Friends who have read "Water Margin" know that one hundred and eight generals in Liangshan joined the water village for various reasons, and embarked on the road of robbing the rich, helping the poor, and doing justice for heaven. After I graduated from Nanjing University in 2006

Friends who have watched " Water Margin " know that one hundred and eight generals in Liangshan joined the water village for various reasons, and embarked on the road of robbing the rich, helping the poor, and doing justice for heaven.

After I graduated from Nanjing University in 2006, I heard and witnessed hundreds of PhD students. The reasons why they chose to study for PhDs can be said to be more complicated than that of Liangshan heroes.

Take me for example. I have not been interested in programming since I was a child. When I graduated from high school, I wanted to major in mathematics or physics. However, my father, a mathematics professor, forced me to major in computer science. He had many reasons. , I don’t want my son to follow his old path. Later I heard that there is actually a profession for computer majors that does not require programming - being a professor! From then on, I had an endless desire to be a professor, so I persisted in completing my Ph.D.

My master's tutor Professor Ni Mingxuan (currently Vice President of the University of Macau) once had a senior student. When Professor Ni asked him why he studied for a doctorate, he replied: "My mother asked me to get a doctorate. After graduating with a Ph.D., I went to make money.” This legendary senior brother is now a senior leader of a well-known domestic IT company.

Corresponding to the various reasons for studying for a Ph.D., these Ph.D. (students) also have different life paths. Some have become university professors at home and abroad, some have worked as technical backbones in companies, some have been ordinary engineers, some have dropped out of school in time to work when they encountered difficulties during their Ph.D. studies, and some have worked hard until graduation and then went to work in . Of course, they all ended much better than the heroes of Liangshan.

I often have friends or fellow students ask me: "Do you think it is good to study for a Ph.D.?" "Do you think I should study for a Ph.D.?" These are quite difficult questions to answer.

When we choose whether to study for a Ph.D., we often do not have enough understanding of this major, such as in our junior or senior years. There is a typical "herding effect" among Chinese students: if the students around you (especially the academic models) are preparing to go abroad or study for a Ph.D. in China, then everyone will follow suit and think that studying for a Ph.D. is very honorable. If everyone says that studying for a Ph.D. is meaningless, even students who are interested in scientific research may not study for a Ph.D. because of face issues.

10 years ago, I went to interview with my classmates for the engineer position of Microsoft , and even the sales manager position of Procter & Gamble . Thinking about it now, those positions were neither my strengths nor my interests, but how could I judge them back then?

I would like to give some suggestions to students through a retrospective analysis of the life paths of some PhDs to help them choose whether to study for a PhD. encourages them to choose their own path through independent thinking.

Some misunderstandings about doctoral studies

Young students have some misunderstandings about doctoral studies due to poor information channels or lack of correct guidance.

1. "Ph.D. studies and scientific research are only suitable for people who are extremely smart or have very good grades."

This view is completely wrong. It is as false a proposition as "Only those with great acting talent can be actors." If you think that the academic giants in everyone's mind are all extremely smart people, it would be like saying that the stars who often appear on the homepage of , Sina or Weibo's hot searches all have very good acting skills.

Among the PhDs I know, those who do top-notch scientific research are not necessarily the super smart people among their peers, nor are they necessarily the ones with the highest Grade Point Average (GPA), nor are they necessarily those who participated in ACM programming The masters of the competition, and are just those who have the patience and perseverance to study .

Many doctoral students can produce top-notch research as long as they choose a mentor, choose a research direction, and are willing to spend time. Many scholars who have made great achievements in overseas academic circles told me in private conversations that their academic performance was not good.

2. “Three generations will be poor in scientific research, but a PhD will ruin their lives."

