arXiv subjects

Yi Zhou

Publications and source records attributed to Yi Zhou.

At least 91 records · Page 5Linked to original sources

arXiv2026

Seeing What Matters: Visual Preference Policy Optimization for Visual Generation

Reinforcement learning (RL) has become a powerful tool for post-training visual generative models, with Group Relative Policy Optimization (GRPO) increasingly used to align generators with human preferences. However, existing GRPO pipelines rely on a single scalar reward per sample, treating each image or video as a holistic entity and ignoring the rich spatial and temporal structure of visual content. This coarse supervision hinders the correction of localized artifacts and the modeling of fine-grained perceptual cues. We introduce Visual Preference Policy Optimization (ViPO), a GRPO variant that lifts scalar feedback into structured, pixel-level advantages. ViPO employs a Perceptual Structuring Module that uses pretrained vision backbones to construct spatially and temporally aware advantage maps, redistributing optimization pressure toward perceptually important regions while preserving the stability of standard GRPO. Across both image and video benchmarks, ViPO consistently outperforms vanilla GRPO, improving in-domain alignment with human-preference rewards and enhancing generalization on out-of-domain evaluations. The method is architecture-agnostic, lightweight, and fully compatible with existing GRPO training pipelines, providing a more expressive and informative learning signal for visual generation.

cs.CV↗

arXiv2026

Dual-axis attribution of zebrafish tectal microcircuits for energy-efficient and robust neurocomputing

Biological neural circuits contain specialized substructures that support distinct computational functions, yet many bio-inspired neural networks borrow biological motifs without identifying their circuit-level origins. In this study, we investigate whether zebrafish tectal microcircuits can be attributed along two computational axes: energy-efficient information processing and robustness-preserving stabilization. We reconstruct a directed zebrafish-inspired retinotectal microcircuit graph and verify retinotectal signal propagation through dynamic simulation. A leaky integrate-and-fire spiking neural network is then used as a nonlinear perturbation testbed, where predefined subcircuits are selectively ablated and evaluated using the Energy Sensitivity Index and the Robustness Sensitivity Index.The results reveal a functional dissociation between two tectal subcircuits.The \textit{ns\_TIN} subcircuit shows a low spike footprint but a measurable influence on prediction error, suggesting a role as a spike-efficient internal information gate.In contrast, the \textit{superficial\_TIN} subcircuit produces the highest robustness sensitivity, suggesting a feedback-like role in maintaining system-level stability.We further transfer these attributed functions into ResNet18-based artificial neural networks and evaluate them on CIFAR-10 under inference-budget reduction and Gaussian noise corruption. The \textit{ns\_TIN}-inspired module improves performance preservation under reduced computation, whereas the \textit{superficial\_TIN}-inspired module improves robustness under input noise. These findings provide a subcircuit-level route for linking biological circuit organization with bio-inspired neural architecture design.

cs.NE↗

arXiv2026

Coarse-to-Real: Generative Rendering for Populated Dynamic Scenes

Traditional rendering pipelines rely on complex assets, accurate materials and lighting, and substantial computational resources to produce realistic imagery, yet they still face challenges in scalability and realism for populated dynamic scenes. We present C2R (Coarse-to-Real), a generative rendering framework that synthesizes real-style urban crowd videos from coarse 3D simulations. Our approach uses coarse 3D renderings to explicitly control scene layout, camera motion, and human trajectories, while a learned neural renderer generates realistic appearance, lighting, and fine-scale dynamics guided by text prompts. To overcome the lack of paired training data between coarse simulations and real videos, we adopt a two-stage synthetic-real domain-hedging strategy that first learns a strong generative prior from large-scale real footage, and then introduces controllability by using a small amount of paired synthetic coarse-to-fine data to anchor shared implicit spatio-temporal features across domains. The resulting system supports coarse-to-fine control, generalizes across diverse CG and game inputs, and produces temporally consistent, controllable, and realistic urban scene videos from minimal 3D input. We will release the model and project webpage at https://gonzalognogales.github.io/coarse2real/.

cs.CV↗

arXiv2026

Apparent double-$T_c$ from a single BKT transition in anisotropic phase-only models

