Searcharxiv⌕ Search

arXiv subjects

Qian Yu

Publications and source records attributed to Qian Yu.

At least 37 records · Page 2Linked to original sources
arXiv2026

Molecular Identifier Visual Prompt and Verifiable Reinforcement Learning for Chemical Reaction Diagram Parsing

Reaction diagram parsing (RxnDP) is critical for extracting chemical synthesis information from literature. Although recent Vision-Language Models (VLMs) have emerged as a promising paradigm to automate this complex visual reasoning task, their application is fundamentally bottlenecked by the inability to align visual chemical entities with pre-trained knowledge, alongside the inherent discrepancy between token-level training and reaction-level evaluation. To address these dual challenges, this work enhances VLM-based RxnDP from two complementary perspectives: prompting representation and learning paradigms. First, we propose Identifier as Visual Prompting (IdtVP), which leverages naturally occurring molecule identifiers (e.g., bold numerals like 1a) to activate the chemical knowledge acquired during VLM pre-training. IdtVP enables powerful zero-shot and out-of-distribution capabilities, outperforming existing prompting strategies. Second, to further optimize performance within fine-tuning paradigms, we introduce Re3-DAPO, a reinforcement learning algorithm that leverages verifiable rewards to directly optimize reaction-level metrics, thereby achieving consistent gains over standard supervised fine-tuning. Additionally, we release the ScannedRxn benchmark, comprising scanned historical reaction diagrams with real-world artifacts, to rigorously assess model robustness and out-of-distribution ability. Our contributions advance the accuracy and generalization of VLM-based reaction diagram parsing. We will release data, models, and code on GitHub.

cs.CV↗
arXiv2026

Small ball probability of collision local time for symmetric stable processes

In this article, the small ball probability is obtained for the collision local time of two independent symmetric $α-$stable processes with parameters $α_1,α_2\in(0,2]$ satisfying $\max\{α_1,α_2\}>1$. The proof is based on obtaining the asymptotic behavior of moment generating function by contour integration.

math.PR↗
arXiv2026

Strong solutions to SDEs with singular drifts driven by fractional Brownian motions

In this paper, we establish the strong well-posedness of SDEs with merely integrable time-dependent drifts driven by fractional Brownian motions with Hurst parameter H<1/2. Our result holds over the entire subcritical regime and can be regarded as an extension of (Krylov and Rockner, Probab. Theory Relat. Fields, 131(2): 154-196 (2005)) to the fractional case. Furthermore, we prove the existence of stochastic flows of Sobolev diffeomorphisms for this class of SDEs, which generalizes a result in (Mohammed et al., Ann. Probab. 43, 1535-1576 (2015)). The approach adopted in our work is based on a compactness criterion for random fields in Wiener spaces.

math.PR↗
arXiv2026

RxnCaption: Reformulating Reaction Diagram Parsing as Visual Prompt Guided Captioning

Large-scale chemical reaction datasets are crucial for AI research in chemistry. However, existing chemical reaction data often exist as images within papers, making them not machine-readable and unusable for training machine learning models. In response to this challenge, we propose the RxnCaption framework for the task of chemical Reaction Diagram Parsing (RxnDP). Our framework reformulates the traditional coordinate prediction driven parsing process into an image captioning problem, which Large Vision Language Models (LVLMs) handle naturally. We introduce a strategy termed BBox and Index as Visual Prompt (BIVP), which uses our state-of-the-art molecular detector, MolYOLO, to pre-draw molecular bounding boxes and indices directly onto the input image. This turns the downstream parsing into a natural-language description problem. Extensive experiments show that the BIVP strategy significantly improves structural extraction quality while simplifying model design. We further construct the RxnCaption-15k dataset, an order of magnitude larger than prior real-world literature benchmarks, with a balanced test subset across four layout archetypes. Experiments demonstrate that RxnCaption-VL achieves state-of-the-art performance on multiple metrics. We believe our method, dataset, and models will advance structured information extraction from chemical literature and catalyze broader AI applications in chemistry. We will release data, models, and code on GitHub.