This sentence may be true in other fields, but it is wrong in the computer field. In the computer field, there are many people who make money by studying for Ph.D.! The characteristic of computer science is that scientific research and industry are very closely integrated. In particular, the most cutting-edge scientific research has a great promotion effect on the industry and generates economic benefits. A typical example is that Page and Brin, the founders of Google, are both doctoral students and come from doctoral families. Although they started a business before finishing their Ph.D. (Brin claimed that he is now working as a Ph.D.), if they had not studied for a Ph.D., it would be difficult to create Google's core technology

The current darling of the industry's big data system is Spark. Developed by professors and Ph.D.s at the University of California, Berkeley. Lead developer Matei Zaharia did not completely give up his teaching position at

in the field of network technology even with the opportunity to make big money. , the core " software-defined network " that the industry is currently paying attention to was also developed by professors and PhDs in universities. Google, VMware (VMware), Databricks (Spark), Nicira (OpenFlow) and countless top technology companies have hired many. Thousands of people graduate with Ph.D.s in computer science, because Ph.D.s are usually closer to core technologies than other employees.

If you want to realize your technical dreams and make a lot of money, studying for a Ph.D. is a good choice. "The path of life" does not make any sense.

3. "Most of the things produced in scientific research are waste paper and have no practical help at all. "

This view is also incorrect in the computer field. Computer science is not the skill to slay the dragon. The prototype of the computer network that billions of people are inseparable from today was developed from the University of California, Los Angeles. My senior (my mentor) Professor Leonard Kleinrock (mentor) developed the prototype of the Internet in the ARPANET (ARPANET) project and sent the Internet's first data packet in 1969. In the 1970s, I Professor Simon Lam, his mentor and an academician of the American Academy of Engineering, analyzed the method of resolving link layer conflicts in his doctoral thesis, which was finally adopted by Ethernet and integrated into our lives

Professor Lam in the 1990s. The first implementation system of Secure Sockets Layer (SSL) was developed. Today, Secure Sockets Layer is used in the browser of every computer and mobile phone. Not only computer networks, but also technology in almost every computer field is led by scientific research.

According to an article published by Dr. Qian Xuehai, a postdoctoral researcher at the University of California, Berkeley, in the "Communications of the Chinese Computer Society" (CCCF)1, the development of computer architecture in recent years is closely related to the papers at the four major academic conferences in this field. Inseparable.

On the other hand, although many papers may not necessarily reflect direct value in the real world, you can learn a lot in the process of writing articles, such as the ability to ask questions, solve problems, and express language, which all make People benefit a lot.

4. “It is not as easy to find a job as an undergraduate or a master’s degree graduate. "

This may happen in other majors, because many things do not require people with a Ph.D. degree to complete. A bachelor's degree is more than enough, but this view is wrong in the computer field.

Advanced technology companies in the United States often have a strong focus on computer majors Ph.D.s are eager for talent, because Ph.D. masters advanced technology and can bring economic benefits to the company.

When I recently visited Intel headquarters, Intel researchers mentioned that 20 years ago, Intel was recruiting masters. Research, however, recently Intel basically only recruits PhDs. In the United States, if a PhD's scientific research field happens to be what the company needs, many companies will offer an annual salary of US$150,000 and company shares to recruit talents.

This situation is not common in China at present. The reason may be that some domestic companies are still in the imitation stage rather than the innovation stage. As domestic technology levels continue to improve, more and more companies will be hungry for computer science PhDs who have mastered scientific research skills and core technologies. It is estimated that the situation will improve within five years.

What can you learn from a PhD?

Studying for a doctorate is not only about learning knowledge and technology, but also cultivating other professional skills. What you gain from reading the blog is not just the few pages of the paper, but also what you learned in the process of writing the paper. For Chinese students, does not lack the ability to solve problems and program at work, but lacks the ability to express and communicate.

For example, if you have an idea that can double the performance of a certain software in the company, then you need to convince the company's senior leaders that your idea is meaningful and worthy of investing resources in development. The abilities required to achieve this goal have always been lacking among Chinese students, but they can be trained through doctoral studies.

1. The ability to grasp the needs of the company's senior leaders and users

If the company's senior leaders are not interested in the performance of a certain software, of course you cannot propose ideas to improve the performance of the software. You need to know what technologies and products the company needs most currently. This ability is not something you are born with, but is something you develop over time by pondering and exercising.

In the process of studying for a Ph.D., you must know the topics that your tutor likes and is good at, and understand the presentation methods and results that the reviewers like. This is an excellent training opportunity for you, because when you are studying for a Ph.D., you are allowed to fail; but in the workplace, the price of failure is very high.

