VectorGym: A Multitask Benchmark for SVG Code Generation, Sketching, and Editing
为解决SVG生成、编辑等缺乏专业级基准问题，提出VectorGym，采用多任务强化学习方法优化，并提供人类标注数据和评估指标。
3 citationsRead paper
Selected work
Representative Papers
CSC: Calibrated Simplicity for Conflict-Aware Social Bot Detection in the LLM Era
为解决大型语言模型时代社交机器人检测难题，提出CSC框架，通过简化图专家、校准融合及轻量级不一致专家提高检测准确性。
0 citationsRead paper
REARL: A Closed-loop Autonomous Driving Simulation Enhancement Framework with Real Traffic Data and Large Language Models
为解决自动驾驶模拟中真实交通复杂性捕捉难题，REARL框架结合真实交通数据与大语言模型动态调整模拟配置，有效减少速度分布和车辆间距的误差。
0 citationsRead paper
SA-Bench: Evaluating Semantic Alignment in LLM-Based Paper Reproduction
本文针对LLM生成的论文复现代码科学不忠实的问题，提出SA-Bench方法，通过定义和评估语义对齐单元来诊断和量化代码与原论文之间的偏差。
0 citationsRead paper
ReproAgent: Contract-Guided Paper-to-Code Reproduction
研究解决了论文到代码复现难题，通过引入ReproAgent，一个基于持续实现合同的四阶段管道，利用实现需求和参考证据两个通道来生成和修复代码。
0 citationsRead paper
Recent publications
Latest Papers
CSC: Calibrated Simplicity for Conflict-Aware Social Bot Detection in the LLM Era
为解决大型语言模型时代社交机器人检测难题，提出CSC框架，通过简化图专家、校准融合及轻量级不一致专家提高检测准确性。
0 citationsRead paper
REARL: A Closed-loop Autonomous Driving Simulation Enhancement Framework with Real Traffic Data and Large Language Models
为解决自动驾驶模拟中真实交通复杂性捕捉难题，REARL框架结合真实交通数据与大语言模型动态调整模拟配置，有效减少速度分布和车辆间距的误差。
0 citationsRead paper
SA-Bench: Evaluating Semantic Alignment in LLM-Based Paper Reproduction
本文针对LLM生成的论文复现代码科学不忠实的问题，提出SA-Bench方法，通过定义和评估语义对齐单元来诊断和量化代码与原论文之间的偏差。
0 citationsRead paper
ReproAgent: Contract-Guided Paper-to-Code Reproduction
研究解决了论文到代码复现难题，通过引入ReproAgent，一个基于持续实现合同的四阶段管道，利用实现需求和参考证据两个通道来生成和修复代码。
0 citationsRead paper
Beyond Peak Backlog: Conditional Energy and Temporal Geometry in Capacity-Constrained Delayed Bandit Optimization
This study addresses the dependence of regret bounds on peak backlog in capacity-constrained multi-armed bandits with delayed feedback. We propose a scheduler-side conditional energy interface that decouples rate adaptation from disturbance filtering to optimize delay complexity. By revealing how temporal geometry influences regret, we demonstrate that minimax regret can differ polynomially under identical delay statistics, thereby overcoming the limitations of aggregated metrics. Leveraging semi-transparent oracles and strong convexity analysis, we derive a parameter-free regret bound with a delay term of o(√(e_c d_tot)) and establish a lower bound for capacity-scarce regimes. These contributions significantly refine the theoretical granularity of regret analysis in delayed feedback settings beyond conventional aggregate measures.
0 citationsRead paper