arXiv subjects

Jing Huang

Publications and source records attributed to Jing Huang.

At least 73 records · Page 4Linked to original sources

arXiv2026

JW-VL: A Vision-Language Model for Solar Physics

Vision-Language Models (VLMs) have achieved breakthrough progress in general knowledge domains, yet adaptation to specialized scientific fields remains challenging due to multimodal representation shifts and the limited integration of domain-specific knowledge. To address the limitations of general-purpose VLMs when applied to solar physics image recognition, analysis, and reasoning, we propose JinWu Vision-Language (JW-VL), a fine-tuned foundation model tailored for solar physics. The model integrates multi-wavelength observational data from both space-based and ground-based telescopes, encompassing representative spectral bands spanning the photosphere, chromosphere, and corona. Built upon a cross-modal alignment knowledge distillation framework, JW-VL learns a joint visual-semantic embedding that enables end-to-end modeling from raw solar observational data to downstream tasks, including solar image recognition, solar activity analysis via image-based question answering, and optical character recognition (OCR), while also supporting the construction of a multi-band, cross-instrument solar image benchmark dataset. Furthermore, as a demonstration of interdisciplinary applicability, we developed a "Daily Solar Activity Reports" agent comprising core modules for solar activity level assessment, significant active region characterization, magnetic field complexity analysis, potential space weather impact assessment, and identifying active regions for targeted observation. While JW-VL may not yet meet the rigorous, high-precision demands of operational solar physics, it bridges raw observations and diverse downstream tasks, establishing a valuable methodological framework for applying multimodal deep learning to the field.

astro-ph.SR↗

arXiv2026

The Separating Noether Number of Finite Abelian Groups

For a finite abelian group $G$, let $β_{\mathrm{sep}}(G)$ denote its separating Noether number. We determine $β_{\mathrm{sep}}(G)$ exactly for every finite abelian group $ G \cong C_{n_1}\oplus \cdots \oplus C_{n_r}$ with $ 1<n_1 \mid \cdots \mid n_r. $ If $r=2s-1$, then $$ β_{\mathrm{sep}}(G)=n_s+n_{s+1}+\cdots+n_r, $$ whereas if $r=2s$, then $$ β_{\mathrm{sep}}(G)=\frac{n_s}{p_1}+n_{s+1}+\cdots+n_r, $$ where $p_1$ denotes the smallest prime divisor of $n_1$. Our proof is additive-combinatorial in nature. It avoids the Davenport-equality assumption $\mathsf{D}(n_sG)=\mathsf{D}^{*}(n_sG)$ used in previous works. The key ingredients are a geometric reduction of auxiliary sequences via the novel construction of geodesic surrogates, alongside a uniform lifting procedure for relation groups. As an application, we prove that if $r\ge 2$, then every extremal separating atom $A$ over $G_0$ with $|G_0|\le r+1$ satisfies $|\supp(A)|=|G_0|=r+1$. Equivalently, the conjectured support conclusion of Schefler, Zhao, and Zhong holds for all finite abelian groups of rank at least $2$. By contrast, the rank-$1$ case is exceptional: for cyclic groups, the analogous conjectural conclusion is false.

math.AC↗

arXiv2026

Self-Distilled Reasoner: On-Policy Self-Distillation for Large Language Models

Knowledge distillation improves large language model (LLM) reasoning by compressing the knowledge of a teacher LLM to train smaller LLMs. On-policy distillation advances this approach by having the student sample its own trajectories while a teacher LLM provides dense token-level supervision, addressing the distribution mismatch between training and inference in off-policy distillation methods. However, on-policy distillation typically requires a separate, often larger, teacher LLM and does not explicitly leverage ground-truth solutions available in reasoning datasets. Inspired by the intuition that a sufficiently capable LLM can rationalize external privileged reasoning traces and teach its weaker self, we introduce On-Policy Self-Distillation (OPSD), a learning algorithm where a single LLM acts as both teacher and student with different contexts. The teacher policy conditions on privileged information (e.g., verified reasoning traces) while the student policy sees only the question; training minimizes the per-token divergence between these distributions over the student's own rollouts. We demonstrate the efficacy of our method on multiple mathematical reasoning benchmarks, achieving superior token efficiency compared to reinforcement learning methods and better performance over off-policy distillation methods. Code repo: https://github.com/siyan-zhao/OPSD.