Transport experiments on two-dimensional superconductors often yield direction-dependent transition temperatures, raising the question of whether such a ``double-$T_c$'' reflects a true thermodynamic splitting or a transport artifact. To establish a baseline, we study a minimal anisotropic phase-only Josephson-junction array in equilibrium and under resistively shunted junction dynamics with fluctuating twist boundary conditions. The equilibrium model exhibits a single Berezinskii--Kosterlitz--Thouless (BKT) transition. Out of equilibrium, anisotropic Josephson couplings and anisotropic dissipation reshape the linear $R$--$T$ curves in a finite-size, finite-current crossover regime, so that curve-shape criteria such as Halperin--Nelson fits and fixed-resistance thresholds yield an apparent double-$T_c$. In contrast, critical-scaling criteria -- the universal exponent $α=3$ and dynamic finite-size scaling -- remain consistent with the single $T_{\mathrm{BKT}}$. A robust splitting that persists in the nonlinear critical scaling, such as that recently reported at KTaO$_3$ interfaces, therefore points to physics beyond this clean anisotropic baseline.

cond-mat.supr-con↗

arXiv2026

How Well Do LLMs Perform on the Simplest Long-Chain Reasoning Tasks: An Empirical Study on the Equivalence Class Problem

Large Language Models (LLMs) have achieved great improvements in recent years. Nevertheless, it still remains unclear how good LLMs are for reasoning tasks, especially for long-chain ones. In this paper, we evaluate LLMs' performance on the simplest yet long-chain reasoning task, namely the Equivalence Class Problem (ECP), i.e., determining whether two variables are equal given a set of randomly generated equivalence relations. We consider both reasoning and non-reasoning representative LLMs over a large variety of problem instances, ranging over different numbers of variables, connectivity probabilities, prompts, and other factors. The experimental results show that non-reasoning LLMs fail ECP, while reasoning models are significantly better but still struggle to completely solve this problem. Interestingly, considering various connectivity probabilities with a fixed number of variables, we observe that, for non-reasoning models, the hardest problem instances coincide with the phase transition point of ln n/(n-1), suggesting the chaos of the problem; in contrast, for reasoning models, the hardest ones coincide with the biggest diameter, suggesting the reasoning difficulty of the problem.

cs.AI↗

arXiv2026

SemEval-2026 Task 7: Everyday Knowledge Across Diverse Languages and Cultures

We present our shared task on evaluating the adaptability of LLMs and NLP systems across multiple languages and cultures. The task data consist of an extended version of our manually constructed BLEnD benchmark (Myung et al. 2024), covering more than 30 language-culture pairs, predominantly representing low-resource languages spoken across multiple continents. As the task is designed strictly for evaluation, participants were not permitted to use the data for training, fine-tuning, few-shot learning, or any other form of model modification. Our task includes two tracks: (a) Short-Answer Questions (SAQ) and (b) Multiple-Choice Questions (MCQ). Participants were required to predict labels and were allowed to submit any NLP system and adopt diverse modelling strategies, provided that the benchmark was used solely for evaluation. The task attracted more than 140 registered participants, and we received final submissions from 62 teams, along with 19 system description papers. We report the results and present an analysis of the best-performing systems and the most commonly adopted approaches. Furthermore, we discuss shared insights into open questions and challenges related to evaluation, misalignment, and methodological perspectives on model behaviour in low-resource languages and for under-represented cultures.

cs.CL↗

arXiv2026

Geometric structure of singular free boundary points for the logarithmic obstacle problem

In the previous work [Interfaces Free Bound., 19, 351--369, 2017], de Queiroz and Shahgholian established the optimal $C^{1,\log}_{\mathrm{loc}}$ regularity of solutions for the obstacle problem with singular logarithmic forcing term $$-Δu = \log u\,χ_{\{u>0\}} \quad \text{in } Ω,$$ where $Ω\subset\mathbb{R}^d$ ($d\geq 2$) is a smooth bounded domain. In our earlier work [arXiv:2408.08104, 2024], we proved the $C^{1,α}$ regularity of the free boundary $Ω\cap\partial\{u>0\}$ near regular points. In this paper, we investigate the more delicate structure of the \emph{singular} free boundary. Since the nonlinearity $-\log u$ is singular near the free boundary and destroys the scaling invariance, so that neither the classical blow-up arguments nor the standard epiperimetric inequality [Weiss, Invent.\ Math., 138, 23--50, 1999] apply directly; moreover, the Weiss type monotonicity formula requires a variable-parameter correction that introduces non-integrable remainder terms into the energy estimates. Motivated by Colombo--Spolaor--Velichkov [Geom.\ Funct.\ Anal., 28, 1029--1061, 2018], we develop a new \emph{log-epiperimetric inequality} for the modified Weiss energy, also proved by the direct method. A key novelty is the introduction of an auxiliary correction term $T$ that absorbs the non-integrable errors. As consequences, we establish a logarithmic energy decay, uniqueness of blow-ups at singular points, and a $C^{1,\log}$-type geometric description of the singular strata. In dimension two, the logarithmic modulus improves to a Hölder modulus.

