Skip to content

如何做学术

Resources

蒋炎岩：读博那些事

阅读：开眼界，看看其他领域的人再用什么方法解决什么问题

  1. Book: "Parallel and Distributed Computer: Numerical Algorithms"
  2. Magazine: "Communications of the ACM (CACM)"

课程

  1. 课程清单
    1. CS 自学指南
  2. 软件分析
    1. 南大软件分析
    2. THU: 软件分析与验证
  3. 操作系统
    1. MIT 6.S081
    2. NJU OS: Operating System Design and Implementation
    3. 15-712 Advanced and Distributed Operating Systems, Spring 2012
  4. 计算机网络
    1. CS 144: Computer Network
  5. 体系结构
    1. CMU CS15213: CSAPP
  6. 分布式
    1. MIT 6.824: Distributed System
    2. DDIA: 设计数据密集型应用
  7. 数据库
    1. CMU 15-445: Database Systems
  8. 深度学习
    1. 动手学深度学习
    2. 吴恩达：深度学习课程
    3. 李沐：深度学习经典论文精读
    4. 花书
  9. 机器学习
    1. 吴恩达：机器学习课程
    2. 机器学习编译（陈天奇）
  10. 编译原理
    1. Stanford CS143: Compilers
  11. 其他
    1. The Missing Semester of Your CS Education

如何做科研

Here are some tips which will help you to succeed in your career:

• Be a scientist not a data collector. Do data analysis.

• Write an email to your advisor/boss/manager who can get your point in five seconds.

• Report data/results/analysis with your insights. Tell them good or bad. If it is bad, tell them your plan. Do not just tell them I got a number without any evaluation. You will be a scientist not a data collector. For example, you get the conductivity of LSGM, you should tell me: our LSGM has higher/lower/comparable conductivity compared with the reported data. Your advisor/boss/manager does not care about the number.

• Your advisor/boss/manager should be the GPS and you should be the driver. They are not the drivers and you should not be the passenger.

• After performing one experiment/job/activity, you should tell your advisor/manager/boss “I am going to …… ” never “what” or “why”

• Be creative! Work independently! I’m going to inspire you to know how to design experiments not help you design experiments.

It is always hard to survive in the long path towards a good career. You might feel uncomfortable, but you will finally realize above tips are beneficial.