cs.CV↗
arXiv2025

Prune4Web: DOM Tree Pruning Programming for Web Agent

Web automation employs intelligent agents to execute high-level tasks by mimicking human interactions with web interfaces. Despite the capabilities of recent Large Language Model (LLM)-based web agents, navigating complex, real-world webpages efficiently remains a significant hurdle due to the prohibitively large size of Document Object Model (DOM) structures, often ranging from 10,000 to 100,000 tokens. Existing strategies typically rely on crude DOM truncation -- risking the loss of critical information -- or employ inefficient heuristics and separate ranking models, failing to achieve an optimal balance between precision and scalability. To address these challenges, we introduce Prune4Web, a novel paradigm that shifts DOM processing from resource-intensive LLM reading to efficient programmatic pruning. Central to our approach is DOM Tree Pruning Programming, where an LLM generates executable Python scoring scripts to dynamically filter DOM elements based on semantic cues from decomposed sub-tasks. This mechanism eliminates the need for LLMs to ingest raw, massive DOMs, instead delegating traversal and scoring to lightweight, interpretable programs. This methodology achieves a 25x to 50x reduction in candidate elements for grounding, thereby facilitating precise action localization while mitigating attention dilution. Furthermore, we propose a specialized data annotation pipeline and a two-turn dialogue training strategy that jointly optimizes the Planner, Programmatic Filter, and Grounder within a unified framework. Extensive experiments demonstrate state-of-the-art performance. Notably, on our low-level grounding task, Prune4Web dramatically improves accuracy from 46.8% to 88.28%, underscoring its efficacy in real-world web automation.

cs.AI↗
arXiv2025

Asymptotic Properties of the Derivative of Self-Intersection Local Time of Multidimensional Fractional Brownian Motion

Let \{B_t^H,t\geq0\} be a d-dimensional fractional Brownian motion. We prove that the approximation of the first-order derivative of self-intersection local time, defined as α_{\varepsilon,t}^{(1)}(0)=-\int_0^t\int_0^sp_\varepsilon^{(1)}(B_s^H-B_r^H)\d r\d s, where p_\varepsilon^{(1)}(x_1,\cdots,x_d):=\partial _{x_1}p(x_1,\cdots,x_d) and p_\varepsilon(x)=(2π\varepsilon)^{-d/2}e^{|x|^2/2\varepsilon},x\in\mathbb{R}^d, d\geq2 is the heat kernel, exits in L^2 sense if and only if H<\frac{3}{2(1+d)} and satisfies three different central limit theorems when normalized by \varepsilon^{\frac d2+1-\frac1H} for H>\frac12 and d\geq2, normalized by \varepsilon^{\frac d2+\frac12-\frac 3{4H}} for \frac{3}{2(1+d)}<H<\frac12 and d\geq3, and normalized by \log(1/\varepsilon)^{-\frac12} for the critical case H=\frac{3}{2(1+d)} and d\geq3.

math.PR↗
arXiv2025

CAD-Coder: Text-to-CAD Generation with Chain-of-Thought and Geometric Reward

In this work, we introduce CAD-Coder, a novel framework that reformulates text-to-CAD as the generation of CadQuery scripts - a Python-based, parametric CAD language. This representation enables direct geometric validation, a richer modeling vocabulary, and seamless integration with existing LLMs. To further enhance code validity and geometric fidelity, we propose a two-stage learning pipeline: (1) supervised fine-tuning on paired text-CadQuery data, and (2) reinforcement learning with Group Reward Policy Optimization (GRPO), guided by a CAD-specific reward comprising both a geometric reward (Chamfer Distance) and a format reward. We also introduce a chain-of-thought (CoT) planning process to improve model reasoning, and construct a large-scale, high-quality dataset of 110K text-CadQuery-3D model triplets and 1.5K CoT samples via an automated pipeline. Extensive experiments demonstrate that CAD-Coder enables LLMs to generate diverse, valid, and complex CAD models directly from natural language, advancing the state of the art of text-to-CAD generation and geometric reasoning.

cs.GR↗
arXiv2025

On the exponential integrability of the derivative of intersection and self-intersection local time for fractional Brownian motion and a limit theorem related to the self-intersection local time for fractional Brownian motion

We give the correct condition for existence of the $k$-th derivative of the intersection local time for fractional Brownian motion, which was originally discussed in [Guo, J., Hu, Y., and Xiao, Y., Higher-order derivative of intersection local time for two independent fractional Brownian motions, Journal of Theoretical Probability 32, (2019), pp. 1190-1201]. We also show that the $k$-th derivative of the intersection and self-intersection local times of fractional Brownian motion are exponentially integrable for certain parameter values. In addition, we show convergence in distribution when the existence condition is violated for the $k$-th derivative of self-intersection local time of fractional Brownian motion under scaling.

math.PR↗
arXiv2025

Lumina-DiMOO: An Omni Diffusion Large Language Model for Multi-Modal Generation and Understanding