2. Language expression ability

During my study in Hong Kong, Professor Liu Yunhao (currently Dean of the School of Software, Tsinghua University) repeatedly emphasized the importance of expression ability. As a scientific researcher, you need to write papers that can convince others; as a software engineer, you need to write technical documents that reflect your contribution; as a project leader, you need to write good project applications and summaries Report and let the company know your value.

Many Chinese students have very weak language expression ability. This is not closely related to their foreign language proficiency, but due to the lack of accuracy and logic training in writing. The process of reading and writing papers in can cultivate logical expression skills.

3. The ability to overcome difficulties

Being rejected by others can hurt your self-esteem. As a proud college student, apart from pursuing the opposite sex, I am afraid that I rarely encounter the experience of being rejected. And the cruel workplace is full of rejection. How to face these things with a correct mentality can also be trained by studying for a Ph.D.

Many students are very angry and ashamed when their first submitted paper is rejected. There are even many smart students who gave up on academic pursuits in anger after their papers were rejected several times.

In fact, the reason why the paper is rejected is not necessarily because you did a bad job, it may just be a little bit of luck and the integrity of the work. In the workplace, you may also be rejected for job applications and promotions, and the reason is not necessarily due to your own shortcomings. Reading a PhD in can help you recognize and understand these things, and face setbacks with a normal mind.

Who is suitable for a PhD?

I think that studying for a PhD does not require innate special abilities or acquired conditions. The abilities we usually talk about, such as programming, mathematics and English, can only be called catalysts for PhD study, rather than "reactants" for chemical reactions. Therefore, the question of is not "who is suitable for a PhD", but "who is not suitable for a PhD".

1. People who are impatient are not suitable for PhD study

A PhD in computer science needs to become a world-class expert in his own research direction.

No one can become an expert in a short time. Even if we see some doctoral students publishing top papers in the first or second year, they actually do not fully understand the direction as a whole. The topic selection may be Rely on the help of the tutor, or even be designated by the tutor.

Most important work, especially in areas involving systems (such as operating systems, networks, distributed systems, architectures, programming languages), cannot be achieved in a short time. An important system paper usually takes nearly a year to experiment and write, and the accumulation of previous knowledge takes even longer.

The period before the article is published is the most difficult. Most failed doctoral students fall at this stage. Most of those who can survive this stage with can continue to publish more papers regardless of their previous background. "Patience" is the most important condition for studying for a Ph.D.

2. People who do not appreciate science and technology are not suitable for PhD study

A successful computer PhD must know how to appreciate (or at least be willing to appreciate) the excellent results published by others, appreciate the classic designs in textbooks, and appreciate the important questions and answers raised by oneself or the instructor. Solution. If you don't appreciate technology at all, it's impossible to produce good scientific research results.

What should you pay attention to when studying for a PhD?

1. Choosing the right supervisor

The supervisor is closely related to the fate of a doctoral student. Two doctoral students with the same conditions may have completely different fates if they meet different supervisors.

Each tutor has a different background, has different resources at hand, and the number of resources he is willing to give to his students is also different.

Suppose a senior professor A’s resources are 10 and a young professor B’s resources are 5. But Professor A has a huge laboratory with 10 students, and can only give you 1 resource; while Professor B only has you as a student, and can give you 5 resources. Obviously Professor B is the more ideal choice. Of course the actual situation is not that simple.

For example, in the United States, there is usually not a big gap in resources between senior and young professors. In China, resource differences can be huge—although the gap is now narrowing. It is also possible that you are outstanding among your classmates, and the senior professor is particularly fond of you and can give you 8 resources.

It is also possible that a young professor is good at teaching and has very close relationships with many colleagues and the industry. At this time, his resources exceed those of senior professors. Therefore, how to choose a mentor is definitely not as simple as whether they are experienced or have a good temper.

2. Choosing the right topic

The topics that can be chosen vary widely in the same field. Some topics have no research value. Some topics are difficult, and even if you do them carefully, you may not be able to achieve good results.

