Bible of the researcher (1) (ZZ from Mit Ai Lab)
Sun jul 27 09:51:39 2003
Title: The Bible of the researchers (1) (ZZ from Mit Ai Lab)
Sending station: 海星 云 (Fri Jul 25 11:54:09 2003)
Author: Alan Bundy, Ben du Bulay, Jim Howe and Gordon Plotkin
Version: October 11, 1993
Translator: Liuquan Bo Beijing Normal University Information College 2000 PhD
Summary
Get a master's or doctoral degree [1] is very difficult, this document gives a full range of suggestions. The first section describes
Question - What is the paper? The second part states the formal requirements for the graduation thesis. Third, four sections describe
The disorders that students may encounter in this process. 5. Six sections give the proposition and project implementation. Seven, eight and nine
The section is concerned about reading and writing. Ten sections describe the review process of learning and how to deal with reference. most
Afterwards, the list of journals submitted by the artificial intelligence is given.
1 What is a graduation asis?
To get a Ph.D. or a master's degree, you have to complete a graduation papers and through oral reply (research on artificial intelligence "
There is no written test [2]). In the verbal defense, the general asked is required to clarify some of the issues in the paper.
Based on the most important evaluation is the graduation thesis.
Qualified graduation thesis (at least EDINBURGH University believes in qualified) definition, you can see the relevant graduate students in the school calendar
Part: 3.2.6 and 3.2.7. The extracted is as follows:
The basic requirements for doctoral thesis doctoral degrees are:
i. Dr. Applicant must pass the expression of the paper and prove that he can engage in the original innovation in the area in the defense.
New research, able to combine specific research projects with the knowledge system in the art, can be a rigorous and academic side
Express your research results.
II. Graduation thesis must include the original innovation, these work make a clear understanding of the knowledge and understanding of this research
Due to contribution, graduation thesis contains materials worthy of publishing; the paper shows authors to the knowledge and related literature
Separately master; the paper shows that the author can rigorously judgment on its work and other scholars in the field;
The work that is reflected in the work can be completed during the three-year graduate study and research; the text is qualified, the reference is charged
Appropriately, with an understandable intrinsic structure, other scholars in the art can understand the purposes present in the paper, background
, Methods and conclusions.
Iii. The level of graduation thesis cannot exceed 100,000 words. But if the instructor is recommended, the committee recognizes the space
The extension is indeed necessary, and the space can be expanded.
The basic requirements of master's thesis master's thesis are:
i. Applicants expressed through the expression of the paper and the written / oral reply, indicating that they have gained and understood advanced in the field.
Knowledge, able to link knowledge of knowledge and wider research on specific topics, can be used in rigorous and academic ways
Knowledge.
Ii. The paper must be the original work of the qualified records of applicants to study or pass the qualifications of the knowledge in the art.
Rigorous investigation and research. The paper shows that applicants have the ability to use appropriate research methods, with sufficient domain knowledge.
The methods used in papers or statements are completed independently. The literature statements are qualified, including appropriate references.
Iii. The parameters of the paper cannot exceed 60,000 words, and some of the cases described in some instructions, or by mentor recommendations, members
The extension of the context will indeed be required to deal with the topic of the paper, and the space can be appropriately expanded.
How big is the innovation and meaning of Doctoral research? "The materials contained in the" "is a simple judgment"
Break the rules. Extraction of the essence of the paper is an article that can be published in journals. Of course, there is no certain
Standard, the final decision depends on the members of the Defense Committee. The useful information contained in the definition of master's thesis is less. In the artificial intelligence, we don't need a written test. Many master's theories
The text is a record record rather than a rigorous investigation, of course, the latter can also. Similarly, extract the essence of master's thesis
Come, it should be a short paper. But should not be requested to be published in the journal, the Working Paper in the department can
.
If you still don't understand the standards of journal papers or Working Paper - quickly read some! Also read some graduation
paper. Don't be scared by the graduation thesis of American universities, their students have to spend five to six years, and our English
The country is generally three to four years.
2 Procedures for studying phases
At the University of Edinburgh, full-time doctoral students should study at least 33 months, non-full-time students are at least 45 months, for Master
The full-time student is at least 21 months, and non-full-time students should be at least 33 months. During this time, students are generally in love.
Dingbao. Generally demanded three years of doctoral studies, completed master's degree in two years. Students are only valid during this period. Doctoral
The longest learning period is five years, and Master is four years.
In the past, postgraduate students with artificial intelligence rarely submit their own papers within the prescribed time limit. In order to reduce the time spent
Each student's research work will be organized in the form of the No. 16 notified, and each stage is listed in the notice.
