal VLA components and identify two VLA-specific challenges. First, the VLM maintains broad semantic representations, making it vulnerable to capacity exhaustion, whereas the ActionHead refines semantics into localized velocity patterns that are highly sensitive to perturbations. Second, the final velocity decoder serves as a readout layer. Freezing it forms an output-stage expressivity bottleneck, while updating it risks overwriting previous velocity mappings. To this end, we propose OrthoSkillVLA, a parameter-efficient framework for continual skill learning in pretrained VLA models without demonstration replay. Given the representation heterogeneity, we impose separate subspace constraints on the VLM and ActionHead, preserving reusable semantic capacity while protecting localized velocity patterns. For the output layer, we introduce a lightweight feature-aware MoE decoder, where each skill is allocated a compact expert and a training-free router selects the expert according to feature-space affinity. Extensive simulated and real-world evaluations, together with ablations, demonstrate that OrthoSkillVLA better preserves prior skills while acquiring new ones.

cs.RO↗
arXiv2026

Global solutions of compressible Navier-Stokes equations with small viscosity

In this paper, we study the Cauchy problem for the compressible Navier-Stokes system in $\mathbb{R}^3$. Suppose that the viscosity coefficients satisfy $0<\max\{μ, ν=λ+2μ\}<1$, and set $\varepsilon=\min\{μ, ν=λ+2μ\}$. We establish the global existence of classical solutions when the initial perturbations of the density and the curl-free part of the velocity are smaller than $\varepsilon^{\frac12+}$ (up to a logarithmic loss), while the divergence part of the initial velocity is smaller than $\varepsilon$. This improves the classical global existence result of Matsumura-Nishida \cite{MaN80}, which requires all the initial data to be smaller than $\varepsilon (<1)$. We expect that this result is representative of general Shizuta-Kawashima systems arising in physical applications. The improvement of the index from $1$ to $\frac12+$ relies on exploiting the hidden Kawashima-type dissipation for the density and controlling the spacetime trace norm of the solution at the scale $\sqrt{\varepsilon}$. These two ingredients are obtained through a weighted trace inequality and a Morawetz-type inequality for the perturbed sound speed and the divergence of the velocity.

math.AP↗
arXiv2026

Multi-Agent Forensic Reasoning for Generalizable Deepfake Video Detection

The malicious use of generative artificial intelligence to create highly realistic deepfake videos raises serious ethical concerns and poses substantial challenges to AI safety. However, existing deepfake video benchmarks provide limited coverage of recent synthesis methods and generally lack reliable fine-grained textual annotations. Meanwhile, conventional detectors and multimodal large language models (MLLMs), whether operating as a single model or relying on a single analytical perspective, often fail to capture subtle forgery artifacts, limiting their generalization to emerging AI-generated methods. To address these limitations, we introduce FaceVid-Forensics-100K, a large-scale deepfake video dataset comprising 100,000 videos and spanning 33 synthesis methods across face swapping, face reenactment, and entire-face synthesis, including recent generators such as Seedance 2.0. The dataset provides fine-grained textual annotations of visual observations and verdict-consistent forensic explanations, automatically synthesized through a multi-model aggregation and conflict-resolution pipeline powered by advanced MLLMs. Building on this benchmark, we propose a multi-agent forensic reasoning framework that employs four specialized domain-expert agents to independently analyze forgery cues from four perspectives: texture, lighting, motion, and physics. A judge agent then reconciles their reports to produce a final prediction together with an explanation. Extensive evaluations on out-of-domain test sets show that, despite being composed entirely of small open-source MLLMs, our framework outperforms all methods including closed-source GPT and Gemini models and ranks first across all reported metrics on this benchmark. The project page is available at https://xavierjiezou.github.io/ARGUS/.