cs.LG↗

arXiv2026

A Proof of a Conjecture of Móricz and Nagy on Rational-Value Sums

Móricz and Nagy introduced the problem of maximizing the number of $r$-element subsets with rational sums in an $n$-element set of irrational numbers, and showed that it is equivalent to an extremal zero-sum problem. They determined the exact maximum in several cases. For the remaining range, they presented an explicit construction of an $n$-element set of irrational numbers containing exactly $m\binom{n-m}{r-1}$ such subsets, where $m=\lfloor n/r\rfloor$. They conjectured that this construction is always optimal for any $1<r<n$. In this paper, we confirm that conjecture. Our proof combines an order-theoretic antichain argument for zero-sum subsets with a sharp maximization of the resulting binomial expressions. As a consequence, we determine exactly the maximum number of $r$-term zero-sum subsequences in sequences of $n$ nonzero integers.

math.CO↗

arXiv2026

Vector sketch animation generation with differentiable motion trajectories

Sketching is a direct and inexpensive means of visual expression. Though image-based sketching has been well studied, video-based sketch animation generation is still very challenging due to the temporal coherence requirement. In this paper, we propose a novel end-to-end automatic generation approach for vector sketch animation. To solve the flickering issue, we introduce a Differentiable Motion Trajectory (DMT) representation that describes the frame-wise movement of stroke control points using differentiable polynomial-based trajectories. DMT enables global semantic gradient propagation across multiple frames, significantly improving the semantic consistency and temporal coherence, and producing high-framerate output. DMT employs a Bernstein basis to balance the sensitivity of polynomial parameters, thus achieving more stable optimization. Instead of implicit fields, we introduce sparse track points for explicit spatial modeling, which improves efficiency and supports long-duration video processing. Evaluations on DAVIS and LVOS datasets demonstrate the superiority of our approach over SOTA methods. Cross-domain validation on 3D models and text-to-video data confirms the robustness and compatibility of our approach.

cs.GR↗

arXiv2026

Chart-R1: Chain-of-Thought Supervision and Reinforcement for Advanced Chart Reasoner

Chart reasoning presents unique challenges due to its inherent complexity -- requiring precise numerical comprehension, multi-level visual understanding, and logical inference across interconnected data elements. Existing vision-language models often struggle with such reasoning tasks, particularly when handling multi-subchart scenarios and numerical sensitivity. To address these challenges, we introduce Chart-R1, a chart-domain vision-language model that leverages reinforcement fine-tuning for advanced chart reasoning. We first propose a programmatic data synthesis approach to generate high-quality step-by-step reasoning data with verifiable answer formats, covering diverse chart types and complexity levels. Our two-stage training strategy includes: (1) Chart-COT, which decomposes complex reasoning into interpretable subtasks through chain-of-thought supervision, and (2) Chart-RFT, which employs group relative policy optimization with numerically sensitive rewards tailored for chart-specific reasoning. Experiments on open-source benchmarks and our proposed ChartRQA dataset demonstrate that Chart-R1 significantly outperforms existing chart-domain methods and rivals large-scale open/closed-source models.

cs.AI↗

arXiv2026

Breaking the SFT Plateau: Multimodal Structured Reinforcement Learning for Chart-to-Code Generation

While reinforcement learning (RL) has proven highly effective for general reasoning in vision-language models, its application to tasks requiring deep understanding of information-rich images and structured output generation remains underexplored. Chart-to-code generation exemplifies this challenge, demanding complex reasoning over visual charts to produce structured code. Supervised fine-tuning (SFT) alone is often insufficient, highlighting the need for effective RL strategies tailored to structured outputs. In this paper, we systematically investigate the performance plateau of SFT through large-scale experiments and propose Multimodal Structured Reinforcement Learning (MSRL) for chart-to-code generation. We construct the largest training corpus to date, with 3 million chart-code pairs curated from real-world tables in arXiv papers, addressing the limitations of previous synthetic datasets. Despite achieving state-of-the-art performance, our experiments show that simply increasing SFT data eventually leads to diminishing improvements. To break this plateau, MSRL employs a multi-granularity reward system that integrates both textual and visual feedback. At the textual level, rule-based rewards validate fine-grained code details, while at the visual level, a model-based reward assesses the structural similarity between rendered code and ground-truth charts. We implement a two-stage curriculum training strategy, first optimizing the model with textual rewards and then incorporating visual signals for further enhancement. Experimental results demonstrate that MSRL substantially breaks the SFT plateau, improving high-level metrics by 6.2% and 9.9% on ChartMimic and ReachQA benchmarks, respectively. Notably, our method outperforms all existing approaches in the chart domain and achieves competitive results with advanced closed-source models.