The requirements that graduate should meet. These requirements will have some subtle adjustments from time to time, so this article does not contain it
Inside.
3 Common issues of graduate students
For the first study, the research process is full of traps. Understanding this helps avoid these traps.
3.1 Solve big problems
Many students' topics are too big, especially for artificial intelligence. You must know that there are many people who are for humans.
Things, it is too difficult to model the model. Obviously, help you choose the right papers to fall in your mentor shoulder
on. In addition, you have to read the literature, exchange with peers, and understand what is the latest developments. "Further"
The work "part is a great source that produces ideas. Reading the literature should be rigorous. Re-completing the original unsuccessful work, too
A source of topic.
3.2 heaven pumps
After choosing a good question, what is going on next? In the house, it is dangerous, holding a pen and paper, looking forward to a good idea from the day, this
It is obviously unrealistic. What you should do is:
a) Read the literature and take a problem reading, for example, is there a mistake in it, how should I use this document?
,and many more.
b) Talk more to others, don't be stupid. Don't think that your idea is too shame, others are too smart.
Where is it?
c) Try the simplified problem of the selected topics.
d) Write down your own ideas in the form of a working paper. Imagine yourself to explain your thoughts to others. You will be surprised
It was found that the original unclear thinking was formed, and the defect was also revealed.
e) Report your ideas to a group, the effect is the same.
3.3 indulge in the computer all day
The computer is very attractive. A computer is obsessed with the time to debug the program, adjust the input / output routine.
When the bug is cleared or the output is very beautiful, the computer idiots will have a sense of accomplishment. This is illusory! As a research
In the case of your program, your program must be interpreted in a higher level than the code, so that you can make real knowledge.
contribution. Prior to the theoretical program before being implemented. If some parts must be verified, verify
After you finish, you should leave your computer and continue to design theoretically. If you think this is too difficult, try to
His friends, describe their working principles in the way of the paper or in the seminar. If people can't understand, that is you
Problem - Come on.
3.4 Homemade a language
Another "indulging in the computer" is to write a computer programming language. Yes, the existing language must not fully meet your requirements, but this does not become a reason for your own language. You can always find the right
Alternatives, just satisfy your requirements. To write a useful new language, it is not only necessary to have an encyclopedia for AI.
Solution, there must be experience in system programming. No one will use the language you wrote - even you will not! all of you
Energy will spend the language design, and your project is forgotten. If you really think that the existing language is not
In order to meet the requirements, write a papers in detail the defect of the language. If your opinion does make sense,
There will be a language to enhance it overnight.
3.5 Issue
The process of collecting experimental data is full of fun, there may be various phenomena. To make sure you know what you want to get according to the experiment
What kind of result.
a) explain to others that you think what results will show.
b) Suppose the test has been done, the data is ready, what do you want to do with it?
c) Not only do one or two people, but also to analyze. Don't dream trials will be present.
3.6 ivory tower
Independent thinking is a good thing, but if it is completely closed with others, it is wrong. Only keep in touch to understand the most in the field
new progress. Talk to others. Selectively participate in the discussions and reports. Time to read comments a week
, Abstract and selected papers.
3.7 Unintended genius
It is easy to think that others don't understand your research because you are a genius, others are grass bags. In fact, there is still its
Its possible:
l Preferences. The artificial intelligence field is full of a lot of jokes: try to express its own ideas with ordinary English; taste
Try to re-express your ideas with others. Try it, is the effect is not the same?
l If I can do it, it is definitely negligible. Once the problem is found, it will feel insignificant.
Then think: This is too easy, let me try again. This is endless! Your solution is concerned about others
It is not insignificant (may be wrong or too complicated) and can be used as a basis for further work. motto
: Do it first, then gradually increase the difficulty.
l Like to play complex. Write a complex program is not a virtue, and it is very annoying. Do things in the simplest way
situation. The OCCAM principle is completely applicable.
3.8 Lost in abstraction
In order to reflect value, your research must be in order to understand some attributes of intelligence, for example, search control, knowledge
Representation and learning. But if you want to succeed, you must grasp the abstract properties in the specific situation, that is, you must
The program must be built to try the search for control, knowledge representation, and learning to learn. I want to grasp the problem in the abstraction, only
There will be no things, eventually fail.
3.9 Over the target, no matter
There is a high standard for the final result, but if you set up a high goal with a high goal at the beginning, it is very possible.
Can't start at all. First make simple things, then apply your high standards, making it gradually refined as valuable results
.
