arXiv subjects

Yiting Wu

Publications and source records attributed to Yiting Wu.

14 recordsLinked to original sources

arXiv2026

GVR-Coder: A Visual-Feedback Framework for Structured SVG Generation in Complex Document and Meeting Scenarios

In demanding professional environments and meeting review scenarios, lengthy text often imposes a high cognitive load. To facilitate efficient information communication, transforming verbose text into logically clear diagrams is essential. Scalable Vector Graphics (SVG) provide an effective representation for this purpose due to their editability and resolution independence. However, current research on Text-to-SVG generation remains hindered by three major challenges: (1) the scarcity of datasets for complex, logic-rich diagrams; (2) the absence of explicit layout priors, which leads to chaotic spatial arrangements; and (3) the lack of fine-grained visual feedback to validate rendered outputs and correct aesthetic defects. To address these challenges, at the data level, we introduce DocMeetSVG-100K, a large-scale SVG dataset tailored for document authoring and meeting review scenarios. At the model level, we propose GVR-Coder, a novel framework designed to generate high-quality logical diagrams from lengthy professional texts. Specifically, we adopt a curriculum-driven rejection sampling fine-tuning to progressively enhance the model's capability in modeling complex structures, while explicitly incorporating layout constraint knowledge during training. In addition, we introduce reinforcement learning from dual rendering feedback, a mechanism that provides implicit feedback through reward signals to jointly optimize structural complexity and visual aesthetics. Furthermore, we design a generate-verify-repair agent loop, which improves generation quality through explicit, fine-grained feedback and targeted refinement. Extensive experiments demonstrate that GVR-Coder outperforms competitive baselines and reliably produces logically coherent and visually appealing diagrams. Code and data are available at https://github.com/CurryaNa/GVR-Coder.

cs.LG↗

arXiv2026

Hollow Needle Puncture Mechanics for Biopsy Sampling

Biopsy sampling relies on hollow needles that puncture soft tissues by propagating and opening a cylindrical crack, yet the mechanics governing this coring process remain only partially understood. Motivated by this gap, we develop a simple, energy based model for puncture by blunt hollow needles, grounded in brittle fracture mechanics and extended to include frictional interactions at the needle tissue interface. The model describes puncture as the competition between the fracture energy and the elastic energy. This energetic balance is controlled by the interplay among needle geometry (radius and wall thickness), material properties (toughness and elastic modulus), and interfacial parameters (adhesion and friction). This model provides semi-analytical predictions for five key quantities: core size, frictionless force, frictional force slope, critical insertion depth, and critical insertion force. Model predictions are validated against experiments, demonstrating that friction significantly improves force estimation and alters the puncture regime. These results offer quantitative insight into the mechanics of tissue coring and force generation during biopsy, providing a predictive foundation for needle design, sampling performance, and real time control in robotic biopsy and needle insertion systems.

cond-mat.soft↗

arXiv2024

Deploying Multi-task Online Server with Large Language Model

In the industry, numerous tasks are deployed online. Traditional approaches often tackle each task separately by its own network, which leads to excessive costs for developing and scaling models, especially in the context of large language models. Although multi-task methods can save costs through parameter sharing, they often struggle to outperform single-task methods in real-world applications. To tackle these challenges, we present a three-stage multi-task learning framework for large language models. It involves task filtering, followed by fine-tuning on high-resource tasks, and finally fine-tuning on all tasks. We conducted comprehensive experiments in single-task and multi-task settings. Our approach, exemplified on different benchmarks, demonstrates that it is able to achieve performance comparable to the single-task method while reducing up to 90.9\% of its overhead.

cs.CL↗

arXiv2024

Nonstationary Discounted Stochastic Games under Prospect Theory with Applications to the Smart Grid

This paper considers the discounted criterion of nonzero-sum decentralized stochastic games with prospect players. The state and action spaces are finite. The state transition probability is nonstationary. Each player independently controls their own Markov chain. The subjective behavior of players is described by the prospect theory (PT). Compared to the average criterion of stochastic games under PT studied firstly in 2018, we are concerned with the time value of utility, i.e., the utility should be discounted in the future. Since PT distorts the probability, the optimality equation that plays a significant role in proving the existence of equilibrium does not exist. On the other hand, the games change into Markov decision processes (MDPs) with nonstationary payoff function when fixing others' stationary Markov strategies, then the occupation measure and the linear programming of stationary MDPs are no longer suitable. Therefore, we explore a new technique by constructing the marginal distribution on the state-action pairs at any time, and establish the existence of Nash equilibrium. When the initial state is given, there exists a Markov Nash equilibrium. Furthermore, this novel technique can be extended to the finite horizon criterion. Then, we present an algorithm to find a Markov $\varepsilon$-equilibrium. Finally, the model is applied to a noncooperative stochastic game among prosumers who can produce and consume energy in the smart grid, and we give some simulation results.