My suggestion is that PhD students should always read papers from top conferences. Even if the research group has never published a paper at a top conference in history, even if the instructor does not require it, and even if you can graduate without publishing a paper at a top conference, you still have to understand it. Otherwise, the research will easily be out of touch with reality, and naturally it will not be recognized.

Career choices after the PhD

I think that studying for a PhD in computer science is a choice that can be advanced or retreated. Because in the computer field, academia and industry are closely connected, and no field is inseparable from computing, the career options for PhD students are quite diverse.

Some of my PhD brothers and sisters at the University of Texas have become tenured professors at Ivy League universities, some have become directors of Bell Labs , some have become project leaders of Google's , and some have become founders of start-up companies. .

Some of my PhD brothers and sisters at Hong Kong University of Science and Technology have become vice presidents of JPMorgan Chase (JPMorgan Chase), some have become upstarts in prestigious domestic universities, and some have held important positions in the research institutes of the Ministry of Public Security.No matter how unsatisfactory their career choice is, PhD students in computer science rarely worry about their livelihood. In the eyes of people in other majors, these people are elites in all walks of life.

If you don’t waste your time, the knowledge you learn during your PhD will be useful in your future career.

If your advantage is writing papers, you can choose to become a university professor; if your advantage is solving technical problems, you can choose to become a researcher in the company; if your advantage is programming, you can become a software engineer; if your advantage is programming, you can become a software engineer; It's modeling, you can join financial companies, etc.

The knowledge of computer majors is all practical, which is a unique advantage of computer majors. Even if the process of studying for a Ph.D. is very unsatisfactory, you can still jump out of the "stop loss" in time. The society will not regard you as a loser. The company still welcomes Ph.D. dropouts - because they have also learned technology.

If one hundred and eight heroes were doomed to have a rough life the moment they set foot on Liangshan, then on the contrary, getting a doctorate in certain majors can guarantee your future life, and on this basis you will have Diversity of options to use to your advantage.

This sentence may be true in other fields, but it is wrong in the computer field. In the computer field, there are many people who make money by studying for Ph.D.! The characteristic of computer science is that scientific research and industry are very closely integrated. In particular, the most cutting-edge scientific research has a great promotion effect on the industry and generates economic benefits. A typical example is that Page and Brin, the founders of Google, are both doctoral students and come from doctoral families. Although they started a business before finishing their Ph.D. (Brin claimed that he is now working as a Ph.D.), if they had not studied for a Ph.D., it would be difficult to create Google's core technology

The current darling of the industry's big data system is Spark. Developed by professors and Ph.D.s at the University of California, Berkeley. Lead developer Matei Zaharia did not completely give up his teaching position at

in the field of network technology even with the opportunity to make big money. , the core " software-defined network " that the industry is currently paying attention to was also developed by professors and PhDs in universities. Google, VMware (VMware), Databricks (Spark), Nicira (OpenFlow) and countless top technology companies have hired many. Thousands of people graduate with Ph.D.s in computer science, because Ph.D.s are usually closer to core technologies than other employees.

If you want to realize your technical dreams and make a lot of money, studying for a Ph.D. is a good choice. "The path of life" does not make any sense.

3. "Most of the things produced in scientific research are waste paper and have no practical help at all. "

This view is also incorrect in the computer field. Computer science is not the skill to slay the dragon. The prototype of the computer network that billions of people are inseparable from today was developed from the University of California, Los Angeles. My senior (my mentor) Professor Leonard Kleinrock (mentor) developed the prototype of the Internet in the ARPANET (ARPANET) project and sent the Internet's first data packet in 1969. In the 1970s, I Professor Simon Lam, his mentor and an academician of the American Academy of Engineering, analyzed the method of resolving link layer conflicts in his doctoral thesis, which was finally adopted by Ethernet and integrated into our lives

Professor Lam in the 1990s. The first implementation system of Secure Sockets Layer (SSL) was developed. Today, Secure Sockets Layer is used in the browser of every computer and mobile phone. Not only computer networks, but also technology in almost every computer field is led by scientific research.

According to an article published by Dr. Qian Xuehai, a postdoctoral researcher at the University of California, Berkeley, in the "Communications of the Chinese Computer Society" (CCCF)1, the development of computer architecture in recent years is closely related to the papers at the four major academic conferences in this field. Inseparable.