cs.CV↗
arXiv2026

CoCo: Code as CoT for Text-to-Image Preview and Rare Concept Generation

Recent advancements in Unified Multimodal Models (UMMs) have significantly advanced text-to-image (T2I) generation, particularly through the integration of Chain-of-Thought (CoT) reasoning. However, existing CoT-based T2I methods largely rely on abstract natural-language planning, which lacks the precision required for complex spatial layouts, structured visual elements, and dense textual content. In this work, we propose CoCo (Code-as-CoT), a code-driven reasoning framework that represents the reasoning process as executable code, enabling explicit and verifiable intermediate planning for image generation. Given a text prompt, CoCo first generates executable code that specifies the structural layout of the scene. The code is then executed in a sandboxed environment to render a deterministic draft image. Subsequently, the model refines this draft through fine-grained image editing to produce the final high-fidelity result. To support this training paradigm, we construct CoCo-10K, a curated dataset containing structured draft-final image pairs designed to teach both structured draft construction and corrective visual refinement. Empirical evaluations on StructT2IBench, OneIG-Bench, and LongText-Bench show that CoCo achieves improvements of 68.83%, 54.8%, and 41.23%, respectively, over direct generation, while also outperforming other CoT-enhanced generation methods. These results demonstrate that executable code is an effective and reliable reasoning paradigm for precise, controllable, and structured text-to-image generation.

cs.AI↗
arXiv2026

PD-GS: Phoneme-Driven 3DGS for Audio-Driven Talking Heads

3D Gaussian Splatting (3DGS) enables fast, photorealistic talking-head rendering, yet accurate lip articulation remains elusive: mouth motion is often over-smoothed and may violate hard articulatory constraints such as bilabial closures, producing the notorious ``leaky mouth'' artifact. A key difficulty is that brief, discrete articulatory events are inferred from a continuous acoustic embedding under a regression objective, which biases predictions toward averaged mouth configurations. While modern self-supervised speech encoders provide rich prosodic and phonetic cues, they do not provide an explicit, frame-aligned linguistic target that reliably disambiguates closure-level events. We propose \textbf{Phoneme-Driven Gaussian Splatting (PD-GS)}, which augments a 3DGS talker with time-aligned phoneme tokens obtained from an automatic ASR and forced-alignment pipeline. Our core component, the \textbf{Linguistic Fusion Module (LFM)}, adaptively fuses continuous audio context with discrete phoneme embeddings through a learned gate, allowing the model to preserve smooth audio-driven dynamics while strengthening phoneme guidance on articulation-critical segments. PD-GS is trained purely from monocular video using image reconstruction and lip landmark supervision. On HDTF, PD-GS achieves the best lip geometry among the compared baselines (LMD 2.66) and qualitatively reduces closure violations in challenging phoneme sequences, yielding more linguistically faithful neural avatars.

cs.AI↗
arXiv2026

Exploring the Optimal Size of Grid-forming Energy Storage in an Off-grid Renewable P2H System under Multi-timescale Energy Management

Utility-scale off-grid renewable power-to-hydrogen systems (OReP2HS), typically comprising photovoltaic plants, wind turbines, electrolyzers (ELs), and battery energy storage system (BESS), requires at least one component, generally the BESS, working with grid-forming ability to provide frequency and voltage references and regulate them through transient power support. However, existing designs of OReP2HS based on the energy management strategies (EMSs) with 5-min or hourly resolution fail to capture fast transients and may underestimate the BESS size required to ensure adequate grid-forming ability. This paper first proposes a framework of multi-timescale EMS covers from those for power system transient behaviors to second-level EL load adjustments and minute-level intra-day scheduling to coordinate renewable power, BESS, and ELs. Then, an iterative search procedure based on high-fidelity simulation is employed to determine the cost-effective BESS size that satisfies grid-forming, long-term energy balancing over 8760 hours, and emergency support requirements. Case studies based on a planned OReP2HS project in Inner Mongolia, China, show that the proposed EMS yields a base-case LCOH of 33.212 CNY/kg, with CAPEX of BESS accounting for 17.83% of total investment. The optimal BESS capacity represents 13.6% of the rated hourly renewable output and shows a yearly degradation of 4.87%. Sensitivity analysis reveals that reducing the electrolytic load adjustment time step from 90 to 5 seconds and increasing its ramping limit from 1% to 10% rated power per second, the BESS size decreases by 53.57%, and the LCOH decreases to 25.458 CNY/kg. Considering the cost of designing and manufacturing utility-scale ELs with fast load regulation capability, a load adjustment time step of 5 to 10 seconds and a ramping limit of 4-6% rated power per second are recommended.

math.OC↗
arXiv2026

Dynamic Modeling and Control of Multi-Stack Alkaline Water Electrolysis Systems with Shared Gas Separators and Lye Circulation:Industrial Data-Based Validation and Simulation

