Not everyone is suitable for a PhD, and there is no need for everyone to do a PhD. As long as people follow the path of success on their own. But now that you've decided to go for a PhD, you might want to have a clear understanding of what that means.
Quitting now, you don't have to wait until you are finally expelled, you can choose to leave on your own initiative. Why is it called permanent brain damage? Because after you finish the defense, you will find that your perspective on the problem will be different from that of many people, and even unique. In the midst of the usual noise, when everyone is giggling and discussing a problem, you will often find details that make a difference.
When you are discussing whether to choose A or B or C, if you are a Ph.D., you may say that the topic is wrong and logically problematic. That's what permanent brain damage can do for you. The damage cannot be repaired, and it is difficult to return to normal after the injury. You will find that except for you, everyone else is different from you and very different.
So what makes you different? What makes you make a difference? What does it take to become a qualified graduate? An excellent graduate? Or even if you don't graduate, you're still able to make important scientific achievements? Don't be afraid, in fact, whether you can graduate or not is not the primary consideration of doctoral students at Tsinghua University, my personal opinionThe most important thing for doctoral students at Tsinghua University is whether you can do what you should do, and the most important part of this is the cultivation of academic interests.
Three questions:Cultivation of academic interests
Many students don't know why they want to do a PhD. Maybe it's because everyone around you is doing a PhD, maybe it's because your parents want you to do a PhD. Have you ever wondered what you're going to be like in five years? This is an important question. What am I going to do if you were just living today? I have an exam tomorrow, and I have to submit a ** deadline the day after tomorrow, this is not right, this is a tactical level. Tactically you can win a small battle, but you can lose the entire war。You may write a lot, a lot, only to find out that you didn't learn much during your PhD, you may do a lot of work, you write a lot, and you end up feeling that doing a PhD here didn't achieve your desired goal. Why? The first problem comes because you didn't think it through in the first placeWhat you're going to be in five years' time。It is more appropriate, of course, to say that I want to graduate in five years, and if you don't want to graduate, this goal is not enough. I want to be an excellent graduate, that's a goal. In five years' time, I'm going to be able to do great things, and that's a good goal. Some classmates said that in five years, I want to be a rich person, well, this is also a goal, but you may choose the wrong direction, and you should not go to a doctoral program.
The next two questions you should ask yourself,:What do you want to be in 15 years? Are you working in these directions? These three questions were asked by my mentor when I was interviewed a long time ago. I was lucky enough to think about it at the time, so I was able to become a PhD student. I didn't think it was anything special at the time, I think it's a question that everyone thinks. It wasn't until I had the opportunity to talk about this lesson that I realized that no, many people don't think about these things. Sometimes you will find that you are so different from everyone else. It doesn't matter, each of you is different, but these three questions are worth thinking about.
Academic interest is reflected in whether you are really interested in studying a problem. This is also my personal opinion. Each teacher may have a different point of view. You don't go to graduate school to be rich, you don't do research to be famous, essentially you go to graduate school because you're curious, because you don't understand, and even because you're uncomfortable. Because you are in a lot of pain in your heart, you will be unable to sleep all day because of some problems, and you are very desperate to solve this problem. I saw that some people on the Internet would be surprised, saying that how well-known experts are, gray-haired, are actually very serious about listening to a certain report, and finally ask some questions from time to time, wow. Is there any mistake? This is what scholars themselves should do, so why shouldn't it? When you go to listen to a report, if your only purpose is to get people together, why should you listen? Just like if you come to listen to my lecture today, if your only purpose is for credit, for class attendance, I say seriously, no. You have to be honest with yourself, and you should ask yourself, are you interested in doing research?
Three conditions:Earn a Ph.D. degree
Conditions for obtaining a doctorate
You may or may not know by now that there are a few requirements required to get a PhD. Some students said that I want to publish enough articles, I want to have enough results, I want to write a good degree**, I want to review and get good opinions, and at the same time I need to find a good job, and finally I start to apply for graduation defense. Yes, both of them. None of that is quite right. These are natural processes, and these are not the main ones. I think you should ask yourself,First, are you able to teach the teacher something?