We introduce Lumina-DiMOO, an open-source foundational model for seamless multi-modal generation and understanding. Lumina-DiMOO sets itself apart from prior unified models by utilizing a fully discrete diffusion modeling to handle inputs and outputs across various modalities. This innovative approach allows Lumina-DiMOO to achieve higher sampling efficiency compared to previous autoregressive (AR) or hybrid AR-Diffusion paradigms and adeptly support a broad spectrum of multi-modal tasks, including text-to-image generation, image-to-image generation (e.g., image editing, subject-driven generation, and image inpainting, etc.), as well as image understanding. Lumina-DiMOO achieves state-of-the-art performance on multiple benchmarks, surpassing existing open-source unified multi-modal models. To foster further advancements in multi-modal and discrete diffusion model research, we release our code and checkpoints to the community. Project Page: https://synbol.github.io/Lumina-DiMOO.

cs.CV↗
arXiv2025

SafeToolBench: Pioneering a Prospective Benchmark to Evaluating Tool Utilization Safety in LLMs

Large Language Models (LLMs) have exhibited great performance in autonomously calling various tools in external environments, leading to better problem solving and task automation capabilities. However, these external tools also amplify potential risks such as financial loss or privacy leakage with ambiguous or malicious user instructions. Compared to previous studies, which mainly assess the safety awareness of LLMs after obtaining the tool execution results (i.e., retrospective evaluation), this paper focuses on prospective ways to assess the safety of LLM tool utilization, aiming to avoid irreversible harm caused by directly executing tools. To this end, we propose SafeToolBench, the first benchmark to comprehensively assess tool utilization security in a prospective manner, covering malicious user instructions and diverse practical toolsets. Additionally, we propose a novel framework, SafeInstructTool, which aims to enhance LLMs' awareness of tool utilization security from three perspectives (i.e., \textit{User Instruction, Tool Itself, and Joint Instruction-Tool}), leading to nine detailed dimensions in total. We experiment with four LLMs using different methods, revealing that existing approaches fail to capture all risks in tool utilization. In contrast, our framework significantly enhances LLMs' self-awareness, enabling a more safe and trustworthy tool utilization.

cs.CR↗
arXiv2025

VQualA 2025 Challenge on Engagement Prediction for Short Videos: Methods and Results

This paper presents an overview of the VQualA 2025 Challenge on Engagement Prediction for Short Videos, held in conjunction with ICCV 2025. The challenge focuses on understanding and modeling the popularity of user-generated content (UGC) short videos on social media platforms. To support this goal, the challenge uses a new short-form UGC dataset featuring engagement metrics derived from real-world user interactions. This objective of the Challenge is to promote robust modeling strategies that capture the complex factors influencing user engagement. Participants explored a variety of multi-modal features, including visual content, audio, and metadata provided by creators. The challenge attracted 97 participants and received 15 valid test submissions, contributing significantly to progress in short-form UGC video engagement prediction.

cs.CV↗
arXiv2025

Rethinking Cross-Subject Data Splitting for Brain-to-Text Decoding

Recent major milestones have successfully reconstructed natural language from non-invasive brain signals (e.g. functional Magnetic Resonance Imaging (fMRI) and Electroencephalogram (EEG)) across subjects. However, we find current dataset splitting strategies for cross-subject brain-to-text decoding are wrong. Specifically, we first demonstrate that all current splitting methods suffer from data leakage problem, which refers to the leakage of validation and test data into training set, resulting in significant overfitting and overestimation of decoding models. In this study, we develop a right cross-subject data splitting criterion without data leakage for decoding fMRI and EEG signal to text. Some SOTA brain-to-text decoding models are re-evaluated correctly with the proposed criterion for further research.

cs.CL↗
arXiv2025

PeTTO: Leveraging GPUs to Accelerate Topology Optimization with the Pseudo-Transient Methods

We present a Pseudo-Transient Topology Optimization (PeTTO) approach that can leverage graphics processing units (GPUs) to efficiently solve single-material and multi-material topology optimization problems. By integrating PeTTO with phase field methods, the partial differential equations (PDEs) constrained optimization problem in topology optimization is transformed into a set of time dependent PDEs, which can be analyzed using the knowledge of transient physics. The sensitivities with respect to the design variable are calculated with the automatic differentiation which help avoid tedious and error-prone manual derivations. The overall system of equations is efficiently solved using a hybrid of the pseudo-transient method and the accelerated pseudo-transient method, balancing the convergence rate and numerical stability. A variety of numerical examples are presented to demonstrate the effectiveness and efficiency of the proposed PeTTO approach. These examples cover different physics scenarios including mechanical and thermal problems, as well as single-material and multi-materials cases in both 2D and 3D. The numerical results show a 40- to 50-fold speedup when running the same PeTTO code on a single GPU compared to desktop CPUs. This work helps bridge the gap between high-performance computing and topology optimization, potentially enabling faster and better designs for real-world problems.