3.10 Methodology is not a paper
Since artificial intelligence is a relatively relatively young discipline, in a sense, it is essentially a cross-discipline. people
Workers have no universally accepted research framework or to define good methods. As an artificial intelligence
Students in the industry, one of the difficulties you face is how to form a research method suitable for the study. Form a combination
Before the methodology, there will be many other methodological or philosophical perspectives, many of whom are unbearable or
It is easy to cause misleading. However, you will find these views of supporters, and maybe sit in your neighbor.
Therefore, the formation of methodology is often formed in a fierce confrontation with the peers. In this struggle, you have gradually formed a worldview of artificial intelligence and related philosophical issues, reflecting in your reading, participating in seminars and other work
in. In the years, this will become the guidelines you believe, but even some micro changes. But when you start
When writing a graduation papers, you may feel that you need to expand your life philosophy. To restrain yourself - the defense committee
Members may not be able to meet you with you. A brief summary summary of the hypothesis of the methodology she holds, for appropriate
The argument or framework should be given to the reference, and the expression restriction of its own methodology is indeed contributing to the spirit of understanding the thesis.
Within the range. If your point of view is too radical, you may need a fifty page space to elaborate, and always
It is so convincing.
3.11 The process of discovery is not a reason
In the process of project implementation, it will gradually form some views on technical issues, some are novel, most of them have
Conceptual review (or new understanding). In the description of the paper, keep in mind: support a certain point of view, technology,
How do you believe that the process of this view is different. The members of the reader or defense committee are rarely interested in the latter - he
It is interested in supporting the general argument of viewpoints. If you start to convince a point of view, it is easy to think that you find it.
The process is the best reason. At this point you need you calm down, then give a reasonable elaboration in the argument, especially when your point of view
At the time of conflict with others.
4 psychological barriers
Like the creation of novels, writing dramas or paintings, engaged in research will also encounter psychological problems.
4.1 mentality
One of the skills required by a researcher is to properly treat your work. An important factor in maintaining mentality is
The belief of the work. Have a little pride! I believe that the problem that I discussed is important, you solve the problem
Significant contribution. Otherwise, how can you have a move with your way? Get confidential first step
It is a research topic that chooses yourself (see Section 5). Of course, it is impossible to have everything else, can't hear any batch.
Comment. Be a good job once you find your own ideas, there is a mission-building preparation.
4.2 Wildun in the study [1]
The lack of research success will make the researchers feel weak, in fact, this is the result of self-implied. Self-energy
The doubt, will put themselves in dangerous situations: the investment and enthusiasm of obtaining research results have disappeared without trace
. The way to get rid of this situation is to recognize that the research ability is not relying on a mysterious existence. It is acquired after heaven.
skill. According to this booklet, you can also do the research of primitive innovation.
4.3 Coping to criticism
It is always difficult to accept criticism, but this is not equal to ignore the criticism. If you want to make progress in research, you will
It is necessary to learn to ask criticism and think carefully, and must learn to distinguish useful and useless criticism. If you don't distinguish it
, Ask your friends to help out your mind. If the criticism is unreasonable, maybe because you misunderstand your explanation, you can do this
Do you improve?
You must learn to withstand some combat: the papers of the journal are denied; the pain when suspected. At this point, learn to smile
In the face, don't give up - I don't see, how many famous scientists have endured unbearable accusations and criticism. real
In the event, sometimes criticism and accuses are completely due to the contradiction between individuals - every scientist wants to win
Others have a head. You have to take this as a comprehensive test of you.
4.4 Early morning - calmly
Almost all people have found that it is difficult to settle down to start a day. But once you start, everything is normal. solution
There are two methods for decisions:
1. Adhere to the regular working time, not necessarily nine nights, but must have a fixed start time every day. otherwise
You will find yourself that you haven't worn. 2. Do some attractive work. For example, don't write a more difficult part, you can deal with you, you will be a small dish.
Do something or draw a chart.
4.5
If you have no research methodology in the selection, you are sure you want to create one yourself. For mathematicians, this
The desire is to prove the theorem. For engineers, they are kept trying to try. Be careful! only
Theorem that does have related to your research is to prove. For example, prove the result of a process if
The process is helpful to the research, otherwise it is not related - there is nothing necessary to prove.
4.6 fear exposure
After having a good idea, you need to test it through the agency certificate, write programs or explains to a friend to test it. but
Yes, sometimes I don't know why you stagnate, you find it hard to start working. Think about it, is it?
