武亚强

武亚强 (Adam Wu)

邮箱： wuyqe@lenovo.com

谷歌学术： Google Scholar

领英： LinkedIn

我是一名的资深人工智能研究与落地领导者，深耕人工智能、机器学习、多媒体信息处理及大语言模型（LLM）领域的创新落地，累计授权发明专利超百项，4 项专利获联想「最有价值专利（MVP）」，目前在联想人工智能技术中心带领中美团队，推动 AI 技术在全球企业场景中实现高价值落地。

最新动态

2026年3月29日

与西安交通大学联合申报的神经-符号协同的抽象图文推理技术与应用荣获吴文俊人工智能科学技术进步奖二等奖。

2026年1月8日

联想在CES发布首款个人超级AI智能体Qira，3大核心技术中「模型编排」「多智能体协作」2项由我团队研发。 [链接]

2025年11月10日

西交-联想产业AI实验室3篇论文被AAAI 2026录用，其中联想主导1篇，西安交大主导2篇。

任职与学会

现任职务

  • 联想集团 高级总监，人工智能中心企业AI负责人
  • 西安交通大学 兼职教授，行业智能实验室主任
  • 中国科学院 正高级工程师

专业学会

  • 中国计算机学会 杰出会员
  • 电气电子工程师学会 高级会员
  • 中国计算机学会 多媒体技术专委会 常务委员
  • 中国图象图形学学会 文档分析与识别专委会 常务委员

核心技术方向

核心技术方向包括多媒体技术、人机交互、人工智能，深耕企业级AI领域。技术专长覆盖：

  • 大语言模型（LLM/AIGC）：含LLM微调、提示工程
  • 数字人与虚实融合（XR）：空间计算、多设备自然融合交互
  • 多媒体信息处理：计算机视觉、深度学习
  • AI架构：端云一体化平台、异构知识系统

教育经历

西安交通大学

学位：博士学位、硕士学位、学士学位  |  专业：计算机与信息工程

Adam Wu

Adam Wu (武亚强)

Email: wuyqe@lenovo.com

Google Scholar: Google Scholar

LinkedIn: LinkedIn

I am a highly accomplished AI Engineering leader in driving innovation across HCI, Multimedia Information Processing, and Large Language Models (LLM). I lead the enterprise AI at Lenovo AI Technology Center and work with my team to transform cutting-edge research into high-value products.

Latest News

Mar 29, 2026

The neural-symbolic collaborative abstract visual-text reasoning technology and its applications, jointly declared with Xi'an Jiaotong University, won the Second Prize of the Wu Wenjun Artificial Intelligence Science and Technology Progress Award.

Jan 8, 2026

Lenovo launched its first personal super AI agent Qira at CES; 2 of the 3 core technologies—Model Orchestration and Multi-agent Collaboration—were developed by my team. [Link]

Nov 10, 2025

3 papers from the XJTU-Lenovo Industrial AI Lab were accepted by AAAI 2026 (1 led by Lenovo, 2 led by Xi'an Jiaotong University).

Positions & Societies

Current Positions

  • Lenovo Group AI Technology Center, Excutive Director
  • Xi'an Jiaotong University, School of Computer Science, Adjunct Professor
  • Chinese Academy of Sciences, Professor-level Senior Engineer

Professional Societies

  • China Computer Federation (CCF), Distinguished Member
  • IEEE, Senior Member
  • Chinese Association for Artificial Intelligence (CAAI), Senior Member
  • CCF Multimedia Technology Committee, Standing Committee Member
  • CSIG Document Analysis and Recognition Committee, Standing Committee Member

Core Technical Directions

Core technical directions include multimedia technology, human-computer interaction, and artificial intelligence, with a focus on enterprise AI. Areas of expertise include:

  • Large Language Models (LLM/AIGC): LLM fine-tuning, prompt engineering
  • Digital Human and Virtual-Real Fusion (XR): spatial computing, multi-device natural fusion interaction
  • Multimedia Information Processing: computer vision, deep learning
  • AI Architecture: end-cloud integrated platforms, heterogeneous knowledge systems

