Funding & Projects · Ming Wang

01 / Funded research

Funding科研资助

01Principal Investigator · 2026基于多智能体的火灾调查智能辅助询问系统研发辽宁省科技计划联合计划（自然科学基金-面上项目）子课题

Principal Investigator项目负责人

辽宁省科技计划联合计划（自然科学基金-面上项目）子课题

¥26,000

2026

02Technical Lead · 2025 — PresentBuilding Reflection Competencies for Human–AI Collaboration: A Multi-Agent Training SystemResilient Workforces Institute Grant

Technical Lead技术负责人

Resilient Workforces Institute Grant

S$70,000

2025 — Present

03Principal Investigator · 2025 -- 2026基于大型语言模型的火灾调查数字化智能案卷管理系统中央公益性研究所基本科研业务费项目(子课题)

Principal Investigator项目负责人

中央公益性研究所基本科研业务费项目(子课题)

20,000

2025 -- 2026

01PINNED · Core Contributor & Co-builderLangGPT: Rethinking Structured Reusable Prompt Design Framework for LLMs from the Programming Language核心贡献者 / 共建者12,475GitHub Stars

Core Contributor & Co-builder核心贡献者 / 共建者

A personified structural prompt framework inspired by reusable, object-oriented programming design. Contributed to its theoretical system and paper writing.受可复用、面向对象程序设计启发的人格化结构化提示框架；参与理论体系完善与论文写作。

Paper ↗Code ↗

02PINNED · Project LeadMinstrel: Structural Prompt Generation with Multi-Agents Coordination for Non-AI Experts项目负责人153GitHub Stars

Project Lead项目负责人

A multi-agent system that decomposes prompt generation across coordinated working groups and automates reusable structural prompt design.通过多个生成式智能体工作组协作分解提示生成任务，自动生成可复用的结构化提示。

Paper ↗Code ↗

03PINNED · Project LeadT-COL: Generating Counterfactual Explanations for General User Preferences on Variable Machine Learning Systems项目负责人2GitHub Stars

Project Lead项目负责人

A link-tree-based counterfactual explanation method that selects and recombines local feature subsets to reflect general user preferences.基于链接树选择并组合局部特征子集，面向通用用户偏好生成反事实解释。

Paper ↗Code ↗

04Key ContributorMM-BigBench核心参与者32GitHub Stars

Key Contributor核心参与者

A multimodal benchmark with complementary metrics for model performance, relative gains, stability, and adaptability.面向多模态模型的综合评测基准，覆盖最佳性能、相对增益、稳定性与适应性等指标。

Paper ↗Code ↗

05Key ContributorPICA核心参与者98GitHub Stars

Key Contributor核心参与者

A multi-turn dialogue model designed to provide emotionally engaged and empathetic responses; contributed to data construction and supervised fine-tuning.面向情感参与与共情回应的多轮对话模型；参与数据构建与监督微调。

Code ↗