👋 Hello, I’m Maocai Dai👋 你好,我是戴茂才

I am Maocai Dai (戴茂才), a master’s student at the Chongqing University of Posts and Telecommunications (CQUPT), advised by Prof. Ye Wang. My research focuses on Image Aesthetic Assessment (IAA) and Image Quality Assessment (IQA), with an interest in building visual models that better align with human perception and subjective preference.我是戴茂才,现为重庆邮电大学计算机科学与技术专业硕士研究生,导师为王烨教授。我的研究方向包括图像美学评估(IAA)与图像质量评估(IQA),致力于构建更贴合人类感知与主观偏好的视觉模型。

You can reach me at dahat364@gmail.com or dahat123@163.com. My code and projects are available on GitHub.你可以通过 dahat364@gmail.com 或 dahat123@163.com 联系我;代码与项目请见 GitHub。

🔥 News🔥 新闻动态

  • 2026.05 — 🎉 Our paper When Attributes Disagree: Gradient Conflict in Image Aesthetic Assessment was selected as an ICML 2026 Spotlight paper.🎉 论文《When Attributes Disagree: Gradient Conflict in Image Aesthetic Assessment》入选 ICML 2026 Spotlight。

🔬 Research Interests🔬 研究方向

  • Image Aesthetic Assessment: modeling image aesthetics through composition, style, attributes, and subjective preference.图像美学评估:从构图、风格、属性与主观偏好等角度建模图像美学。
  • Image Quality Assessment: estimating perceptual visual quality and building robust quality representations.图像质量评估:估计感知视觉质量,并构建鲁棒的质量表征。
  • Human-aligned Vision: connecting computational image understanding with nuanced human visual judgments.人类对齐视觉:连接计算机图像理解与细腻的人类视觉判断。

📝 Publications📝 论文成果

ICML 2026 · Spotlight
Overview of the AGREE framework

When Attributes Disagree: Gradient Conflict in Image Aesthetic Assessment当属性产生分歧:图像美学评估中的梯度冲突

Ye Wang, Maocai Dai, Jiang Xie, Xiuli Bi*, Fei Tao, Xiao Li, Hong Yu

International Conference on Machine Learning (ICML), 2026 · Spotlight国际机器学习大会(ICML),2026 · Spotlight

We propose AGREE, a plug-and-play framework that mitigates attribute-level gradient conflicts in Image Aesthetic Assessment through attribute-aware routing. It consistently improves diverse IAA baselines across five benchmarks.我们提出 AGREE,一种可插拔框架,通过属性感知路由缓解图像美学评估中的属性级梯度冲突,并在五个基准上持续提升多种 IAA 基线模型的表现。

Project Page项目主页

💼 Service💼 学术服务

Teaching Assistant助教

  • Data Structures, School of Computer Science and Technology, Chongqing University of Posts and Telecommunications.数据结构,重庆邮电大学计算机科学与技术学院。
  • Natural Language Processing, School of Computer Science and Technology, Chongqing University of Posts and Telecommunications.自然语言处理,重庆邮电大学计算机科学与技术学院。

🎖 Honors and Awards🎖 荣誉与获奖

  • 2026.10 China National Scholarship for Graduate Students (2025–2026).研究生国家奖学金(2025–2026 学年)。
  • 2026.05 ICML 2026 Spotlight Paper — When Attributes Disagree: Gradient Conflict in Image Aesthetic Assessment.ICML 2026 Spotlight 论文——When Attributes Disagree: Gradient Conflict in Image Aesthetic Assessment。
  • 2025.10 The Second Scholarship offered by Chongqing University of Posts and Telecommunications.重庆邮电大学二等奖学金。
  • 2023.12 The China National Encouragement Scholarship (2022–2023).国家励志奖学金(2022–2023 学年)。
  • 2023.10 The Second Scholarship offered by Chongqing University of Posts and Telecommunications.重庆邮电大学二等奖学金。
  • 2023.02 The Excellent Student Cadre offered by Chongqing University of Posts and Telecommunications.重庆邮电大学优秀学生干部。
  • 2022.12 The China National Encouragement Scholarship (2021–2022).国家励志奖学金(2021–2022 学年)。
  • 2022.10 The Second Scholarship offered by Chongqing University of Posts and Telecommunications.重庆邮电大学二等奖学金。
  • 2022.02 The Excellent Student Cadre offered by Chongqing University of Posts and Telecommunications.重庆邮电大学优秀学生干部。
  • 2021.10 The Third Scholarship offered by Chongqing University of Posts and Telecommunications.重庆邮电大学三等奖学金。

🤝 Collaborators🤝 合作者

  • Guanmeng Xian (贤冠萌) A longtime friend and fellow researcher from our undergraduate days at CQUPT.本科时期相识的挚友与科研同行者。
  • Haokun Ren (任豪坤) A fellow graduate student and research colleague in our group at CQUPT.研究生阶段的同门与科研伙伴。
  • Xuyang Zhou (周绪洋) A fellow researcher and collaborator in our group at CQUPT.同组低一届的师弟与科研伙伴。

📖 Educations📖 教育经历

  • 2024.09 – Present, M.Sc. student in Computer Science and Technology, Chongqing University of Posts and Telecommunications. Advisor: Prof. Ye Wang.重庆邮电大学计算机科学与技术专业硕士研究生。导师:王烨教授。
  • 2020.09 – 2024.06, B.Eng. in Intelligent Science and Technology, Chongqing University of Posts and Telecommunications.重庆邮电大学智能科学与技术专业,工学学士。