Education

Xi'an Jiaotong University

Degrees: Ph.D., M.S., B.S.  |  Major: Computer and Information Engineering

最新动态

2025

10月21日

「超写实数字人技术」项目顺利通过北京市科技计划综合绩效评价。

5月9日

受邀在2025中国图象图形大会「文字识别与文档智能」论坛作题为“从任务专用到通用化：基础模型时代的图像分割演进”的特邀报告。

2024

9月25日

受邀在首届中国数字人大会主论坛做特邀报告：数字化伙伴：生成式技术开拓数字人新纪元。

6月25日

自研向量模型Zhihui_LLM_Embedding以76.7分登顶MTEB检索榜单（全球第一）。

3月31日

受邀在首届中国具身智能大会（CEAI 2024）「具身智能与智能汽车」分论坛做技术报告。

3月17日

「智慧教学环境构建中虚实融合关键技术及重大应用」项目获2023年产学研合作创新成果一等奖。

1月9日

团队研发的车载智能助手在CES 2024首发，搭载自研数字人生成、知识增强车载大模型。

2023

9月27日

「数实融合精准导学关键技术及应用」项目获CCF科技进步一等奖。

7月12日

受邀在首届CCF黄河论坛「AI+硬科技」产业创新分论坛做技术报告。

6月19日

团队斩获国际文档分析与识别大会（ICDAR）主办的通用文档理解（DUDE）学术竞赛冠军。

2022

12月27日

团队智慧教育解决方案获评量子位年度AI最佳解决方案TOP 10。

Latest News

2025

Oct 21

The "Ultra-realistic Digital Human Technology" project successfully passed the comprehensive performance evaluation of Beijing Science and Technology Program.

May 9

Invited to deliver a keynote speech titled "From Task-Specific to Generalization: Evolution of Image Segmentation in the Foundation Model Era" at the "Text Recognition and Document Intelligence" forum of China Graphics Conference 2025.

2024

Jun 25

Self-developed vector model Zhihui_LLM_Embedding scored 76.7, topping the MTEB Retrieval Leaderboard (1st Place).

Mar 31

Invited to deliver a technical report at the "Embodied AI and Smart Vehicles" sub-forum at the 1st China Embodied AI Conference (CEAI 2024).

Mar 17

Project "Key Technology and Major Applications of Virtual-Real Fusion in Smart Teaching Environment Construction" won First Prize of the 2023 Industry-University-Research Cooperation Innovation Award.

Jan 9

Team's R&D of the In-Vehicle Smart Assistant debuted at CES 2024, featuring self-developed Digital Human generation and knowledge-enhanced in-car LLM.

2023

Sep 27

Project "Digital-Real Fusion for Precise Guided Learning Key Technologies and Applications" won First Prize of the CCF Science and Technology Progress Award.

Jun 19

Team won the Document Understanding of Everything (DUDE) academic competition organized by ICDAR (International Conference on Document Analysis and Recognition).

2022

Dec 27

Team's Smart Education Solution was named a Qubit Annual AI Best Solution TOP 10.

项目与荣誉

主导项目

2024至今 公司AI战略拆解执行，从0开始构建联想AI多智能体与大模型编排平台研发

  • 核心方向： 主导大模型/AIGC在核心AI智能体平台的研发与应用。
  • 关键成果：
    • 领导和设计公司多智能体协作统一架构，实现最大化ROI的多智能体自主规划与执行
    • 技术成果已落地联想AI PC、混合云等个人及企业级AI项目，赋能全场景智能应用。
    • MTEB榜单第一： 自研向量模型 Zhihui_LLM_Embedding 以76.7分登顶2024年6月MTEB检索榜单。

2018至2023 从0到1构建并规模化落地联想智慧教育一体化解决方案

  • 角色： 研发负责人、首席架构师
  • 落地规模： 全球部署至 1万+所学校，学生学习效率提升 10%+。
  • 奖项： 项目获2024年中国产学研合作创新成果一等奖。

2016至2023 联想Document AI核心技术攻关与全产品线落地

  • 技术突破： 攻克复杂文档识别、屏幕文本理解、高精度文档定位等核心技术，中文识别精度达业界第一。
  • 赛事成果： 斩获 ICDAR、ICPR 等国际顶级文档智能赛事10余项全球第一名，其中含ICDAR 2023 DUDE通用文档理解竞赛冠军。
  • 产业落地： 相关Document AI技术已全面落地联想全线产品，实现技术规模化商业应用。