math.OC↗

arXiv2023

AgentSims: An Open-Source Sandbox for Large Language Model Evaluation

With ChatGPT-like large language models (LLM) prevailing in the community, how to evaluate the ability of LLMs is an open question. Existing evaluation methods suffer from following shortcomings: (1) constrained evaluation abilities, (2) vulnerable benchmarks, (3) unobjective metrics. We suggest that task-based evaluation, where LLM agents complete tasks in a simulated environment, is a one-for-all solution to solve above problems. We present AgentSims, an easy-to-use infrastructure for researchers from all disciplines to test the specific capacities they are interested in. Researchers can build their evaluation tasks by adding agents and buildings on an interactive GUI or deploy and test new support mechanisms, i.e. memory, planning and tool-use systems, by a few lines of codes. Our demo is available at https://agentsims.com .

cs.AI↗

arXiv2023

A Tale of Two Approximations: Tightening Over-Approximation for DNN Robustness Verification via Under-Approximation

The robustness of deep neural networks (DNNs) is crucial to the hosting system's reliability and security. Formal verification has been demonstrated to be effective in providing provable robustness guarantees. To improve its scalability, over-approximating the non-linear activation functions in DNNs by linear constraints has been widely adopted, which transforms the verification problem into an efficiently solvable linear programming problem. Many efforts have been dedicated to defining the so-called tightest approximations to reduce overestimation imposed by over-approximation. In this paper, we study existing approaches and identify a dominant factor in defining tight approximation, namely the approximation domain of the activation function. We find out that tight approximations defined on approximation domains may not be as tight as the ones on their actual domains, yet existing approaches all rely only on approximation domains. Based on this observation, we propose a novel dual-approximation approach to tighten over-approximations, leveraging an activation function's underestimated domain to define tight approximation bounds. We implement our approach with two complementary algorithms based respectively on Monte Carlo simulation and gradient descent into a tool called DualApp. We assess it on a comprehensive benchmark of DNNs with different architectures. Our experimental results show that DualApp significantly outperforms the state-of-the-art approaches with 100% - 1000% improvement on the verified robustness ratio and 10.64% on average (up to 66.53%) on the certified lower bound.

cs.LG↗

arXiv2022

DualApp: Tight Over-Approximation for Neural Network Robustness Verification via Under-Approximation

The robustness of neural networks is fundamental to the hosting system's reliability and security. Formal verification has been proven to be effective in providing provable robustness guarantees. To improve the verification scalability, over-approximating the non-linear activation functions in neural networks by linear constraints is widely adopted, which transforms the verification problem into an efficiently solvable linear programming problem. As over-approximations inevitably introduce overestimation, many efforts have been dedicated to defining the tightest possible approximations. Recent studies have however showed that the existing so-called tightest approximations are superior to each other. In this paper we identify and report an crucial factor in defining tight approximations, namely the approximation domains of activation functions. We observe that existing approaches only rely on overestimated domains, while the corresponding tight approximation may not necessarily be tight on its actual domain. We propose a novel under-approximation-guided approach, called dual-approximation, to define tight over-approximations and two complementary under-approximation algorithms based on sampling and gradient descent. The overestimated domain guarantees the soundness while the underestimated one guides the tightness. We implement our approach into a tool called DualApp and extensively evaluate it on a comprehensive benchmark of 84 collected and trained neural networks with different architectures. The experimental results show that DualApp outperforms the state-of-the-art approximation-based approaches, with up to 71.22% improvement to the verification result.

cs.SE↗

arXiv2022

Provably Tightest Linear Approximation for Robustness Verification of Sigmoid-like Neural Networks