cs.AI↗

arXiv2026

The Unlearning Mirage: A Dynamic Framework for Evaluating LLM Unlearning

Unlearning in Large Language Models (LLMs) aims to enhance safety, mitigate biases, and comply with legal mandates, such as the right to be forgotten. However, existing unlearning methods are brittle: minor query modifications, such as multi-hop reasoning and entity aliasing, can recover supposedly forgotten information. As a result, current evaluation metrics often create an illusion of effectiveness, failing to detect these vulnerabilities due to reliance on static, unstructured benchmarks. We propose a dynamic framework that stress tests unlearning robustness using complex structured queries. Our approach first elicits knowledge from the target model (pre-unlearning) and constructs targeted probes, ranging from simple queries to multi-hop chains, allowing precise control over query difficulty. Our experiments show that the framework (1) shows comparable coverage to existing benchmarks by automatically generating semantically equivalent Q&A probes, (2) aligns with prior evaluations, and (3) uncovers new unlearning failures missed by other benchmarks, particularly in multi-hop settings. Furthermore, activation analyses show that single-hop queries typically follow dominant computation pathways, which are more likely to be disrupted by unlearning methods. In contrast, multi-hop queries tend to use alternative pathways that often remain intact, explaining the brittleness of unlearning techniques in multi-hop settings. Our framework enables practical and scalable evaluation of unlearning methods without the need for manual construction of forget test sets, enabling easier adoption for real-world applications. We release the pip package and the code at https://sites.google.com/view/unlearningmirage/home.

cs.AI↗

arXiv2026

HiFloat4 Format for Language Model Inference

This paper introduces HiFloat4 (HiF4), a block floating-point data format tailored for deep learning. Each HiF4 unit packs 64 4-bit elements with 32 bits of shared scaling metadata, averaging 4.5 bits per value. The metadata specifies a three-level scaling hierarchy, capturing inter- and intra-group dynamic range while improving the utilization of the representational space. In addition, the large 64-element group size enables matrix multiplications to be executed in a highly fixed-point manner, significantly reducing hardware area and power consumption. To evaluate the proposed format, we conducted inference experiments on several language models, including LLaMA, Qwen, Mistral, DeepSeek-V3.1 and LongCat. Results show that HiF4 achieves higher average accuracy than the state-of-the-art NVFP4 format across multiple models and diverse downstream tasks.

cs.LG↗

arXiv2026

Critical Numbers for Restricted Sumsets: Rigidity and Collapse in Finite Abelian Groups

This paper establishes a classification of the critical numbers for restricted sumsets in finite abelian groups, determining them exactly for even-order groups and bounding them for odd-order groups, while revealing a fundamental structural dichotomy governed by parity. For groups of even order, we prove a universal rigidity theorem: the index-$2$ subgroup creates an immutable arithmetic barrier at density $1/2$, fixing the critical number at $|G|/2+1$ regardless of the group's internal structure. In sharp contrast, we demonstrate that for groups of odd order, this barrier vanishes, causing the critical threshold to collapse to significantly lower densities bounded by index-$5$ obstructions or the smallest prime divisor. These results unify and vastly generalize previous work on cyclic groups, providing a definitive structural theory for the transition from sparsity to saturation. As a decisive application, we resolve a conjecture of Han and Ren in algebraic coding theory. By translating the additive rigidity at density $1/2$ into a geometric constraint, we prove that for all sufficiently large $q$, any subset of rational points on an elliptic curve $E/\mathbb{F}_q$ generating an MDS code must satisfy the tight bound $|P|\le|E(\mathbb{F}_q)|/2$.

math.CO↗

arXiv2026

TreeCUA: Efficiently Scaling GUI Automation with Tree-Structured Verifiable Evolution