An emerging approach for large-scale renewable hydrogen production is integrating multiple alkaline water electrolysis (AWE) stacks into one balance-of-plant (BoP) system, sharing gas-lye separation and lye circulation components. While this configuration, termed $N$-in-1, reduces cost and complexity, its dynamic performance under fluctuating power remains unclear compared with conventional 1-in-1 systems. This paper develops a state-space model of the multi-stack AWE system, capturing lye circulation, temperature, and hydrogen-to-oxygen (HTO) dynamics, calibrated via experiments on a 4,000 Nm$^3$/h-rated 4-in-1 system. A mixed-integer quadratic programming (MIQP)-based predictive controller is then designed to coordinate inter-stack current distribution, lye flow, and cooling for load tracking and operational stability. Simulations on the experimentally validated model show that a $4$-in-1 system achieves similar performance compared to four parallel 1-in-1 systems under continuous operation. Differences in load-tracking, temperature stabilization errors, and specific energy consumption remain below 0.015 MW, 0.346 K, and 0.001 kWh/Nm$^3$ under wind power supply when all stacks remain online.

math.OC↗
arXiv2026

Harmonic Cancellation in Multi-Electrolyzer P2H Plants via Phasor-Modulated Production Scheduling

Thyristor rectifiers (TRs) are cost-effective power supplies for hydrogen electrolyzers (ELZs) but introduce harmonic distortion that may violate grid codes. This letter proposes a self-governing harmonic mitigation strategy through coordinated operation of multiple ELZs in large power-to-hydrogen (P2H) plants. First, the harmonic model of TR-powered ELZs is derived, revealing a natural harmonic cancellation mechanism among them. Based on this, a system-level operation scheme based on phasor modulation is developed and integrated into plant scheduling. Case studies demonstrate that the proposed method reduces harmonic currents by 21.2%-39.7% and ensures grid-code compliance, with only a 0.25% loss in hydrogen output, while increasing total revenue by over 21% compared to production-oriented strategies.

math.OC↗
arXiv2026

Topological Defects in Triple-$Q$ Magnetic Orders: A Fixed-Lattice Homotopy Classification

Multiple-$Q$ magnetic orders combine continuous spin rotations with discrete crystalline sectors associated with translations and point-group transformations, producing a richer defect structure than conventional single-$Q$ magnets. We classify the bulk defects of all seven stable phases for $N=2$ and $3$ in the $M$-point triple-$Q$ Ginzburg--Landau theory with $(\Vfour\rtimes\Dthree)\times\OO(N)$ symmetry, where $\Vfour$ is the translation-generated Klein four-group. The atomic lattice is treated as a prescribed background, with lattice dislocations and disclinations excluded and the three Fourier fields retaining their physical $M$-point labels. The parent-group transformations continuously connected to the identity form $G_0=\{e\}\times\SO(N)$. For a reference-state stabilizer $H$, the connected component containing the reference state is $G_0/(H\cap G_0)$, not the quotient obtained by projecting $H$ onto spin space. This distinction gives the orthogonal triple-$Q$ phase the full manifold $\OO(3)$, with chirality walls and Abelian $\ZZ_2$ frame vortices rather than non-Abelian binary-polyhedral vortices. Every connected component of the $\OO(2)$ phases supports an integer $2π$ vortex, whereas fractional windings close only when attached to a discrete-domain wall and are linearly confined at nonzero wall tension. Translation symmetry further forbids cross-gradient bilinears, reducing the quadratic elastic sector to an isotropic and an $M$-point-locked anisotropic stiffness. The classification separates free internal defects, crystalline domain walls, and wall-bound composites in triple-$Q$ magnets.

cond-mat.str-el↗
arXiv2026

Bogoliubov sum rule and the Knight-shift ellipsoid in spin-locked superconductors

