TheBrave · Sun Yat-sen University · CS Sophomore

康凯

广袤天地大有可为

There is vast space ahead, and much worth doing.

我是中山大学计算机科学与技术专业大二学生，英文名 TheBrave，正在系统学习计算机基础与人工智能相关知识。我希望未来围绕 AI、具身智能与智能体系统持续深造，在更高密度的学术环境中继续积累研究能力与工程能力。

I am a sophomore majoring in Computer Science and Technology at Sun Yat-sen University, also known as TheBrave. I am building a solid foundation in computing and AI, and I hope to pursue deeper study in AI, embodied intelligence, and agent systems within a stronger research environment.

中山大学 · 计算机科学与技术 Sun Yat-sen University · Computer Science AI / Embodied Intelligence

查看项目方向View Projects 查看简历Open Resume 访问 GitHubVisit GitHub

About Me

把计算、感知与行动连接起来

Connecting computation, perception, and action.

你好，我是康凯，英文名 TheBrave，目前就读于中山大学计算机科学与技术专业。对我来说，计算机不仅是编程和系统，更是一种理解智能、构建智能的方式。我想做的不只是“会用模型”，而是逐步走近那些真正能把感知、推理与行动连接起来的智能系统。

Hello, I am Kang Kai, also known as TheBrave, currently studying Computer Science and Technology at Sun Yat-sen University. To me, computing is not only about programming and systems, but also a way to understand and build intelligence. I want to move beyond simply using models and get closer to systems that truly connect perception, reasoning, and action.

我特别关注人工智能与具身智能，因为它们不只是“让模型回答问题”，而是在研究一个系统如何感知环境、做出决策并与真实世界交互。未来我希望通过继续深造，进入更高密度的研究环境，在 AI 与 embodied intelligence 方向积累更扎实的能力。

I am especially interested in AI and embodied intelligence because they are not only about generating answers, but about how a system perceives the world, makes decisions, and interacts with real environments. In the future, I hope to continue my studies in a stronger research setting and build deeper capability in AI and embodied intelligence.

身份 Status 大二学生 Sophomore

学校 University 中山大学 Sun Yat-sen University

英文名 English Name TheBrave

目标 Goal 继续深造 Further Study

方向 Orientation 研究导向成长 Research-oriented Growth

Personal Statement

我希望把本科阶段视作能力成形的关键窗口: 一边夯实基础课程与工程实践，一边通过论文复现、数学建模和项目探索逐步建立自己的研究直觉。相比“尽快做出很多成果”，我更在意能否持续进入问题、理解问题，并在真实约束下把事情向前推进。

I see my undergraduate years as a key stage for shaping my abilities: strengthening fundamentals and engineering practice while building research intuition through paper reproduction, mathematical modeling, and project exploration. Rather than chasing fast results, I care more about entering a problem, understanding it well, and pushing it forward under real constraints.

Research Interests

我正在持续靠近的方向

Directions I am steadily moving toward.

人工智能基础

AI Foundations

系统打牢机器学习、深度学习、数据结构、算法与数学基础，为后续科研训练和复杂问题建模做准备。

Building a strong base in machine learning, deep learning, algorithms, and mathematics for future research training and complex problem modeling.

具身智能

Embodied Intelligence

关注智能体如何在真实或模拟环境中完成感知、规划、决策与执行的闭环，以及动作与环境反馈如何共同塑造智能。

Interested in how agents form a full loop of perception, planning, decision-making, and execution in real or simulated environments.

多模态与智能体

Multimodality and Agents

希望探索语言、视觉、动作之间的联动，以及更自主、更可落地的 Agent 系统如何被构建。

Exploring the link between language, vision, and action, and how more autonomous and practical agent systems can be built.

Skills

正在建立的能力地图

Capabilities I am building.

编程与工程

Programming and Engineering

计算机基础

Computer Science Fundamentals

AI 学习方向

AI Learning Directions

Preparation

我正在持续补强的学术准备

Academic preparation I am strengthening.

基础能力

Core Foundations

研究准备

Research Preparation

当前阶段我更重视基础打牢、复现能力和技术表达能力的同步积累，为后续进入更深入的科研训练做好准备。

At this stage, I care most about strengthening fundamentals, reproduction ability, and technical communication in parallel, so that I will be better prepared for deeper research training.

后续补充

Next Additions

后续这里可以继续补充课程成绩、更多项目、科研经历或阅读过的重要论文，逐步把网站沉淀成完整学术主页。

This section can later be expanded with coursework records, more projects, research experience, or important papers I have studied.

Reading

正在形成的论文阅读脉络

Papers shaping my current interests.

Foundation Model

Attention Is All You Need

这篇论文帮助我更系统地理解了 Transformer 架构的核心思想，也让我开始意识到表示学习、注意力机制与大模型范式之间的联系。

This paper helped me build a clearer understanding of the Transformer architecture and the connection between representation learning, attention mechanisms, and modern foundation models.