That's not what I thought, aren't students here to learn? Later, when I became a teacher, I felt that there was something wrong with this idea. It's good that the teacher contributes time, energy, and experience to the class. You graduated, and if you didn't teach your teacher anything, didn't that teacher pay for money, time, energy, and experience, and get nothing in the end? This is not right, very wrong. If you don't teach your teacher anything in the end, I don't think you're a qualified PhD student. You should teach your teacher at least one thing. After you were able to finish speaking, your teacher said yes, yes, I didn't think about this thing, thank you, I learned. This is your goal. Second, write an essay and let your peers say yes. You don't have to publish TAC, Automatica, PAMI, Proceedings of the IEEE, you don't have to write these "CNS" articles. The first article on the "Kalman Filter" was not posted on this either, but you should write a good article. You can post it in a very obscure journal or conference, but it must be a good article. And this good word should not be said by your teacher, not by your classmates. It should be a peer, from the bottom of my heart, and I sincerely admire and say yes. Third, do not disgrace the teacher。It's just that a teacher will go out in the future, and he won't feel disgraced because he says that this person is my student, but he should be very proud, this is my student!
None of this is what I said. When I was interviewed, the teacher said, I've finished asking you all the questions, what questions do you have? I said, let me ask a question, if I go to graduate school, how can I graduate? And then there are these three sentences. First, you should teach me something. Second, you should write an article and let our peers say yes. Third, do not disgrace the teacherPlease take a look, are there any quantitatives in these three? The first, one thing, 1. The second, an article, is also 1. The third article is more general, what does it mean to not disgrace the teacher? It's all a qualitative concept. A good criterion is not quantitative, it is qualitative. We are also in the process of translating to this aspect. In the past, the graduation requirements of our students were very clear and relatively high, and they were judged from a quantitative point of view. Now it is more and more qualitative, which really reflects the improvement of the level of our students at Tsinghua University, and we don't need those numbers to judge, not how many articles you have written, but to believe in peer review.
Three criteria:How to choose a topic
How to choose a topic After consideration, everyone participated in the interview, and finally successfully entered the school. One question that soon came was the topic selection. Assuming that you can all successfully pass the doctoral qualification examination, how to choose the topic? Give everyone some advice. My food is so bad that it's hard to eat, but that doesn't affect me from judging how well other people's food is cooked. So when you do research, you may not be able to produce particularly good results yourself, but you should have a good taste. This is the kind of taste that you need to train the most during your graduate studies. There is a joke that children never eat anything too salty, and when they grow up, they don't taste anything they eat. If your taste has deteriorated during your time in graduate school, and you eat too much salt, it may be difficult for you to become a particularly good researcher in the future. So when choosing a topic, you should ask yourself three questions first.
First, is this question of concern to engineers? It doesn't matter if you care about it or not. Does the engineer care? We are a graduate student in engineering, you are not science, you are not business, you are not art, you are engineering, so you must try to answer whether engineers care? Engineers may not care now, but they will care in the future, and that's fine. Someone goes to study the influence of articles published by MIT, and many articles do not have citations (citations) in the first five years of publication, and students will think that they are finished, and they will not have citations when they graduate, and they will not be rated excellent, right? These articles tend to turn out to be good results. There are also articles that have been published for ten years without citation, and these articles generally turn out to be particularly good results. There are also very few articles that have been sent out, and there has been no citation for 15 years, which is very interesting, and these articles are basically Nobel Prize-level. Some students have a high **citation, they should be happy, after all, there are still some people whose articles have not been cited for a lifetime, but don't be too happy. Otherwise, you were born in the right era, and your peers can know your work very well. Otherwise, it will only mean that what you are doing is a hot issue, but not a fundamental issue. Everyone knows that it is easy to get high when doing a hot topic, but as a basic question, others may not be able to understand for a long time why they should study this issue. On the other hand, a high citation is not necessarily a good thing, and it may also be cited as a negative example because his results are wrong. It is better to care about the topic selection than to care, whether the engineer will care.