Effectively scaling GUI automation is essential for computer-use agents (CUAs); however, existing work primarily focuses on scaling GUI grounding rather than the more crucial GUI planning, which requires more sophisticated data collection. In reality, the exploration process of a CUA across apps/desktops/web pages typically follows a tree structure, with earlier functional entry points often being explored more frequently. Thus, organizing large-scale trajectories into tree structures can reduce data cost and streamline the data scaling of GUI planning. In this work, we propose TreeCUA to efficiently scale GUI automation with tree-structured verifiable evolution. We propose a multi-agent collaborative framework to explore the environment, verify actions, summarize trajectories, and evaluate quality to generate high-quality and scalable GUI trajectories. To improve efficiency, we devise a novel tree-based topology to store and replay duplicate exploration nodes, and design an adaptive exploration algorithm to balance the depth (\emph{i.e.}, trajectory difficulty) and breadth (\emph{i.e.}, trajectory diversity). Moreover, we develop world knowledge guidance and global memory backtracking to avoid low-quality generation. Finally, we naturally extend and propose the TreeCUA-DPO method from abundant tree node information, improving GUI planning capability by referring to the branch information of adjacent trajectories. Experimental results show that TreeCUA and TreeCUA-DPO offer significant improvements, and out-of-domain (OOD) studies further demonstrate strong generalization. All trajectory node information and code will be available at https://github.com/UITron-hub/TreeCUA.

cs.CV↗

arXiv2026

OCRVerse: Towards Holistic OCR in End-to-End Vision-Language Models

The development of large vision language models drives the demand for managing, and applying massive amounts of multimodal data, making OCR technology, which extracts information from visual images, increasingly popular. However, existing OCR methods primarily focus on recognizing text elements from images or scanned documents (Text-centric OCR), neglecting the identification of visual elements from visually information-dense image sources (Vision-centric OCR), such as charts, web pages and science plots. In reality, these visually information-dense images are widespread on the internet and have significant real-world application value, such as data visualization and web page analysis. In this technical report, we propose OCRVerse, the first holistic OCR method in end-to-end manner that enables unified text-centric OCR and vision-centric OCR. To this end, we constructe comprehensive data engineering to cover a wide range of text-centric documents, such as newspapers, magazines and books, as well as vision-centric rendered composites, including charts, web pages and scientific plots. Moreover, we propose a two-stage SFT-RL multi-domain training method for OCRVerse. SFT directly mixes cross-domain data to train and establish initial domain knowledge, while RL focuses on designing personalized reward strategies for the characteristics of each domain. Specifically, since different domains require various output formats and expected outputs, we provide sufficient flexibility in the RL stage to customize flexible reward signals for each domain, thereby improving cross-domain fusion and avoiding data conflicts. Experimental results demonstrate the effectiveness of OCRVerse, achieving competitive results across text-centric and vision-centric data types, even comparable to large-scale open-source and closed-source models.

cs.CV↗

arXiv2026

Ultrastrong light-matter coupling in near-field coupled split-ring resonators revealed by photocurrent spectroscopy

Landau polaritons arising from the coupling between cyclotron resonance and terahertz split-ring resonators (SRRs) have served as a central platform for exploring ultrastrong light-matter interaction for more than a decade. Over this period, a wide variety of SRR architectures, differing in size, geometry, and even material composition, have been investigated. However, the regime of near-field coupled SRRs has remained largely unexplored. Here, we demonstrate ultrastrong coupling using photocurrent spectroscopy in two prototypical near-field configurations: a SRR dimer and a topological SRR chain. The measurements reveal hybridization not only with bright resonant modes but also with optically dark modes and topological edge modes, highlighting the exceptional sensitivity of the photocurrent spectroscopy. Moreover, the engineered near-field interactions allow the study of multi-mode ultrastrong coupling and the interplay between topological band structure and cavity quantum electrodynamics.

cond-mat.mes-hall↗

arXiv2026

Beyond Symbols: Motion Perception Cues Enhance Dual-Task Performance with Wearable Directional Guidance

Directional cues are crucial for environmental interaction. Conventional methods rely on symbolic visual or auditory reminders that require semantic interpretation, a process that proves challenging in demanding dual-tasking scenarios. We introduce a novel alternative for conveying directional cues on wearable displays: directly triggering motion perception using monocularly presented peripheral stimuli. This approach is designed for low visual interference, with the goal of reducing the need for gaze-switching and the complex cognitive processing associated with symbols. User studies demonstrate our method's potential to robustly convey directional cues. Compared to a conventional arrow-based technique in a demanding dual-task scenario, our motion-based approach resulted in significantly more accurate interpretation of these directional cues ($p=.008$) and showed a trend towards reduced errors on the concurrent primary task ($p=.066$).

cs.HC↗

arXiv2026

Diffusion Epistemic Uncertainty with Asymmetric Learning for Diffusion-Generated Image Detection