Due to the subconscious, you may have no valuable fear. Difficult experience tells you, problem solving
The law always arrived in midnight, and the time of dawn disappeared without a trace. Take some courage! Research is this
I have ahead of ten steps, and I will refund the nine steps. The sooner you can use the experiment to test your own ideas, you can find it earlier.
Foot and problem.
Sender: Embedded (in the end), the letter area: paper
Title: Bible of the researchers (3) (ZZ from Mit Ai Lab)
Sending station: 海星 云 (Fri Jul 25 11:56:53 2003)
5 Select Research Project
The selected research project must meet the following criteria:
1) desire.
2) The problem to be solved is a doctoral study level.
3) It must be the latest, that is, it can be done in three years.
4) Some people are willing to guide you.
The importance of the first one is not too much in any case. You need to pilot all your enthusiasm for a difficult,
Keeping power during the study of loneliness and irregularities. Choose those questions that you think the most important, choose those
It can take advantage of the field you are good at, for example, for mathematicians, with mathematical reasoning, for language
For the scholars, the understanding of the natural language. Choose a new field to be careful, soon you find yourself need to go
Learn the most foundation.
After choosing a domain or problem of interest, you must find the relevant items. This is the need to play a role.
The place. Find a tutor in the department in the field. He may provide some projects for you to participate
Test, you can also give you opinions from value and feasibility. Overall, the study of research and research
Health always underestimates the value of the project, the feasibility of the estimated project - even if it looks a lot of mature projects
Tone a lot. So I have to listen carefully to the advice of the instructor, don't choose too big.
Please advise some reading materials. You can find the right from the "further work" part of the paper and graduation thesis.
project. Problem to continue to study others is a very good research method, find some difference you think.
Work, follow yourself to think the correct thinking is over again. Maybe you can simplify the other's procedures and work with others.
Tailor, or develop a more powerful program.
There are a lot of ideas that want to be thriving, and build a hierarchy of research targets. This not only formed a frame of work
Configuration, and when you find (inevitable), someone can act as a protective network [1] when you have tried a deeper level.
The project must avoid the following may lead to various situations of "bad" research: the problem of research and solving is independent, and the study is not
There is based on the former work.
6 research methodology
You must choose one from a number of research methods. Many people will try different ways when they start, but eventually become one
. For example, you may first analyze certain aspects of human performance, then use others theory; maybe you will rationally reconstruct others theory and find out and disadvantages. "Reconstruction of rationality" is always effective, this
It is also because a regrettable situation often exists: in the published papers, how often is how to achieve work and
How to do more performance, and there are very few descriptions of assumptions and principles. However, no matter what, you must first find
Find a theory.
The following is an example of methodology (provided by AB).
6.1 stage 1
Imagine the following situation - The output of the design program indicates that the program has modeling capabilities, and in mathematics reasoning
Proof, instance conversations in natural language and scenario identification in the visual field.
6.2 stage 2
What kind of theoretical hypothesis helps to process the above situation, the theoretical hypothesis may include the process and data structure
" It is necessary to consider problems as much as possible. The problem encountered is considered a general case. If it is not true
Resolved questions, do not use special processing methods, if possible, try to use existing artificial intelligence methods
. It is also a scientific research achievement that proves that existing methods can be solved.
6.3 stage 3
Consider further situations, see if your planned solution can handle the above hypothesis, according to the hypothesis of the solution
Row constraint, generalization, expansion, and debugging.
6.4 stage 4
If you feel that your solution is reliable, choose a suitable programming language to implement the program, and
Debug according to stage 1-3.
6.5 stage 5
Run the program on some instances that have not been considered before. Modify your own procedure until it is strong enough -
It is to say that the program can run in many instances and will not crash each time there is a new input.
6.6 stage 6
Evaluate your program by running in some instances. Statistics on the number of success / failures, and analyze each
The reason for the failure, and all discovery is described in the paper.
6.7 stage 7
Describe your procedure with a language that is unrelated to a specific implementation. Do you have to introduce any new technologies and be used in this field?
The technology is compared. If time is allowed, you should also apply your new technology to other areas.
Note: Stage 4 and 5 take time far beyond your imagination - year-in-year - so to keep enough time.
7 writes
Research papers is our most important product, which is a quasi-rope measuring the progress and achievement of our personal / research group. It is because of the paper
The status is so important, you should use most of the time during graduate students to write the paper. Write the paper is with other artificial wisdom
The most important way to researchers is also a good way to clarify and debug your ideas.
Books, graduation papers and journal papers, not only different thicknesses, and there is also a difference in writing. First of all, you must clear this system.