We establish an exact Bogoliubov sum rule for any Hermitian single-particle operator $O$: at each momentum, its particle-hole and particle-particle matrix-element weights sum to the single-particle trace $\mathrm{Tr}_{s}(O^{2})$. The result follows solely from Hilbert--Schmidt-norm invariance under a canonical Bogoliubov transformation and contains neither excitation-energy denominators nor occupations; physical response identities therefore require additional assumptions. At zero field, a fully gapped helicity-diagonal state with both helicity sheets present and a spin-orbit splitting asymptotically larger than the gap obeys $χ_{μν}(0)/χ_N=δ_{μν}-Π_{μν}+o(1)$, where $χ_N$ is the normal-state Pauli susceptibility and $Π=\langle\hat{\mathbf n}_{\mathbf k}\hat{\mathbf n}_{\mathbf k}\rangle_{\rm FS}$ is the Fermi-surface average of the unit spin-locking texture. The eigenvalues of $Π$ form a simplex, while the normalized spin Knight-shift tensor defines an ellipsoid whose semi-axes are the residual principal responses. Full cubic invariance of both the superconducting state and locking texture fixes $Π=\mathbb I/3$ and hence $χ(0)/χ_N=2\mathbb I/3+o(1)$; cubic crystal symmetry alone does not. For zero-field $s$-wave pairing in the reference-Fermi-surface regime, we obtain the exact closed-form kernel $F_s(λ)=1-\operatorname{asinh}λ/[λ\sqrt{1+λ^2}]$, valid for arbitrary $λ=|\mathbf g|/Δ$. In a finite Zeeman field, the zero-field helicity reduction generally fails, so the equilibrium magnetization and differential response require a self-consistent BdG calculation rather than a field-dependent locking-tensor substitution. Applied to the $^{75}$As data on K$_2$Cr$_3$As$_3$, the framework identifies a field-dependent axial suppression pattern at $8$--$16$~T.

cond-mat.supr-con↗
arXiv2026

Who Wins Where? Conformal Model Comparison for Local Superiority

Standard model comparison is global, aggregating losses across the covariate space to declare a single winner. This can obscure heterogeneous performance, where different models are preferable in different regions. We introduce conformalized local model comparison, a split-sample framework for constructing calibrated local best-model maps. Given a model comparison score, such as the difference between two squared losses, the method uses three disjoint splits to fit competing models, estimate local centers and scales from out-of-sample scores, and conformally calibrate residual uncertainty. At a target point, the procedure declares a local winner only when a one-sided conformal bound excludes a tie, with the score's sign determining the favored model. We prove finite-sample marginal control for one-sided erroneous declarations on the realized future comparison score, establish pointwise consistency of the localized mean-score estimator away from tie boundaries, show that aggregate comparison can disagree sharply with the prevalence of local superiority, and derive a squared-loss bias--variance decomposition that clarifies how model structure affects local wins. Synthetic and real-data experiments show that the method recovers heterogeneous winner regions, abstains under uncertainty, and yields higher conditional gain than global selection.

cs.LG↗
arXiv2026

Outage Analysis of Backscatter-Based Ambient IoT Device Classes with Energy Buffering

This paper presents an analytical framework for evaluating the outage probability of ambient Internet of Things (A-IoT) device classes communicating directly with a base station. Device 1 is a passive backscatter device with minimal storage, while Device 2 is equipped with a supercapacitor that enables energy buffering and optional amplification. The proposed framework jointly accounts for carrier-detection sensitivity, energy harvesting constraints, supercapacitor energy dynamics, and an energy-aware amplification policy, while the energy evolution of buffered devices is modeled using a discrete-time Markov chain (DTMC). The results highlight the interplay between energy availability and communication reliability. Device 2 achieves superior performance in energy-rich regimes due to buffering and amplification gains, whereas Device 1 becomes more robust in energy-constrained regimes, particularly at larger distances or under high payload requirements. These findings highlight that the optimal device choice depends critically on the operating regime and application demands.

eess.SP↗
arXiv2026

Pokrovsky--Talapov and Berezinskii--Kosterlitz--Thouless Phase Transitions in Bilayer Superconducting Films under an In-Plane Magnetic Field

We study finite-temperature phase transitions in a Josephson-coupled bilayer superconducting film with compact layer phases under an in-plane magnetic field. At zero temperature, where thermally excited layer vortices are absent, the relative-phase sector undergoes a Pokrovsky--Talapov (PT) commensurate--incommensurate (C--IC) transition from a commensurate Fulde--Ferrell (C/FF) state to an incommensurate Bloch superconducting (IC/Bloch SC) state. At finite temperature, compactness separates two distinct defect mechanisms. The C--IC boundary remains a PT soliton-entry line: interlayer Josephson vortex--antivortex-pair solitons enter with the square-root onset $ρ_{\rm sol}\propto [k_0-k_0^c(T)]^{1/2}$. Thermal melting is instead Berezinskii--Kosterlitz--Thouless (BKT)-like, with correlation exponent $η=1/4$ at the boundary, but the active vortex channel changes across the phase diagram. Josephson locking suppresses elementary layer vortices in the C/FF state and selects a same-vorticity layer-pair BKT channel, whereas elementary layer vortices control melting of the IC/Bloch SC state.

cond-mat.supr-con↗