The rapid progress of diffusion models highlights the growing need for detecting generated images. Previous research demonstrates that incorporating diffusion-based measurements, such as reconstruction error, can enhance the generalizability of detectors. However, ignoring the differing impacts of aleatoric and epistemic uncertainty on reconstruction error can undermine detection performance. Aleatoric uncertainty, arising from inherent data noise, creates ambiguity that impedes accurate detection of generated images. As it reflects random variations within the data (e.g., noise in natural textures), it does not help distinguish generated images. In contrast, epistemic uncertainty, which represents the model's lack of knowledge about unfamiliar patterns, supports detection. In this paper, we propose a novel framework, Diffusion Epistemic Uncertainty with Asymmetric Learning~(DEUA), for detecting diffusion-generated images. We introduce Diffusion Epistemic Uncertainty~(DEU) estimation via the Laplace approximation to assess the proximity of data to the manifold of diffusion-generated samples. Additionally, an asymmetric loss function is introduced to train a balanced classifier with larger margins, further enhancing generalizability. Extensive experiments on large-scale benchmarks validate the state-of-the-art performance of our method.

cs.CV↗

arXiv2026

Chordal signed graphs and signed bigraphs

Chordal graphs and chordal bigraphs enjoy beautiful characterizations, in terms of forbidden subgraphs, vertex/edge orderings, vertex/edge separating sets, and tree-like representations. In this paper, we introduce chordal signed graphs and chordal signed bigraphs. Interestingly, chordal signed graphs are equivalent to strict chordal digraphs studied by Hell and Hernández-Cruz. A forbidden subdigraph characterization of strict chordal digraphs can be translated to a forbidden subgraph characterization of chordal signed graphs. We give a forbidden subgraph characterization of chordal signed bigraphs. The forbidden subgraphs for chordal signed bigraphs are analogous to those for chordal signed graphs but the proofs are much more complicated and intriguing.

math.CO↗

arXiv2026

An AI-guided mechanotyping instrument for fully automated oocyte quality assessment

The mechanical properties of oocytes are regarded as important indicators of their developmental potential. During fertilization, deviations from the normal mechanical range can hinder sperm penetration, ultimately reducing fertilization efficiency and compromising embryo quality. However, current methods for measuring oocyte mechanics often suffer from serious cellular damage, low automation levels, and large measurement errors. To address these limitations, we developed an AI-guided micronewton-scale mechanical measurement system for safe and automated oocyte quality assessment. The system integrates voice interaction with automated experimental workflows to control a magnetically actuated microgripper, which applies defined loading forces to induce micron-scale compressive deformation of the oocyte. Combined with AI-assisted object detection and image segmentation algorithms, the system captures cellular deformation in real time, enabling precise calculation of the oocyte's compressive modulus. This measurement system enables automated, quantitative, and non-destructive evaluation of oocyte mechanical properties, providing an effective approach for oocyte quality screening in in vitro fertilization (IVF) and other assisted reproductive technologies (ART).

physics.bio-ph↗

arXiv2026

Digital Twin AI: Opportunities and Challenges from Large Language Models to World Models

Digital twins, as precise digital representations of physical systems, have evolved from passive simulation tools into intelligent and autonomous entities through the integration of artificial intelligence technologies. This paper presents a unified four-stage framework that systematically characterizes AI integration across the digital twin lifecycle, spanning modeling, mirroring, intervention, and autonomous management. By synthesizing existing technologies and practices, we distill a unified four-stage framework that systematically characterizes how AI methodologies are embedded across the digital twin lifecycle: (1) modeling the physical twin through physics-based and physics-informed AI approaches, (2) mirroring the physical system into a digital twin with real-time synchronization, (3) intervening in the physical twin through predictive modeling, anomaly detection, and optimization strategies, and (4) achieving autonomous management through large language models, foundation models, and intelligent agents. We analyze the synergy between physics-based modeling and data-driven learning, highlighting the shift from traditional numerical solvers to physics-informed and foundation models for physical systems. Furthermore, we examine how generative AI technologies, including large language models and generative world models, transform digital twins into proactive and self-improving cognitive systems capable of reasoning, communication, and creative scenario generation. Through a cross-domain review spanning eleven application domains, including healthcare, aerospace, smart manufacturing, robotics, and smart cities, we identify common challenges related to scalability, explainability, and trustworthiness, and outline directions for responsible AI-driven digital twin systems.

cs.AI↗