Distinguishment, details of the reference. No. 8 notice. The following is a non-formal description, not necessarily reflecting the latest
Happening.
The writing should be written as part of the schedule life. Record everything you make: your own thinking
Road, program document, lecture note, paper notes. There are several purposes: assist memory (you will be amazed to forget yourself
This fast), clear thinking (you will often find when you try to explain something to yourself or others, ask
The question appeared or resolved) and the starting point for writing a technical paper. To ensure that you record clearly, you can read it later.
Read, but also to ensure that the archive is correct for future findings. I personally discover, it is very effective to enter the note into the computer (for this purpose)
You should learn Troff, Latex - or uncommon-TEX).
7.1 research papers
Refers to the paper submitted to journals, meetings, and books. Other types of papers can also be attributed to this
. This provision encourages the teachers and students from being published outside the department, and provide the pre-copy of the papers for other researchers.
7.2 Teaching papers (previously known as papers called special occasions)
Refers to those teaching materials, such as lectures, programming preliminary guides, etc.
7.3 Soft Papers (previously known as technical papers)
Refers to the software related to the software developed, such as program design preliminary guidelines [1], manuals and editing tools, etc. Also included
The special soft / hardware described in this system, for example, logo tutle, and robots.
7.4 seminar
Refers to the paper for internal communication and discussion, as well as student projects and the paper plan.
7.5 Working Paper
This is a description of research work, including completed and is being implemented. At this point, don't be afraid to show that it is not yet mature.
Idea. It should be developed into a habit of writing your own ideas into a working paper. If your thinking is unclear, it will soon
Some people will refer to you and provide better ideas.
7.6 Publishing Paper
The paper must pass the advice of the instructor before publishing. When you and your tutors think it is worthwhile to publish
It should be submitted to the journal. Select a journal from the list listed in Section 11. When preparing to publish a papers,
Remember that people who have contributed to the paper have been recognized, for example, your mentor, and any formation of ideas.
People who have contributed, and those who have commented on the draft draft, etc. If a person's contribution is quite obvious (for example
Help your mentor), consider whether you need a joint signature. Remember to thank you for your support. If there is
Some places are not allowed to find mentor discussions.
The number of copies of the paper will generally pass to the journal, and then the journal editor will transfer it to several reviewers review. in case
Rejected, don't be too depressed - happy. Read the reviewer's comments carefully. They are right or
They misunderstood you? Is the selection of journals suitable for your papers? After the submission is rejected, you can consider transferring other issues.
Journal, but must first consider those reviews that they think about the correct.
7.7 conference papers
Conference papers, compared with the journal papers, the status should be some. The meeting often considers the description of the progress. meeting
The paper is usually sent to review as journal papers. The conference papers and the permanent speech at the meeting have a length limit (
From 5-15 pages, 10-30 minutes), so be sure to prepare. Reporting your own papers at the meeting is very valuable:
You can get feedback from more than usual listeners; you have more opportunities to meet with big people, making it easier to apply for fund funding.
.
[Booth75] is a guidelines for the style and statement of scientific papers, which can be found in [PARSONS73].
How to write useful information for graduation thesis. All of these documents can be found in the Forrest Hill Library.
8 Writing Guide
In the process of research project implementation, there is a need to write a lot of documents: the paper plan and the paper outline, research notes, thesis
Post-sensitization, conference papers and journal papers and the graduation papers to be completed. Write journals, conference papers and graduation theory
The standard is not high, but it is the main reason for the three papers. Although the graduation thesis is poor, although it is not necessarily meant
Full failure, but over and over again, it will make you feel exhausted. Writing ability is not good and will make others
It is difficult to understand your work. Therefore, the cultivation of good writing ability is very important. This section describes the improvement of writing
The tricks and rules will never have more comprehensive insights than this. For more information on writing technology reports, please refer to
[Cooper64]; Facing a large number of audience, how to express their thoughts in English, can refer to [Orewll68].
There is no shortcomings in the cultivation of good writing skills, and these rules to mention below are not difficult. However, please note: If you want to break the following rules mentioned in the writing process, you must think twice.
And behind. For example, you want to use terms to live in the town, and don't want them to know what you didn't get.
achievement.
l The paper must be with information content, that is, have its own point of view, and can be proven by his research.
First, you must understand what the information you want to convey. Summarize it in the paper, or talk to friends
A friend listened. The views have been reflected in the title, summary, introduction, conclusion, and the structure of the paper.
l Only to make your own papers are not enough - must also express in a way that will not be understood
. Make sure all the key points are clear, significantly, and often expressed. Don't include several critical in a sentence
point. Don't assume that a key point is very obvious without a description. Say what you want to say, let it come, and then say it.