math.AP↗

arXiv2026

On Global-in-time Solutions of Incompressible MHD Equations with Small Alfvén Numbers

In 1965 Kraichnan pointed out that a sufficiently strong background magnetic field, i.e. the case of small Alfvén number, will reduce the nonlinear interaction and inhibit the formation of strong gradients in the magnetohydrodynamic (abbr. MHD) system with $μ=ν\geqslant 0$, where $μ$ and $ν$ are the coefficients of kinematic viscosity and resistivity resp.. This means that the MHD system with $μ=ν\geqslant 0$ admits global-in-time large perturbation solutions with small Alfvén numbers. The existence of such large perturbation solutions was first mathematically verified in Hölder spaces by Bardos--Sulem--Sulem for the case $μ=ν= 0$ in 1988, and in Sobolev spaces by Cai--Cui--Jiang--Liu for the case $μ=ν> 0$ recently. In this paper, we further found a similar result for the general case ``$μ>0$ and $ν>0$", and provide a rigorous proof by developing a new approach, which includes a key bilinear estimate for dealing with the nonlinear interaction terms. Moreover both additional results for the vanishing behavior of the nonlinear interaction and the small Alfvén number limit of solutions are also established.

math.AP↗

arXiv2026

SemiGDA: Generative Dual-distribution Alignment for Semi-Supervised Medical Image Segmentation

Semi-supervised learning addresses label scarcity and high annotation costs in medical image segmentation by exploiting the latent information in unlabeled data to enhance model performance. Traditional discriminative segmentation relies on segmentation masks, neglecting feature-level distribution constraints. This limits robust semantic representation learning and adaptive modeling of unlabeled data in scenarios with few labels. To address these limitations, we propose SemiGDA, a novel Generative Dual-distribution Alignment framework for semi-supervised medical image segmentation. Our SemiGDA overcomes the reliance of discriminative methods on large labeled datasets by aligning feature and semantic distributions to boost semantic learning and scene adaptability. Specifically, we propose a Dual-distribution Alignment Module (DAM), which employs two structurally distinct encoders to model image and mask feature distributions. It enforces their alignment in the latent space via distributional constraints, establishing structured feature consistency. Moreover, we design a Consistency-Driven Skip Adapter (CDSA) strategy, which introduces dual skip adapters (Image and Mask) to fuse multi-scale features via skip connections. Using a consistency loss, CDSA enhances cross-branch semantic alignment and reinforces fine-grained semantic consistency. Experimental results on diverse medical datasets show that our method outperforms other state-of-the-art semi-supervised segmentation methods. Code is released at: https://github.com/taozh2017/SemiGDA.

cs.CV↗

arXiv2026

A Systematic Approach for Large Language Models Debugging

Large language models (LLMs) have become central to modern AI workflows, powering applications from open-ended text generation to complex agent-based reasoning. However, debugging these models remains a persistent challenge due to their opaque and probabilistic nature and the difficulty of diagnosing errors across diverse tasks and settings. This paper introduces a systematic approach for LLM debugging that treats models as observable systems, providing structured, model-agnostic methods from issue detection to model refinement. By unifying evaluation, interpretability, and error-analysis practices, our approach enables practitioners to iteratively diagnose model weaknesses, refine prompts and model parameters, and adapt data for fine-tuning or assessment, while remaining effective in contexts where standardized benchmarks and evaluation criteria are lacking. We argue that such a structured methodology not only accelerates troubleshooting but also fosters reproducibility, transparency, and scalability in the deployment of LLM-based systems.

cs.AI↗

arXiv2026

Pricing Electric Vehicle Charging and Station Access via Copositive Duality

Optimized charging of electric vehicles (EVs) at public locations consists of two decisions: how much energy to deliver at what times, which is continuous, and where to plug in, which is binary. This makes optimizing EV charging a mixed-integer linear program (MILP). This discreteness undermines traditional marginal pricing methods. In this paper, we develop the first marginal-price-based mechanism for pricing EV charging with binary station access constraints. Using the result of Burer (2009), we express the EV charging as a completely positive program (CPP), whose dual is a copositive program (COP). This convex dual admits valid shadow prices even though the original allocation problem is discrete and nonconvex. By interpreting the COP dual variables as marginal prices, we construct a pricing mechanism that captures EV supply equipment (EVSE) congestion as well as charging-capacity limits. We prove that the resulting mechanism is revenue-adequate for the operator and individually rational for every EV user, in the strong sense that each user maximizes their own welfare by accepting their assigned charging plan rather than deviating to any alternative option. We further develop problem-specific inner-approximation and dimension-reduction techniques that substantially improve the computational tractability of solving the COP in our setting. Numerical experiments on both small and large scale charging instances demonstrate that our pricing mechanism captures discrete congestion effects and aligns user incentives with the system-optimal assignment, outperforming time-of-use (TOU) and convex relaxation benchmarks.