math.NA↗
arXiv2025

Trapping of electrons and $^{40}\textrm{Ca}^+$ ions in a dual-frequency Paul trap

We demonstrate the operation of a dual-frequency Paul trap and characterize its performance by storing either electrons or calcium ions while applying two quadrupole fields simultaneously which oscillate at $Ω_\textrm{fast} = 2π\times 1.6$ GHz and $Ω_\textrm{slow} = 2π\times 2$ MHz. The particles are loaded and stored in the trap under various conditions followed by detection employing an electron multiplier tube. We find that tens of electrons or ions can be trapped for up to ten milliseconds and a small fraction remains trapped even after hundreds of milliseconds. During dual-frequency operation we find that while the number of trapped electrons rapidly decreases with increase of the $Ω_\textrm{slow}$ field amplitude, the number of trapped ions shows no dependence on the $Ω_\textrm{fast}$ field amplitude as supported by our extensive numerical simulations. We aim to use a similar trap for synthesising antihydrogen from antiprotons and positrons. Accordingly, we discuss open challenges such as the co-trapping of oppositely charged species and particle trap duration.

physics.atom-ph↗
arXiv2025

Multi-Object Sketch Animation with Grouping and Motion Trajectory Priors

We introduce GroupSketch, a novel method for vector sketch animation that effectively handles multi-object interactions and complex motions. Existing approaches struggle with these scenarios, either being limited to single-object cases or suffering from temporal inconsistency and poor generalization. To address these limitations, our method adopts a two-stage pipeline comprising Motion Initialization and Motion Refinement. In the first stage, the input sketch is interactively divided into semantic groups and key frames are defined, enabling the generation of a coarse animation via interpolation. In the second stage, we propose a Group-based Displacement Network (GDN), which refines the coarse animation by predicting group-specific displacement fields, leveraging priors from a text-to-video model. GDN further incorporates specialized modules, such as Context-conditioned Feature Enhancement (CCFE), to improve temporal consistency. Extensive experiments demonstrate that our approach significantly outperforms existing methods in generating high-quality, temporally consistent animations for complex, multi-object sketches, thus expanding the practical applications of sketch animation.

cs.CV↗
arXiv2025

The net-regular strongly regular signed graphs with degree 6

In this paper, we study the net-regular strongly regular signed graphs with degree 6 and determine all connected 6-regular and net-regular strongly regular signed graphs. There are three, six and four 6-regular strongly regular signed graphs with net-degree 4, 2 and 0, respectively.

math.CO↗
arXiv2025

Move to Understand a 3D Scene: Bridging Visual Grounding and Exploration for Efficient and Versatile Embodied Navigation

Embodied scene understanding requires not only comprehending visual-spatial information that has been observed but also determining where to explore next in the 3D physical world. Existing 3D Vision-Language (3D-VL) models primarily focus on grounding objects in static observations from 3D reconstruction, such as meshes and point clouds, but lack the ability to actively perceive and explore their environment. To address this limitation, we introduce \underline{\textbf{M}}ove \underline{\textbf{t}}o \underline{\textbf{U}}nderstand (\textbf{\model}), a unified framework that integrates active perception with \underline{\textbf{3D}} vision-language learning, enabling embodied agents to effectively explore and understand their environment. This is achieved by three key innovations: 1) Online query-based representation learning, enabling direct spatial memory construction from RGB-D frames, eliminating the need for explicit 3D reconstruction. 2) A unified objective for grounding and exploring, which represents unexplored locations as frontier queries and jointly optimizes object grounding and frontier selection. 3) End-to-end trajectory learning that combines \textbf{V}ision-\textbf{L}anguage-\textbf{E}xploration pre-training over a million diverse trajectories collected from both simulated and real-world RGB-D sequences. Extensive evaluations across various embodied navigation and question-answering benchmarks show that MTU3D outperforms state-of-the-art reinforcement learning and modular navigation approaches by 14\%, 23\%, 9\%, and 2\% in success rate on HM3D-OVON, GOAT-Bench, SG3D, and A-EQA, respectively. \model's versatility enables navigation using diverse input modalities, including categories, language descriptions, and reference images. These findings highlight the importance of bridging visual grounding and exploration for embodied intelligence.

cs.CV↗