Repeat it.
l Don't try to make all the questions clear in a papers. Adhere to the main point of view, only related to this is essentially
The content is included. Leave other questions to the next paper. Readers should be able to grasp from the first page of the paper
The Main Tropic to the Thesis. The lengthy and scattered profile should be ruthlessly cut.
l The basic framework of science papers is: What is the problem, how do you solve it?
l In order to keep the technical standards of the paper, there should be an imaginary target reader in the mind when writing.
l Write the paper does not have to write down from the head according to the order of the class. Especially the introduction part, when you are clear, you have to get
This part is readily written after the conclusion. Therefore, when writing the paper, first describe the central idea, for example, the main skills
Surgery, process and proof. Then think about the reader to understand what the center must know, and then put this information
In the profile or chapter.
l Use the actual example to clarify a process, but this cannot replace the formal description of the process.
l Clearly explain what is innovative in your work, which is better. Will be related to similar work
Clear comparison.
l If you find that you have repeatedly used a long name word group to represent an entity or thinking, this shows that you need to define a new
Terminology. Do not define new terms unless you really need it.
l Learn to use the keyboard (all 9 fingers), screen editor, word processor, and spellper. Put your own papers (
Direct, or enter the computer according to the note or according to manuscript. In this way, saves when modifying the school.
time. The final papers must be inspected by the spell inspector.
l Please take a look at your paper. Consider spending time to integrate their opinions into the body. If they don't read
Know, it is your problem, not their problem. Haven't considered the opinions of others asked people to read your argument
Wen is a very rude behavior. The draft of the graduation papers must pass through the tutor's review, there may be an inspection of internal reviewers,
Generally do not check the external reviewers.
The following content is related to all types of writing, but mainly about graduation thesis.
l The graduation papers should not be the stack of all materials you know about the topic. On the contrary, there should be a center of the paper
The argument, you must carefully consider how much the paper's role is in the arity. Remember the writing of the graduation thesis and
It is not a member of the Defense Committee. You should write the writing of the graduation thesis as a statement on later researchers. Such as
There is appropriate reasonable arrangement of the paper, and the arguments are clear, they will hit the festival. Also consider reading your paper and reading others
The knowledge status of the teeth is different, so the materials in the paper must be detailed enough to avoid reading for others.
Causing distress. There is also a little in mind that researchers around the world (at least satenation) will judge the level of research in your institutions and department according to your work. Even if you don't think it is, the defense committee will think so.
You must also think so.
l You can write your own graduation papers on top of the top, bottom up or two. Since the top, it means you from notes
Start, then gradually explain the chapter of the paper. Since the beginning, you first describe all aspects of our work,
The posts will become a graduation papers. One of the above methods alone is not particularly successful. Self-top, sometimes
How to play, because in the explanation, you have changed for the completed work. Since the bottom up, it will change
It is hodgepodge that is not related to each other. Therefore, two methods are combined more efficient.
l In the process of doing research, you should write your own ideas and results into a series of notes and work papers. Part of the department
The discipline has a level published at the journal or the meeting. Set these notes and the thesis into a file (paper
Or electromagnetic), this is the graduation thesis. This belongs to the work upheld, which is next to the top of the top.
l Refine the arbitraries of the graduation thesis should have the following characteristics:
- There are a few sentences in the argument, which means that the length of the summary is similar.
- These sentences form each step of the argument.
- Each sentence should outline the contents of a part of the paper (probably one chapter).
Discussion on the title, summary, summary, conclusion, and overall allocation of the paper.
l The thesis point will play a role in the following aspects:
- Ensure that parts of the paper are integrated in a consistent manner. It also shows how different notes and papers should be
The new organization has become a graduation papers.
- It can answer such as "What do I have done?" And "why these work is governed
Bit? "This problem. At this point, you should have a clear point of summary, introduction, conclusions, and title.
Sate.
- He can answer what questions should be discussed in "further work" in 'further work'. " In fact, you must be precise
I know what the role in each chapter is in the whole, that is, I want to prove anything.
l The thesis is very short, and it is easy to edit. You can adjust it without stopping it until you have a surprise.
l Now returne to the upward activity - a number of existing materials, writing new materials, to meet the needs of the argument
want.
In order to experience the above description, an example is given, the famous artificial intelligent researcher fr. Aloysius H
Ph.D. in Acker.