Second, will I do this question? When choosing a topic, a doctoral student may say: "My supervisor chose a topic for me, but I don't know how to do it, and I haven't found anyone at home and abroad who has done it, and there is no literature or keywords, is it deliberately trying to embarrass me?" Can't get along with me. "Think about it, if you can do it, why would you study it? If you know how to do it, you can just write a report today, why spend four or five years? You are now in your 20s, and this is the most precious time in your life, so why do you want to do something else that you are good at? If you're doing something purely tedious but laborious, I suggest you don't do it. You may post a lot of articles in a short period of time, but in the long run, there is no harm in doing you a lot. You use the best time of your life to do a problem that you have already knew, and you have lost too many opportunities to do it.
Third, am I interested in this question? You have to ask yourself whether you are interested or not, and people must do the questions that you are interested in in order to truly develop your potential. So when I talk about it for the third time, I ask you to be honest with yourself. Some students may say that my topic was set by my tutor and I was not given the opportunity to choose the topic at all. I'm sorry, who said that you can only do one question alone, who said that a doctoral student should only do one question. I tell my students that a person should do at least three questions, one that you can graduate from, and one that is worthy of the tuition fees that people pay for you. There may also be a topic that you are looking for a job or that you are interested in for various reasons. You should work on the question that interests you, focus your most time on it, and of course it doesn't mean that you don't care about anything else.
Seven questions:How to do research
Teacher Jia Qingshan introduced seven questions
You are happier, we have fewer students. For example, if we can only recruit one Ph.D. student a year, and only one master's student can recruit every few years, you will occupy more resources for a teacher than others. So how do you do research? Do you want the teacher to assign me tasks? Teacher, you asked me to do three things, I can only do one, you tell me which one should I do? In fact, these problems should be solved by yourselves. It is this process that is what graduate students should receive. You have to figure out which problems are more fundamental and which ones are more superficial and can be done later. Even if you don't do research in the future, this ability will help you a lot.
Here are the seven strands, not created by me. It came from when I was a student, and there was a very famous chair professor in our center, named Mr. Lu Baosen, who proposed it.
First, what's the problem?What is this question that you are going to do? One sentence should be answered. Any question can be addressed to anyone in any short amount of time to the extent you want to go. If you don't do it, it's not because you're doing a lot of work, it's because you're not working hard enough. For example, I'm listening to a report and I haven't understood what he's saying for five minutes, not because I'm stupid, it's because the report is bad. Why? He didn't get the problem clear for five minutes, and he didn't respect the audience. If there is a report that I understand at the end, I will have great respect for this reporter, who belongs to a real bull, and it is not easy. The report is easy to understand, and I can explain its highlights in a very concise way, which is my standard now. Students, it is possible that I will appear on your defense committee in the future, and this is my standard. You don't have to worry about saying that I can finish my work in five minutes, does that mean I haven't done anything? No. I would be very happy. On the other hand, if you talk for 20 minutes, 30 minutes, and I don't understand what you're saying, I feel very bad. Ask yourself what the problem?One sentence should be clear.
Second, why is the problem important?Why is this important? This question is important because my teacher asked me to do it. This question is important, if he is not important, how can I write an article. This question is important, if he is not important, I spend my time in this life every day, how can I treat Jiangdong's father, so he is not important, I have to say that he is important. You're not honest enough. Let's be honest with you, whether you think this question is important or not. If you don't think it's important, don't do it, do it, what are you doing? Every day has the opportunity to change your life. Ma Yun said that he wanted to walk a thousand roads every night, but he still walked the old road when he got up in the morning. Why bother?
Third, why is the problem difficult?Is this question difficult, and why is this question difficult? If you haven't thought about it, think about it. If you've thought about it and found that it's not difficult, don't do it. You can't write an article, you can't get an article into a good journal, you can't get into a good journal, you go to a good journal, and no one will read it. is in a good journal, and others are laughed at when they read it. You should have to do an important question, a difficult question.