The robustness of deep neural networks is crucial to modern AI-enabled systems and should be formally verified. Sigmoid-like neural networks have been adopted in a wide range of applications. Due to their non-linearity, Sigmoid-like activation functions are usually over-approximated for efficient verification, which inevitably introduces imprecision. Considerable efforts have been devoted to finding the so-called tighter approximations to obtain more precise verification results. However, existing tightness definitions are heuristic and lack theoretical foundations. We conduct a thorough empirical analysis of existing neuron-wise characterizations of tightness and reveal that they are superior only on specific neural networks. We then introduce the notion of network-wise tightness as a unified tightness definition and show that computing network-wise tightness is a complex non-convex optimization problem. We bypass the complexity from different perspectives via two efficient, provably tightest approximations. The results demonstrate the promising performance achievement of our approaches over state of the art: (i) achieving up to 251.28% improvement to certified lower robustness bounds; and (ii) exhibiting notably more precise verification results on convolutional networks.

cs.LG↗

arXiv2021

Quantum Secure Multi-party Summation Based on entanglement swapping

In this paper, we present a quantum secure multi-party summation protocol, which allows multiple mutually distrustful parties to securely compute the summation of their secret data. In the presented protocol, a semitrusted third party is introduced to help multiple parties to achieve this secure task. Besides, the entanglement swapping of $d$-level cat states and Bell states is employed to securely transmit message between each party and the semitrusted third party. At last, its security against some common attacks is analyzed, which shows that the presented protocol is secure in theory.

quant-ph↗

arXiv2021

Multi-party quantum key agreement protocol with authentication

Utilizing the advantage of quantum entanglement swapping, a multi-party quantum key agreement protocol with authentication is proposed. In this protocol, a semi-trusted third party is introduced, who prepares Bell states, and sends one particle to multiple participants respectively. After that the participants can share a Greenberger-Horne-Zeilinger state by entanglement swapping. Finally, these participants measure the particles in their hands and obtain an agreement key. Here, classical hash function and Hadamard operation are utilized to authenticate the identity of participants. The correlations of GHZ states ensure the security of the proposed protocol. To illustrated it detailly, the security of this protocol against common attacks is analyzed, which shows that the proposed protocol is secure in theory.

quant-ph↗

arXiv2020

Blow-up for a semilinear heat equation with Fujita's critical exponent on locally finite graphs

Let $G=(V,E)$ be a locally finite, connected and weighted graph. We prove that, for a graph satisfying curvature dimension condition $CDE'(n,0)$ and uniform polynomial volume growth of degree $m$, all non-negative solutions of the equation $\partial_tu=Δu+u^{1+α}$ blow up in a finite time provided that $α=\frac{2}{m}$. We also consider the blow-up problem under certain conditions for volume growth and initial value. The obtained results provide a significant complement to the work by Lin and Wu in earlier paper.

math.AP↗

arXiv2017

Blow-up problems for nonlinear parabolic equations on locally finite graphs

Let $G=(V,E)$ be a locally finite connected weighted graph, $Δ$ be the usual graph Laplacian. In this paper, we study the blow-up problems for the nonlinear parabolic equation $u_t=Δu + f(u)$ on $G$. The blow-up phenomenons of the equation are discussed in terms of two cases: (i) an initial condition is given; (ii) a Dirichlet boundary condition is given. We prove that if $f$ satisfies appropriate conditions, then the solution of the equation blows up in a finite time.

math.AP↗

arXiv2017

The existence and nonexistence of global solutions for a semilinear heat equation on graphs

Let $G=(V,E)$ be a finite or locally finite connected weighted graph, $Δ$ be the usual graph Laplacian. Using heat kernel estimate, we prove the existence and nonexistence of global solutions for the following semilinear heat equation on $G$ \begin{equation*} \left\{ \begin{array}{lc} u_t=Δu + u^{1+α} &\, \text{in $(0,+\infty)\times V$,}\\ u(0,x)=a(x) &\, \text{in $V$.} \end{array} \right. \end{equation*} We conclude that, for a graph satisfying curvature dimension condition $CDE'(n,0)$ and $V(x,r)\simeq r^m$, if $0 2$, then there is a non-negative global solution $u$ provided that the initial value is small enough. In particular, these results are true on lattice $\mathbb{Z}^m$.

math.AP↗

arXiv2016

On-diagonal lower estimate of heat kernel on graphs

The purpose of this paper is to establish a new continuous-time on-diagonal lower estimate of heat kernel for large time on graphs. To achieve the goal, we first give an upper bound of heat kernel in natural graph metric, and then use this bound and the volume growth condition to show the validity of the on-diagonal lower bound.

math.AP↗