Example
"The Computational Modelling of Religious Concepts"
BY
Fr. Aloysius Hacker
1) We understand the religious concept with our thinking.
2) The interpretation of religious concepts, such as sacred triumphants, often encountered philosophical problems.
3) Problem has caused problems due to lack of appropriate term. Calculative terms often provide a suitable near
like.
4) Although there are still some problems still exist, for example, free wishes.
5) We have seen a new computing theology that is appearing.
Every point in the above five points corresponds to one or two chapters in the papers.
The first chapter introduces the general concept of computer modeling, and reflects the ratio between computational concepts and religious concepts.
The causal relationship and non-conditioning relationship between religious status, thereby successfully applied computer modeling to religion. And therefore
Corrected some of the theoretical debates that lasted for more than 2,000 years.
The second chapter describes "related work", and summarizes many concepts related to controversial (for example, the sacred three).
The theoretical insights of the Harrants and pointed out the contradiction between these insights. Third, four chapters are the core of the graduation thesis. Chapter III has made three integration with the triple delivering function, and here
Solving God is a philosophical problem with three entities.
The fourth chapter further expands this kind of compament, regards the universe as a program, God is a programmer, the halve is seen as a program
The runtime patch is inserted during interruption.
The fifth chapter is "further work". The problem of highlighting is pointed out, that is, discussions about the issue of free will, and
Possible computational instructions.
Chapter 6 is the conclusion section. Summarized the results obtained, as well as the computational approach to reveal the success of religious issues. These ones
Work shows that an important and novel theological research method has already begun.
9 Reading Guide
Tight tracking related research is one of the most important sub-objectives of doctoral degrees. "Research difficulties" have been held in the system
At the meeting, there have been some difficulties in reading the habits, quoted as follows:
"Reading is very difficult: With the difference in academic development, the difficulties faced by researchers are different. I don't know when I started.
What (many documents are not published), later read attracted, and became an excuse to escape the actual research. Final study
People lost their time and patience (sometimes afraid of finding others) "
.
Obviously, in addition to reading, there are many ways to keep track of related research, but there are similar problems. One
The researchers must learn to maintain a moderate balance between the work and the promotion of their work.
The research of others can be arranged on different concentric circles in accordance with their own research.
The lower the degree. For example, if you are studying the lookups of the privilege repeat method, then the inner circle should be repeated with the privilege.
The work consisting of the work, the medium circle is composed of natural language and calculating semantics, and the outer ring is used by artificial intelligence and semantics.
Research composition. You can also try to increase other outer circles. Obviously, the tracking of the outer ring research should be compared
There is less inner ring, so different learning techniques should be used for different circles.
9.1 outer ring
The moderate understanding of the research work of the outer ring can be maintained by quick browsing or reading the paper abstract. Recently artificial intelligent papers
Abstract writing level has improved. The publisher and journal lists of publishing AI summary are included in Part 11.4. Take it every week
It is a good idea to go to the library to quickly browse the latest journal. It is a good idea. Another fast acquisition letter
The method of interest is to meet the short-term statement and long tutorial nature reports at the time of participation. With someone else in a coffee
It is also very valuable to talk to their latest ideas on the house or corridors. About AI's background knowledge can be in one
Found in the textbooks, for example, [BODEN77], [Winston77], [Bundy78]. Of course, the above materials are
Very basic speech.
9.2 medium circle
At this point you should spend more time. The methods mentioned above still apply, but it is not enough - you need to come from
Read some papers at the end and have a longer conversation. You have to read some more specialized textbooks and participation in seminars.
Keep a complete record of the read papers, and make some comments, otherwise, with the blur of memory, some body
Will be dissipated with the teachings. Some people think that the index card is useful at this time.
9.3 inner ring
It is impossible to understand the paper, just reading it again. You have to read it several times, and you must go in. See all examples
Son and do some exercises. Contact the author to contact the author. For some questions and / or criticism, talk or write to the author.
Be deeply understanding of a work, the most value method is to teach others. Do a report, whether it is formal or not
formal. The paper often used is sure to hold a copy, if not, copy a copy.
When reading the papers, you will find that if there is a problem in your mind, you have a problem that the paper answers is, then the understanding of the paper will be more profound. The specific problem depends on the situation at the time, which may be like this: "Can I have this technology in my program?
"How did he solve X problem?" "" Is this qualified as a journal paper? " "How should I watch when I discuss
Is this idea? "
Finally, don't be afraid that you are so ignorant because of the problems mentioned. Everyone is very sensitive to the fields of ignorance. people
Worker intelligence is a cross-discipline, so we must have a lot of ignorance. People all like answer questions -
This will make him feel important. Discuss with people who know more about a job, read papers than themselves
Harvests are often more.
