Eric Xing’s advice to Machine Learning students

这是刑波老师去年在上海龙星计划机器学习课程之后给学员们的一封信,今天偶然间又翻出来读,仍旧觉得受益匪浅,转到此处,希望对看到的同学能有帮助。

Dear Students,

It has been a great experience to work with you during the 5-day Dragon Star lectures. I truly appreciate that you have attended all the lectures with patience and enthusiasm, and I am glad most you have enjoyed and liked the lectures. As a teacher, your approval and support are the highest honor I can enjoyed in my career.  I got many questions from you on how to be a good researcher, bellow are some words I wrote to a student i don’t personally know who asked me the same question through email a few years ago. Hopefully this complement Fei-Fei’s excellent essay on how to do good research.

…. as general advices on students interested in pursue a serious science career, I think the following are important:

1) know exactly what your goal is, what is your strength and your weakness, both technical and mental, as they are equally important for success.
Set you goal and plan your moves accordingly.

2) don’t wait if you really believe a change is needed. It is never too late to change career course if the change is serious and a full dedication for the new course is committed. The amount of time wasted in waiting is often longer than the time needed to clinch the new life. For me, I was once at the bottom of my new career because I nearly knew nothing of machine learning when I changed major, but I knew I would love it because of its beauty and power. It took me 5 years starting from that point to
become a professor of machine learning at CMU.

3) you need to work extremely hard. In my graduate and professional career, I worked on average 12 hours per day and 7 days per week, with high concentration and efficiency. (Of course, as I mentioned in our evening chat, I personally do not actually view such activities as “work”, they are fun and exciting, as much as, or better than other entertainments.)

4) you work hard not because you are pushed by your boss, but because you are inspired and you love what you do. Indeed, I often told my students and friends that I “LIVE” my research rather than doing my research. I enjoy doing it more than watching TV and playing card, etc.

5) you can still live a colorful life and be energetic. I am big sports and music fun, and love to do many other things, and I enjoy being with my friends and family.  So I waist no time idling. Whenever I am not working, I play or exercise hard, and be with my family and friends.

6) you need to train yourself to be very creative, and very independent. Honestly almost all the research ideas in my graduate study were not from my advisors, but from myself; what I got from my advisors at Berkeley were moral support, inspiration on research style and taste, sense of honest and pride, and their friendship.

7) being informed of the latest progress in your research area. It is not unusual to read 1 thousand papers per year. Basically, to be on top of the field you need to know EVERYTHING of your field, and in many cases many related fields.

8) when breaking new grounds, on the one hand, you need to respect the intelligence of the authorities in the fields, don’t assume they are stupid and you are smarter, and always ask why they do not do the same as what you want to do years ago. On the other hand, when you can convince yourself that you indeed gain an insight that those authorities had not, you should have the courage to surpass them.

9) Being versatile and flexible. Many problems and techniques are related, don’t hand yourself on a single tree.

10) It is never a bad thing to be a perfectionist and idealist in research. It is absolutely important to have an optimistic character, and to be physically fit and strong. It is very stupid to assume that one can work less because others are less smart. The truth is that, most of the top researchers I know of are not only extremely smart, but also work long hours and are efficient. So even physically you need to have means to compete. I myself used to be a semi-professional athlete when I was in college, and
even now I have no problem consistently working/playing harder and longer than my students 10+ years younger than my age.

11) learn the art of communication. Be ready to discuss and share your ideas with your colleagues and competitors. Learn from not only your colleagues, but also from your competitors.

12) Most importantly, be honest, open, patient, happy, and far-sighted, rather than being sneaky, isolated, anxious, bitter, and short-sighted.

Finally, I wish all of you happy and prosperous in the years to come.

Best regards,
Eric

Advertisements

一个有关“Eric Xing’s advice to Machine Learning students”的想法

发表评论

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / 更改 )

Twitter picture

You are commenting using your Twitter account. Log Out / 更改 )

Facebook photo

You are commenting using your Facebook account. Log Out / 更改 )

Google+ photo

You are commenting using your Google+ account. Log Out / 更改 )

Connecting to %s