On the other hand, although many papers may not necessarily reflect direct value in the real world, you can learn a lot in the process of writing articles, such as the ability to ask questions, solve problems, and express language, which all make People benefit a lot.

4. “It is not as easy to find a job as an undergraduate or a master’s degree graduate. "

This may happen in other majors, because many things do not require people with a Ph.D. degree to complete. A bachelor's degree is more than enough, but this view is wrong in the computer field.

Advanced technology companies in the United States often have a strong focus on computer majors Ph.D.s are eager for talent, because Ph.D. masters advanced technology and can bring economic benefits to the company.

When I recently visited Intel headquarters, Intel researchers mentioned that 20 years ago, Intel was recruiting masters. Research, however, recently Intel basically only recruits PhDs. In the United States, if a PhD's scientific research field happens to be what the company needs, many companies will offer an annual salary of US$150,000 and company shares to recruit talents.

This situation is not common in China at present. The reason may be that some domestic companies are still in the imitation stage rather than the innovation stage. As domestic technology levels continue to improve, more and more companies will be hungry for computer science PhDs who have mastered scientific research skills and core technologies. It is estimated that the situation will improve within five years.

What can you learn from a PhD?

Studying for a doctorate is not only about learning knowledge and technology, but also cultivating other professional skills. What you gain from reading the blog is not just the few pages of the paper, but also what you learned in the process of writing the paper. For Chinese students, does not lack the ability to solve problems and program at work, but lacks the ability to express and communicate.

For example, if you have an idea that can double the performance of a certain software in the company, then you need to convince the company's senior leaders that your idea is meaningful and worthy of investing resources in development. The abilities required to achieve this goal have always been lacking among Chinese students, but they can be trained through doctoral studies.

1. The ability to grasp the needs of the company's senior leaders and users

If the company's senior leaders are not interested in the performance of a certain software, of course you cannot propose ideas to improve the performance of the software. You need to know what technologies and products the company needs most currently. This ability is not something you are born with, but is something you develop over time by pondering and exercising.

In the process of studying for a Ph.D., you must know the topics that your tutor likes and is good at, and understand the presentation methods and results that the reviewers like. This is an excellent training opportunity for you, because when you are studying for a Ph.D., you are allowed to fail; but in the workplace, the price of failure is very high.

2. Language expression ability

During my study in Hong Kong, Professor Liu Yunhao (currently Dean of the School of Software, Tsinghua University) repeatedly emphasized the importance of expression ability. As a scientific researcher, you need to write papers that can convince others; as a software engineer, you need to write technical documents that reflect your contribution; as a project leader, you need to write good project applications and summaries Report and let the company know your value.

Many Chinese students have very weak language expression ability. This is not closely related to their foreign language proficiency, but due to the lack of accuracy and logic training in writing. The process of reading and writing papers in can cultivate logical expression skills.

3. The ability to overcome difficulties

Being rejected by others can hurt your self-esteem. As a proud college student, apart from pursuing the opposite sex, I am afraid that I rarely encounter the experience of being rejected. And the cruel workplace is full of rejection. How to face these things with a correct mentality can also be trained by studying for a Ph.D.

Many students are very angry and ashamed when their first submitted paper is rejected. There are even many smart students who gave up on academic pursuits in anger after their papers were rejected several times.

In fact, the reason why the paper is rejected is not necessarily because you did a bad job, it may just be a little bit of luck and the integrity of the work. In the workplace, you may also be rejected for job applications and promotions, and the reason is not necessarily due to your own shortcomings. Reading a PhD in can help you recognize and understand these things, and face setbacks with a normal mind.

Who is suitable for a PhD?

I think that studying for a PhD does not require innate special abilities or acquired conditions. The abilities we usually talk about, such as programming, mathematics and English, can only be called catalysts for PhD study, rather than "reactants" for chemical reactions. Therefore, the question of is not "who is suitable for a PhD", but "who is not suitable for a PhD".

1. People who are impatient are not suitable for PhD study

A PhD in computer science needs to become a world-class expert in his own research direction.