10 review of graduation thesis
When your graduation thesis has been speaking, the mentor will finally be submitted to the papers after satisfying it. Notice Office
Say you are ready to submit a graduation papers. To ensure that your paper is in line with the relevant provisions. Two points of graduation theory
The text is delivered to the system office.
Your mentor and department will choose a pleasant member of the right and foreign unit, and they may be informally to negotiate with you.
The list of review members. When all the review members have reviewed - it may take several months - the judgment
The committee is responsible for arranging a reply involved in all review members.
The reply is a question and answer between your members and the Defense Committee to last for a few hours. Your tutor can participate as bystanders
This depends on whether the Defense Committee agrees. The defense is usually held in a office in the office. In the defense
Heaven, some foreign review members (may have you) need to come from outside. Wearing for normal work,
The atmosphere of the defense is relatively easy. If you can make you feel more comfortable, you can dress up a little.
Before and after the defense, the defense committee must submit a report to the system. The report after the defense is a joint sign of the Defense Committee.
The department contains the final opinion of the paper. Conclusion is usually one of the following:
1) It is fully accepted.
2) Accepted and require some micro modification.
3) Accepted the paper, but the defense will not pass, and the respondents will be further examined.
4) Consider the extension period and re-submit the paper.
5) Consider the reduction of doctoral degrees as a master's degree.
6) Reject.
The first, 3rd, and six conclusions are rare. You will be informally informally inform the final conclusion, the conclusion is in theory
Or the possibility of overthrowing the school's academic committee (but has not yet happened).
The second conclusion allows for correcting the typesetting and spelling errors in the paper, and the small part needs to be override. You should in the original papers
Copy of the copy, you can paste the new page if you need it. The review member of this department will have completely changed the paper
After checking, then the teacher responsible for handling your paper and inform you the final conclusion. This may need
A month of time.
The fourth conclusion requires that the paper is all rewritten and sometimes needs further research. You have to rewrite, reread, heavy
The new submission of the graduation papers, follow the same defense committee to repeat the process of repayment. This is your last chance. second
The second reply is no longer possible to have the fourth and 5th conclusions.
The fifth conclusion is for those who are considered to be uncomfortable, but feel that they can grant a master's degree. This
It should not be seen as a consolidated prize. Sometimes further research and rewriting. You need to re-bind the paper (
Use the cover of the master's thesis! ), Re-submit, organized by the defense committee composed of the last different committees
reply.
The purpose of the defense is to let the council think that the graduation papers have a Ph.D. / Master. Especially a member of the Defense Committee
They will make various questions in their reply, these questions must be answered in the process of defense.
It must be no longer mentioned in the report after the defense. If the report still has questions, the system is awarded the degree.
Will not consider their opinions.
The defense committee will ask questions about their questions and hope to be satisfactory. Due to time limitations, it is usually started with the most serious and / or most general problems. For example, they may first ask you to use your own words, you think you think
What is the most critical contribution to your paper? Therefore, you must prepare a concise answer to this in advance. you still can
Together with the instructor, I hope that other questions that may be asked. However, in general, your most worried question is already
The apere explanation in the paper, so the problem asked when actually defense is often unexpected. In this way, the defense
It is best to sleep a good night, and you can keep your vitality and alert when you reply. This can be practiced over again.
The answer will not be questioned.
Don't answer the problem there. Pay attention to the questions and instructions of the Defense Committee, give a concise and pertinent answer
. If the defense committee believes that you have a consistent and high level of understanding of the problems in the research, they will tend to be awarded
Your degree, even if there are some small mistakes in the paper.
The defense is a bit like a debugger. Graduation thesis is the program, you are programmers, doctoral / master's standards are language syntax, answer
The debate is an interpreter. During the defense, you have got different error messages. This information is not necessarily true.
- It is possible to be treated seriously due to misunderstanding - but must be taken seriously. Suppose each error message will make your paper
A modification. Of course, you want to only need little modification, but let this expect to blind yourself and ignore the problem.
It is more based. Don't take the defense committee. You can't rely on a big or sweet words, it's just like you can't
It can make a computer with a bug program by power or perseverance. What you have to do is: Rule your own ideas, eliminate
In addition to misunderstanding between your committee, to ensure you know how to modify your own papers, then modify it.
. The reply is a cooperation process, and the Defense Committee wants you, so you should give all the help them need.
.