Fourth, what is the state of the art?What are others doing? People often have this question, how do you know if your literature research is sufficient? You stand here and close your eyes, and you should have an Earth in mind. On this earth, on every continent, the research group that is most relevant to the problem you are doing is in **. Each group can name the representative, and in what era, this is called good. If you jump out and say, "Teacher, I have done this question before and after the ancients, I don't believe it." Because I'm also in our discipline, and if you make a problem like this, there are more or less other people's jobs that have something to do with you. If you want to say that you don't have anything in your head, it's because you're lazy and you haven't done enough literature research.
Fifth, what is your magic touch? Since this matter is very important and difficult, many powerful characters have tried to solve it, but they have not been able to solve it, why did you solve it? what is your magic touch？Your essay only needs to tell one story. Students may say that you underestimate my contributions, and you can't finish an article. Yes, then you write a few articles, one article about one thing. what is your magic touch？Be clear about what your contribution is in one sentence.
Sixth, what are the main results?Okay, so what are your main achievements under the support of this beautiful idea? List it. Three or four is about the same, one or two doesn't matter, five or six is not normal. How can there be so many things? After all, you're an idea.
Seventh, what are the assumptions?There should be assumptions that you get such good results. Some you know, some you don't realize, you should write it out. These assumptions are that when the scholars behind you study this issue again, you can consider relaxing and breaking through to promote, and others will ask if you don't write it. If you don't know, it only means that you don't know enough about your results.
Seven questions:What's the problem? Why does it matter? Why is it difficult? What did someone else do? What are your core contributions? What are the main results? What are the assumptions? You should ask yourself every time you write an essay. When you have answered all seven questions and write an excel sheet, I highly recommend everyone to do so. I also often write an article before writing an excel sheet, two columns, seven rows. There is a column of seven questions on the left and a column of answers on the right, after which you can start writing the article. How to write it?
Seven strands of text:How to write**
The seven-strand essay is still seven questions, and it is a re-organization of the answers to the seven questions just now. Actually, I don't have many articles, you can go to my homepage. If you want to post a lot of articles, I don't think you should follow me as a student or come here to listen to my lessons. I have learned a lot and not many successful experiences. However, I have read many good articles, so I can share with you what the structure of a good article is.
The first part is the introduction. It's about two columns on one page, and it's about two pages on A4 paper. Divided into four paragraphs: the first paragraph: what's the problem? why is the problem important?What is the problem? Why the question is important. In the second paragraph, why is the problem difficult? Why are the questions so hard? It is necessary to clarify the difficulties of the problem. In the third paragraph, briefly mention the overview, what is the state of the art?Then say what is your magic touch?Why did your work solve the previous problem, maybe the work of the predecessors may have ignored such and such problems. Briefly introduce what are your main results?The dots illustrate the breakthrough you've made to this problem. The last paragraph describes the overall organizational structure.
The second section is Literature Review. Literature review, about one to two columns, half to one page of A4 paper would be more reasonable. This section answers the question "What is the state of the art?"."It's a very good opportunity to sort out the academic context. Why you write this article, who you write it for, and who will use it, you have to point it out in this part, so you need to sort out the academic context. Secondly, this part can show how smart you are. I listened to a lecture last week and completely forgot what the presenter's job was, but he completely touched me in the first five minutes. He is so young, and he knows so much about such an ancient direction, it sounds like a living history, which I admire. No one may remember what he said in a few years, but I will always remember this person. Writing an article is also a way to show that you have a clear understanding of how different tools are available to deal with the same problem. He will demonstrate your knowledge of the history of your field, your knowledge of current schools, and your relationship between your work and your existing work. A good literature review can hang your article on a very suitable branch of the subject tree, which not only helps the associate editor to identify the reviewer and the editor to submit it for review, but also helps the reader to locate your job very quickly when searching.