No one can become an expert in a short time. Even if we see some doctoral students publishing top papers in the first or second year, they actually do not fully understand the direction as a whole. The topic selection may be Rely on the help of the tutor, or even be designated by the tutor.

Most important work, especially in areas involving systems (such as operating systems, networks, distributed systems, architectures, programming languages), cannot be achieved in a short time. An important system paper usually takes nearly a year to experiment and write, and the accumulation of previous knowledge takes even longer.

The period before the article is published is the most difficult. Most failed doctoral students fall at this stage. Most of those who can survive this stage with can continue to publish more papers regardless of their previous background. "Patience" is the most important condition for studying for a Ph.D.

2. People who do not appreciate science and technology are not suitable for PhD study

A successful computer PhD must know how to appreciate (or at least be willing to appreciate) the excellent results published by others, appreciate the classic designs in textbooks, and appreciate the important questions and answers raised by oneself or the instructor. Solution. If you don't appreciate technology at all, it's impossible to produce good scientific research results.

What should you pay attention to when studying for a PhD?

1. Choosing the right supervisor

The supervisor is closely related to the fate of a doctoral student. Two doctoral students with the same conditions may have completely different fates if they meet different supervisors.

Each tutor has a different background, has different resources at hand, and the number of resources he is willing to give to his students is also different.

Suppose a senior professor A’s resources are 10 and a young professor B’s resources are 5. But Professor A has a huge laboratory with 10 students, and can only give you 1 resource; while Professor B only has you as a student, and can give you 5 resources. Obviously Professor B is the more ideal choice. Of course the actual situation is not that simple.

For example, in the United States, there is usually not a big gap in resources between senior and young professors. In China, resource differences can be huge—although the gap is now narrowing. It is also possible that you are outstanding among your classmates, and the senior professor is particularly fond of you and can give you 8 resources.

It is also possible that a young professor is good at teaching and has very close relationships with many colleagues and the industry. At this time, his resources exceed those of senior professors. Therefore, how to choose a mentor is definitely not as simple as whether they are experienced or have a good temper.

2. Choosing the right topic

The topics that can be chosen vary widely in the same field. Some topics have no research value. Some topics are difficult, and even if you do them carefully, you may not be able to achieve good results.

My suggestion is that PhD students should always read papers from top conferences. Even if the research group has never published a paper at a top conference in history, even if the instructor does not require it, and even if you can graduate without publishing a paper at a top conference, you still have to understand it. Otherwise, the research will easily be out of touch with reality, and naturally it will not be recognized.

Career choices after the PhD

I think that studying for a PhD in computer science is a choice that can be advanced or retreated. Because in the computer field, academia and industry are closely connected, and no field is inseparable from computing, the career options for PhD students are quite diverse.

Some of my PhD brothers and sisters at the University of Texas have become tenured professors at Ivy League universities, some have become directors of Bell Labs , some have become project leaders of Google's , and some have become founders of start-up companies. .

Some of my PhD brothers and sisters at Hong Kong University of Science and Technology have become vice presidents of JPMorgan Chase (JPMorgan Chase), some have become upstarts in prestigious domestic universities, and some have held important positions in the research institutes of the Ministry of Public Security.No matter how unsatisfactory their career choice is, PhD students in computer science rarely worry about their livelihood. In the eyes of people in other majors, these people are elites in all walks of life.

If you don’t waste your time, the knowledge you learn during your PhD will be useful in your future career.

If your advantage is writing papers, you can choose to become a university professor; if your advantage is solving technical problems, you can choose to become a researcher in the company; if your advantage is programming, you can become a software engineer; if your advantage is programming, you can become a software engineer; It's modeling, you can join financial companies, etc.

The knowledge of computer majors is all practical, which is a unique advantage of computer majors. Even if the process of studying for a Ph.D. is very unsatisfactory, you can still jump out of the "stop loss" in time. The society will not regard you as a loser. The company still welcomes Ph.D. dropouts - because they have also learned technology.

If one hundred and eight heroes were doomed to have a rough life the moment they set foot on Liangshan, then on the contrary, getting a doctorate in certain majors can guarantee your future life, and on this basis you will have Diversity of options to use to your advantage.