math.OC↗

arXiv2026

Low-regularity global well-posedness for the Boltzmann equation near vacuum

We study the Boltzmann equation near vacuum in anisotropic low-regularity Besov spaces. We establish the global existence and uniqueness of strong solutions with the critical regularity index $2/p$ for $p\in[1,\infty)$ in $\mathbb{R}^3$. The proof relies on a new bilinear estimate for the nonlinear collision operator. Combined with a div-curl type lemma we develop, this allows us to close the a priori estimates and thereby obtain global well-posedness.

math.AP↗

arXiv2026

Co-Authoring with AI: How I Wrote a Physics Paper About AI, Using AI

The rapid integration of Large Language Models (LLMs) into scientific writing fundamentally challenges traditional definitions of authorship, responsibility, and scientific integrity. As researchers transition from using computers as deterministic tools to managing them as ``virtual collaborators,'' the nature of human contribution must be re-evaluated. Using the drafting process of a recent computational physics manuscript as a case study, this essay explores the indispensable role of the Human-in-the-Loop (HITL). We demonstrate that while AI excels at structural organization and syntax generation, the human author bears the ultimate responsibility for enforcing rigorous physical logic, maintaining academic diplomacy, and anticipating peer-review critiques. In this paradigm, the human contribution shifts from writing boilerplate text to acting as a Principal Investigator who actively mentors and steers the AI's reasoning. To ensure accountability and preserve the integrity of the scientific record in this new era, I argue that the community must mandate the publication of full, unedited AI interaction transcripts as standard supplementary material.

physics.comp-ph↗

arXiv2026

Enhancing Medical Visual Grounding via Knowledge-guided Spatial Prompts

Medical Visual Grounding (MVG) aims to identify diagnostically relevant phrases from free-text radiology reports and localize their corresponding regions in medical images, providing interpretable visual evidence to support clinical decision-making. Although recent Vision-Language Models (VLMs) exhibit promising multimodal reasoning ability, their grounding remains insufficient spatial precision, largely due to a lack of explicit localization priors when relying solely on latent embeddings. In this work, we analyze this limitation from an attention perspective and propose KnowMVG, a Knowledge-prior and global-local attention enhancement framework for MVG in VLMs that explicitly strengthens spatial awareness during decoding. Specifically, we present a knowledge-enhanced prompting strategy that encodes phrase related medical knowledge into compact embeddings, together with a global-local attention that jointly leverages coarse global information and refined local cues to guide precise region localization. localization. This design bridges high-level semantic understanding and fine-grained visual perception without introducing extra textual reasoning overhead. Extensive experiments on four MVG benchmarks demonstrate that our KnowMVG consistently outperforms existing approaches, achieving gains of 3.0% in AP50 and 2.6% in mIoU over prior state-of-the-art methods. Qualitative and ablation studies further validate the effectiveness of each component.

cs.CV↗

arXiv2026

Balancing Multi-modal Sensor Learning via Multi-objective Optimization

Learning-enabled control systems increasingly rely on multiple sensing modalities (e.g., vision, audio, language, etc.) for perception and decision support. A key challenge is that multi-modal sensor training dynamics are often imbalanced: fast-to-learn sensing channels dominate optimization, while slower channels remain underutilized, degrading reliability under sensing perturbations. Existing balancing strategies are largely heuristic and can require computationally intensive subroutines. In this paper, we reformulate multi-modal sensor learning as a multi-objective optimization (MOO) problem that explicitly prioritizes the worst-performing modality while retaining the nominal multi-modal sensor fusion objective. We then propose a simple gradient-based method, MIMO (multi-modal sensor learning via MOO), for the resulting formulation. We provide convergence guarantees and evaluate the method on standard multi-modal benchmarks. Results show improved balanced performance over state-of-the-art balanced multi-modal learning and MOO baselines, together with up to ~20x reduction in subroutine computation time, highlighting the suitability of MIMO for resource-constrained control pipelines.