获奖与荣誉

团队荣誉

  • 带领团队斩获 ICDAR 国际文档分析与识别竞赛 10 项荣誉
  • 荣获中国产学研合作创新成果一等奖，中国计算机学会科技进步一等奖、吴文俊人工智能科技奖二等奖等省部级学会奖项 5 项
  • 2022 Fast Company，Next Big Thing in Tech，Winner
  • 2022 量子位，人工智能十佳解决方案

个人荣誉

  • 2024 中国产学研合作创新成果一等奖（第1）
  • 2023 中国计算机学会科技进步一等奖（第3）
  • 2021 北京市科学技术进步奖二等奖（第6）
  • 2020 人工智能学会吴文俊人工智能科技进步奖二等奖（第3）
  • 2006～至今，联想研究院：获得 2 次联想研究院杰出领导者（2016，2018），带领团队获得 1 次联想集团优秀团队（2021），5 次联想研究院杰出团队（2016，2017，2020，2022，2023）
  • 2019～至今，10 项国际文档分析与识别竞赛冠军（ICDAR2019，ICDAR2021，ICPR2020，ICDAR2023）

Projects & Honors

As the technical lead and chief architect, my projects have achieved significant commercial and societal impact

Key Projects

2024-Present: Lenovo AI Strategy Execution & AI Agent/LLM Orchestration Platform Development

  • Focus: Leading R&D and application of LLMs/AIGC in the core AI Agent platform.
  • Achievements:
    • MTEB Ranking #1: Self-developed vector model Zhihui_LLM_Embedding scored 76.7, topping the MTEB Retrieval Leaderboard (June 2024).
    • Technologies have been deployed in Lenovo AI PC, Hybrid Cloud and other personal/enterprise AI projects, empowering full-scenario intelligent applications.

2018-2023: Lenovo Smart Education Integrated Solution (From 0 to Large-Scale Deployment)

  • Role: R&D Lead, Chief Architect
  • Impact Metrics: Deployed to 10,000+ schools globally, improving student learning efficiency by 10%+.
  • Awards: Won the First Prize of China Industry-University-Research Cooperation Innovation Award (2024).

2016-2023: Lenovo Document AI Core Technology R&D & Full Product Line Deployment

  • Technical Breakthroughs: Overcame core technologies such as complex document recognition, screen text understanding, and high-precision document positioning, with industry-leading Chinese recognition accuracy.
  • Competition Achievements: Won 10+ first prizes in international top document intelligence competitions (ICDAR, ICPR), including the ICDAR 2023 DUDE championship.
  • Industrial Deployment: Related Document AI technologies have been fully deployed across Lenovo's entire product line, achieving large-scale commercial application.

Awards & Honors

Team Honors

  • Led the team to 10 ICDAR (International Conference on Document Analysis and Recognition) awards
  • First Prize of China Industry-University-Research Cooperation Innovation Award, First Prize of CCF Science and Technology Progress Award, Second Prize of Wu Wenjun AI Science and Technology Award, and 5 other provincial/society-level awards
  • Over 140 patent applications and grants, including 17 U.S. patents; 4 patents awarded Lenovo "Most Valuable Patent (MVP)"
  • Technical achievements recognized by Fast Company "Next Big Thing in Tech," Qubit "AI Top 10 Solutions," and other international and industry accolades

Personal Honors

  • 2024 First Prize, China Industry-University-Research Cooperation Innovation Award
  • 2023 First Prize, CCF Science and Technology Progress Award
  • 2022 Fast Company, Next Big Thing in Tech, Winner
  • 2022 Qubit, AI Top 10 Solutions
  • 2021 Beijing Municipal Science and Technology Progress Award, Second Prize
  • 2020 CAAI Wu Wenjun AI Science and Technology Progress Award, Second Prize
  • 2006–Present, Lenovo Research: 2× Lenovo Research Outstanding Leader (2016, 2018); led team to 1× Lenovo Group Outstanding Team (2021), 5× Lenovo Research Outstanding Team (2016, 2017, 2020, 2022, 2023)
  • 2019–Present, 10 International Document Analysis and Recognition Competition championships (ICDAR 2019, ICDAR 2021, ICPR 2020, ICDAR 2023)

论文与专利

代表论文

以下为部分代表论文，完整列表见 Google Scholar。

SketchVL: Policy Optimization via Fine-Grained Credit Assignment for Chart Understanding and More

M Huang, L Zhang, Y Li, Y Wu, J Liu. arXiv preprint arXiv:2601.05688, 2026.