The third part is problem formulation, which may take some work, one or two pages. You should answer "what is the problem" in its entirety. It's not a sentence, you should introduce in detail how to understand mathematics quantitatively, you must use mathematics, don't have so many words. You need to use some space to express your problem in a mathematical form, and you can attach diagrams to show the architecture of the whole system, so as to help people understand the relationship between the problem you are studying and the existing problem. You can refer to other ** that use a similar problem model. Identify who is relevant to your work? Are you the first person to do this? If you're not, you've used someone else's work and you haven't cited it, which is troublesome. You should have learned, this is academic misconduct. Even if it is a drawing, the things that two people think will definitely be different, and they cannot be plagiarized.
The fourth part is the main results。This is the most important part of the article, and it should account for about 2 3 of your full text. The TAC article is generally twelve pages long, and the short one is usually eight pages, and there should be at least 2 or 3 of the content of this thing. It's "what are the main results?".The full elaboration should be made clear: "What are the assumptions?" ”。This is the core of **, there should be theorems, proofs, and algorithms that can be described. Last weekend, I just read a graduation ** sent to me by a student, and after reading it, I directly said that I don't need to graduate from this. The main part accounts for too little of the whole text, and how much of it is your own work to spend so much ink in front of you to write about the background of other people's work? If it is less than 2 3 in the full text, there is no need to send it for review, and it is self-inflicted.
The fifth part is numerical experiments。If your main work is all about numerical experiments, "main results" are "numerical experiments". Very good, this is a very good signal that your work is not in place and the article should not be pitched. The main results are for you to write theorems, and the "numerical experiments" are for practical applications. Some people may have a little more numerical examples, so you need to use other methods to compare, so please be careful. After that, it is necessary to select the examples, try to select some more recognized algorithms, benchmarks, and select practical problems with rich data volumes. Organize numerical experiments with different groups at multiple levels to analyze how it affects your results for parameter changes, measurement errors, model biases, etc.
Finally, there is the conclusion. It's simple, it doesn't have to be long. About 1 4 pages, half a column of A4 double columns is sufficient. Briefly review what problems you solved, how you solved them, and what kind of contributions you mainly made. What should the people behind do if they follow in this direction? Don't be the same as an introduction and an abstract, they serve different purposes. The introduction is to let people know whether your question is important after reading it, the abstract is to know whether the article is worth reading after reading, and the summary is that after reading it, I will decide whether to continue to study in your direction.
A trick:Practice makes a report
Writing a report is different from writing an article, writing an article can have a lot of mathematical formulas, writing an article can have a very long length, carefully ** the connection between various logical relationships. But giving a report is different, we need to learn how to use a more general way to tell the highlights of your work in a limited time, for different groups of people. This is the only way to get good results if you keep training. After each report, the people who give you a good review are, of course, very important. But the people who give you weak opinions are the ones who really help. We should take that feedback and don't just because you're done with the report and don't have to change it. It should be based on the feedback from the completion of this report, and then you will change it, and keep revising. That way, the next time you do a report, you'll do better in at least some directions than the previous one. Most of you may not be engaged in scientific research in the future, but you will have many opportunities to introduce yourself to others. I hope you will be able to practice this ability to introduce any depth to anyone in any short period of time. If you haven't done that yet, I think you should keep trying, and it might really be really always on the road, you have to keep training yourself.
How do graduate students do academics? I started by asking people what they want to be in five years. What do you want to be in fifteen years? Of course, five years from now, I don't want to see my students say that they haven't graduated in the fifth year, and that they are on campus and should not be graduating. In addition to that, I think there are two things you should learn. The first is to make a report. The second is the selection of topics. The choice of topic is taste, and it allows you to judge what is important? What questions don't matter? This ability is very important, very critical. If you can formalize a question and ask it correctly, you can work with someone else, even if you don't know how to write it yourself. There will be smart people who will help you solve this problem together, but only if you are working on an important problem. In the end, you have written the article and been hired, so you don't have to be too happy, or continue to do your own thing. If your article is rejected, don't really be too discouraged. Because constantly listening to feedback, changing ourselves, and finding the right direction of research is something we should continue to do
Article**: Tsinghua University Xiaoyan**, academic school.