cs.LG↗

arXiv2026

Single-Pulse Study of the Pseudo-nulling Pulsar PSR J1820-0509 Based on FAST Observations

Using two observations obtained with the Five-hundred-meter Aperture Spherical radio Telescope (FAST), we present a detailed single-pulse analysis of the high-nulling pulsar PSR J1820-0509. We measure an exceptionally high nulling fraction of approximately 81.78%, significantly exceeding previous estimates from Parkes observations. The single-pulse energy distribution exhibits a clear bimodal structure, consistent with classical nulling behavior. However, stacking the identified null pulses reveals a statistically significant residual profile above the noise level, indicating that the nulls correspond to a very weak emission state rather than a complete cessation of radio emission. The pulsar shows clustered burst activities spanning several hundred rotation periods, with prominent quasi-periodicities at 1191 +/- 81 and 590 +/- 15 pulse periods in the two observations. Based on temporal clustering and integrated profile morphology, we identify three distinct emission modes (A, B, and C) and a pseudo-null state (D). These modes exhibit systematic differences in pulse morphology, polarization, and energy statistics. The pulse width-energy relations reveal clear transitions between low- and high-energy regimes. The energy distributions of Modes A and C are well described by lognormal functions, while Mode B follows a composite Gaussian-lognormal distribution. These results suggest that the radio emission of PSR J1820-0509 is governed by multiple quasi-stable magnetospheric states. The presence of weak emission during pseudo-nulls, together with systematic mode-dependent variations, supports the interpretation that pulsar nulling reflects transitions between different magnetospheric activity levels rather than a complete shutdown of emission.

astro-ph.HE↗

arXiv2026

Interpretable Battery Aging without Extra Tests via Neural-Assisted Physics-based Modelling

State of health (SoH) is widely used for battery management, but it is a single scalar and offers limited interpretability. Two batteries with similar SoH can exhibit very different degradation behaviors and the lack of interpretability hinders optimal battery operation. In this paper, we propose IBAM for interpretable battery aging modelling with a neural-assisted physics-based framework. IBAM outputs a 2-D aging fingerprint without extra diagnostic tests and uses only routine logs from the battery management system. The fingerprint offers great interpretability by capturing a battery's curve-wide polarization voltage loss and the tail loss near the end-of-discharge. IBAM first creates a physics-based battery model based on a fractional-order equivalent circuit model, and then extracts per-cycle fingerprints from the model using a two-stage least-squares method. IBAM further anchors fingerprints on the SoH axis with physics-guided regression, where the per-cycle SoH is estimated via a bidirectional gated recurrent unit with customized multi-channel voltage features. Across batteries with short-, medium-, and long-lifespans, IBAM consistently yields the best physics model fidelity at different aging stages, and provides clear interpretations of degradation mechanisms and fingerprint patterns about batteries of different lifespans. The resulting fingerprints support interpretable battery health assessment and can inform battery control choices.

eess.SP↗

arXiv2026

LibraGen: Playing a Balance Game in Subject-Driven Video Generation

With the advancement of video generation foundation models (VGFMs), customized generation, particularly subject-to-video (S2V), has attracted growing attention. However, a key challenge lies in balancing the intrinsic priors of a VGFM, such as motion coherence, visual aesthetics, and prompt alignment, with its newly derived S2V capability. Existing methods often neglect this balance by enhancing one aspect at the expense of others. To address this, we propose LibraGen, a novel framework that views extending foundation models for S2V generation as a balance game between intrinsic VGFM strengths and S2V capability. Specifically, guided by the core philosophy of "Raising the Fulcrum, Tuning to Balance," we identify data quality as the fulcrum and advocate a quality-over-quantity approach. We construct a hybrid pipeline that combines automated and manual data filtering to improve overall data quality. To further harmonize the VGFM's native capabilities with its S2V extension, we introduce a Tune-to-Balance post-training paradigm. During supervised fine-tuning, both cross-pair and in-pair data are incorporated, and model merging is employed to achieve an effective trade-off. Subsequently, two tailored direct preference optimization (DPO) pipelines, namely Consis-DPO and Real-Fake DPO, are designed and merged to consolidate this balance. During inference, we introduce a time-dependent dynamic classifier-free guidance scheme to enable flexible and fine-grained control. Experimental results demonstrate that LibraGen outperforms both open-source and commercial S2V models using only thousand-scale training data.

cs.CV↗