Empowering multimodal llms with external tools: A comprehensive survey

W An, J Nie, Y Wu, F Tian, S Lu, Q Zheng. arXiv preprint arXiv:2508.10955, 2025.

Chartsketcher: Reasoning with multimodal feedback and reflection for chart understanding

M Huang, L Zhang, J Ma, H Lai, F Xu, Y Li, W Wu, Y Wu, J Liu. arXiv preprint arXiv:2505.19076, 2025.

Unleashing the potential of model bias for generalized category discovery

W An, H Lin, J Nie, F Tian, W Shi, Y Wu, Q Wang, P Chen. Proceedings of the AAAI Conference on Artificial Intelligence 39 (15), 15365, 2025.

Using Depth-Enhanced Spatial Transformation for Student Gaze Target Estimation in Dual-View Classroom Images

H Miao, P Zhao, Y Sun, F Nan, X Zhang, Y Wu, F Tian. ICASSP 2025, 2025.

LFSRM: Few-Shot Diagram-Sentence Matching via Local-Feedback Self-Regulating Memory

L Zhang, W Wu, J Liu, X Chang, X Hu, Y Zheng, Y Wu, Q Zheng. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2025.

Flipped classroom: Aligning teacher attention with student in generalized category discovery

H Lin, W An, J Wang, Y Chen, F Tian, M Wang, QY Wang, G Dai, J Wang. Advances in Neural Information Processing Systems 37, 60897-60935, 2024.

MRCI: Multi-range Context Interaction for Boundary Refinement in Image Segmentation

Y Wu, W Lyu, X Liang, Q Zheng, J Wei, L Jin. International Conference on Pattern Recognition, 211-226, 2024.

RDLNet: a novel and accurate real-world document localization method

Y Wu, Z Xu, Y Duan, Y Wu, Q Zheng, H Li, X Hu, L Jin. Proceedings of the 32nd ACM International Conference on Multimedia, 9847-9855, 2024.

Programming knowledge tracing based on heterogeneous graph representation

Y Wu, H Zhu, C Wang, F Song, H Zhu, Y Chen, Q Zheng, F Tian. Knowledge-Based Systems 300, 112161, 2024.

部分专利

申请与授权发明专利超过 140 余项，4 次获得联想集团最有价值专利奖。

授权日期专利名称专利编号发明人
2023.05.23图像显示控制方法和装置202310185874.8武亚强等
2023.04.11一种显示处理方法、装置和电子设备202211639256.8武亚强等
2023.02.17学习视频生成方法、装置、电子设备及存储介质CN202111448184.4武亚强等
2022.11.22学生使用电子设备的控制方法、装置及电子设备CN201911358291.0张晓平，武亚强等
2022.08.19一种多媒体数据同步方法及装置、设备CN202010616703.2张晓平，武亚强等
2022.04.22一种内容显示方法及装置CN201811150283.2陈宏星，武亚强等
2021.10.20教室终端系统及其控制方法、控制器与主控设备CN202010128622.8张晓平，武亚强等
2021.05.18一种声音信息的转换方法、装置及设备CN201710465049.8白金才，武亚强等
2020.08.04Information processing method and apparatus, and electronic device and computer readable medium thereofUS10735918B2Yingwen, Yaqiang Wu, et al
2019.10.29一种数据处理方法及电子设备CN201410413299.3武亚强等
2019.02.19Method and apparatus for file processingUS10210148B2Yaqiang Wu, et al
2018.07.03Display method and display deviceUS10013730B2Yaqiang Wu, et al

Papers and Patents

Representative Papers

Selected publications below. Full list on Google Scholar.