Embodied Intelligence

OmniDexGrasp: Generalizable Dexterous Grasping via Foundation Model and Force Feedback

这篇论文把我对具身智能的兴趣从“概念层面”推进到了“工程与复现层面”。它不仅对应我的论文复现实践，也让我开始更具体地思考感知、控制与泛化能力之间的关系。

This paper pushed my interest in embodied intelligence from a conceptual level to an engineering and reproduction level. It also made me think more concretely about perception, control, and generalization in real systems.

Next Step

后续阅读方向

Next Reading Directions

Projects

从复现、排障到真正理解问题

From reproduction and debugging to real understanding.

Embodied AI / Reproduction

OmniDexGrasp 论文复现

OmniDexGrasp Paper Reproduction

论文复现 Paper Reproduction 环境配置 Environment Setup 问题排障 Debugging

这是我围绕具身智能与抓取任务做过的一次论文复现实践。我完成了环境配置，但在实际运行阶段遇到了难以推进的问题，最终借助 Codex 持续排查与修复，才把复现流程真正往前推进。

This was a paper reproduction practice around embodied intelligence and grasping tasks. I completed the environment setup, but ran into stubborn runtime issues and pushed the workflow forward through repeated debugging with Codex.

这次经历让我更直观地认识到，研究复现不仅是“跑代码”，更是理解依赖关系、实验条件、工程细节与论文设定之间差异的过程。

This experience made me realize that research reproduction is not just about running code, but about understanding dependencies, experimental conditions, engineering details, and the gap between paper settings and real execution.

查看原始仓库View Original Repository

Problem Solving / Practice

环境配置与技术排障

Environment Setup and Technical Troubleshooting

依赖管理 Dependency Management 运行链路 Execution Pipeline 报错定位 Error Localization

在复现过程中，我花了很多精力处理环境依赖、运行链路和报错定位问题。这种经历虽然不显眼，但非常锻炼工程耐心、问题拆解能力与对复杂系统的理解。

During the reproduction process, I spent substantial time resolving dependency issues, tracing execution flow, and locating runtime errors. Though less visible than model results, this work trained my patience and systems thinking.

Future / Research Track

下一步研究积累

Next Research-Oriented Steps

Embodied AI Agent Planning Simulation

接下来我希望继续补充机器人仿真、Agent 任务规划、感知决策联动等项目，把当前的兴趣逐步沉淀成更完整的研究主线。

Going forward, I hope to add work in robot simulation, agent planning, and perception-decision coupling, turning current interests into a clearer long-term research track.

Awards & Experience

用结果证明学习能力

Learning ability shown through results.

全国大学生数学建模竞赛

National Undergraduate Mathematical Contest in Modeling

广东赛区省级一等奖

Provincial First Prize, Guangdong Division

获得全国大学生数学建模竞赛广东赛区省级一等奖。这段经历锻炼了我在有限时间内完成问题抽象、模型建立、方案验证和团队协作表达的能力。

I received the Provincial First Prize in the Guangdong Division. This experience strengthened my ability to abstract problems, build models, verify solutions, and collaborate under time pressure.

我看重的能力

Capabilities I Value

无论是数学建模还是论文复现，我都在尝试积累同一种能力：面对陌生而复杂的问题，保持耐心，拆解路径，持续推进，直到找到可行解。

Whether in mathematical modeling or paper reproduction, I am training the same ability: staying patient with unfamiliar and complex problems, decomposing them, and steadily moving toward workable solutions.

当前阶段关键词

Current Keywords

Journey

一条正在展开的成长路径

An academic path still unfolding.

现在

Now

夯实计算机基础课程，建立对 AI 与具身智能方向的系统认知。

Strengthening core computer science coursework and forming a systematic understanding of AI and embodied intelligence.

接下来

积累项目、参与科研训练、完善技术作品与公开表达能力，并为进一步深造持续准备。

Building more projects, joining research training, and improving technical communication while preparing for further study.

未来

Future

希望在更高水平的研究环境中继续深造，在 AI 与具身智能领域参与更接近真实智能系统的问题研究。

Hoping to continue in a stronger research environment and work on questions in AI and embodied intelligence that are closer to real intelligent systems.

Contact

欢迎交流、合作与连接

Open to connection, conversation, and collaboration.

如果你想和我交流 AI、具身智能、课程学习、项目实践或未来研究方向，欢迎联系我。

If you would like to discuss AI, embodied intelligence, coursework, project practice, or future research directions, feel free to reach out.

kangk6@mail2.sysu.edu.cn GitHub / kangk6

友情链接 Friends

hentilix.pages.dev

这个版本已经替换成了你的真实邮箱和 GitHub，可以继续补充更多项目与成果。

This version already uses your real email and GitHub link, and can be expanded with more projects and achievements over time.