SketchVL: Policy Optimization via Fine-Grained Credit Assignment for Chart Understanding and More

M Huang, L Zhang, Y Li, Y Wu, J Liu. arXiv preprint arXiv:2601.05688, 2026.

Empowering multimodal llms with external tools: A comprehensive survey

W An, J Nie, Y Wu, F Tian, S Lu, Q Zheng. arXiv preprint arXiv:2508.10955, 2025.

Chartsketcher: Reasoning with multimodal feedback and reflection for chart understanding

M Huang, L Zhang, J Ma, H Lai, F Xu, Y Li, W Wu, Y Wu, J Liu. arXiv preprint arXiv:2505.19076, 2025.

Unleashing the potential of model bias for generalized category discovery

W An, H Lin, J Nie, F Tian, W Shi, Y Wu, Q Wang, P Chen. Proceedings of AAAI 39 (15), 15365, 2025.

Using Depth-Enhanced Spatial Transformation for Student Gaze Target Estimation in Dual-View Classroom Images

H Miao, P Zhao, Y Sun, F Nan, X Zhang, Y Wu, F Tian. ICASSP 2025, 2025.

LFSRM: Few-Shot Diagram-Sentence Matching via Local-Feedback Self-Regulating Memory

L Zhang, W Wu, J Liu, X Chang, X Hu, Y Zheng, Y Wu, Q Zheng. IEEE TPAMI, 2025.

Flipped classroom: Aligning teacher attention with student in generalized category discovery

H Lin, W An, J Wang, Y Chen, F Tian, M Wang, QY Wang, G Dai, J Wang. NeurIPS 37, 60897-60935, 2024.

MRCI: Multi-range Context Interaction for Boundary Refinement in Image Segmentation

Y Wu, W Lyu, X Liang, Q Zheng, J Wei, L Jin. ICPR, 211-226, 2024.

RDLNet: a novel and accurate real-world document localization method

Y Wu, Z Xu, Y Duan, Y Wu, Q Zheng, H Li, X Hu, L Jin. ACM MM 2024, 9847-9855, 2024.

Programming knowledge tracing based on heterogeneous graph representation

Y Wu, H Zhu, C Wang, F Song, H Zhu, Y Chen, Q Zheng, F Tian. Knowledge-Based Systems 300, 112161, 2024.

Selected Patents

Over 140 patent applications and grants; 4× Lenovo Most Valuable Patent Award.

DateTitleNumberInventors
2023.05.23图像显示控制方法和装置202310185874.8武亚强等
2023.04.11一种显示处理方法、装置和电子设备202211639256.8武亚强等
2023.02.17学习视频生成方法、装置、电子设备及存储介质CN202111448184.4武亚强等
2022.11.22学生使用电子设备的控制方法、装置及电子设备CN201911358291.0张晓平，武亚强等
2022.08.19一种多媒体数据同步方法及装置、设备CN202010616703.2张晓平，武亚强等
2022.04.22一种内容显示方法及装置CN201811150283.2陈宏星，武亚强等
2021.10.20教室终端系统及其控制方法、控制器与主控设备CN202010128622.8张晓平，武亚强等
2021.05.18一种声音信息的转换方法、装置及设备CN201710465049.8白金才，武亚强等
2020.08.04Information processing method and apparatus, and electronic device and computer readable medium thereofUS10735918B2Yingwen, Yaqiang Wu, et al
2019.10.29一种数据处理方法及电子设备CN201410413299.3武亚强等
2019.02.19Method and apparatus for file processingUS10210148B2Yaqiang Wu, et al
2018.07.03Display method and display deviceUS10013730B2Yaqiang Wu, et al

企业博士后招聘

招聘信息

团队与院士团队联合招聘企业博士后。

研究方向

生成式AI（多模态大模型、交互数字人）。

待遇与发展

博士毕业生可选择回归学术界或留任产业界，CV/NLP 方向博士提供有竞争力的薪酬。

Postdoc Recruitment

Recruitment

Our team is jointly recruiting Enterprise Postdoctoral researchers with Academician teams.

Research Direction

Generative AI (Multimodal Large Models, Interactive Digital Human).

Compensation & Career

Graduates can choose to return to academia or remain in industry. Competitive compensation is offered for Ph.D. holders in CV/NLP.