Thesis writing experience
Article directory
foreword
I took the "Thesis Writing" course taught by Teacher Min Fan for half a semester. Although this course is called "Thesis Writing", after all, I haven't written a thesis by hand and have no practical experience. experience.
1. Experience
- First of all, you need to clarify the direction, which direction you want to study. Before writing a thesis, you need to read a lot of high-quality papers, such as Dingkan Dinghui’s articles, like my research direction: large-scale graph neural network GNN, robustness research on small-scale graph neural networks. I will look for articles such as ICML, ICLR, AAAI, KDD, IJCAI, IEEE, NeurIPS, and I used to read the latest papers on arXiv, but the articles on arXiv are mixed, which is not very friendly to a novice like me. , I have read the article on arXiv a few times, and never read it again.
- How to read a paper? Generally, by looking at the abstract of a paper, we can clarify the general starting point of this paper, the innovation points, and what kind of problems it solves. We can tell from the abstract whether this paper is our own research direction and whether it is the paper we are looking for. If so, then we can directly look at the model diagram, innovation point (calculation formula), and experimental part of his paper. Of course, this is based on the fact that you have read your pioneering work in this research direction in detail, and are already very familiar with the previous preliminaries.
- Should I write the paper directly in English or in Chinese? If you only want to publish papers in Chinese, please ignore it. Although Mr. Min Fan said that English papers are better to write than Chinese papers, due to some reasons of my own, I am not familiar with English grammar enough, and it is easy to make grammatical mistakes. I guess I will write Chinese papers first, and then translate them into English papers , but the Chinese culture is extensive and profound. It is true that Chinese is more difficult to write, and it is easy to have ambiguity and unclear expressions.
- In terms of English sentence structure, in fact, short sentences are better in most cases, and long sentences can appear, but not too many, because long sentences seem to be more laborious. A paper should be concise and easy to understand. The better the paper, the clearer it is. Of course proper packaging is required, but not excessive packaging. After all, a reviewer is also a human being, and he (she) doesn’t like to read some gorgeous and excessive sentences. Everyone likes to read papers that are comfortable to read.
- When writing a thesis, don’t write it after completing the experiment. Once you have an idea and start the experimental part, you can start writing the thesis. This needs to be done at the same time. It does not mean that I can only complete one paper in one period. To achieve high quality and high productivity, I need to divide a period into multiple periods and complete the corresponding work in each period. For example, I have already started to work on the experimental part of a thesis, so if I have a new idea, then I can start the outline of the next paragraph.
- Regarding the writing order of the paper, the abstract must be written at the end, after all, this is the summary sentence. First write the introduction and preliminary first, then you can write the experimental part first, then the method part, and then the conclusion.
- It is very important to choose the thesis software for writing. Mr. Min Fan recommended Latex to us. Compared with Word typesetting, Latex needs to set various complex styles and appearances. Using LaTeX typesetting, we don’t need to worry too much about the appearance of their files. is to focus on editing the right content. The appearance format of the document is independently defined by the template style file, so even if we do not have rich experience in typesetting, we can easily produce beautiful documents at the publication level.
- It is inevitable that the editing and derivation of formulas will be involved in the thesis, and we must explain the variables in the formulas in the thesis in the thesis. We can use the Latex editor (the number of formulas is small, the formula is simple, and can be used by those who are proficient in LaTeX grammar); we can also use the formula editor that comes with the office; we can also use MathType (this is relatively simple and easy to label). This method is in In the process of entering the formula, you can directly input a little bit according to the above symbols like a fool.
- How to determine the title of my thesis? When the journal editor of the materials science journal Advanced Materials saw the manuscript titled "A unique DNA structure for drug delivery", he had already begun to prepare to reject the manuscript on the grounds that it did not meet the readership of the journal; when the title When it is changed to "A photosensitizer-loaded DNA origami nanosystem for cancer therapy", the editor will be interested in continuing to see what structure of DNA origami nanomaterials can be used as a carrier to treat cancer, and the probability of being submitted for review is very high. Therefore, a title that is appropriate and catches the eye of the editor will often increase the possibility of being submitted for review. The title must convey the main content of the scientific research project to readers, and should be clear, rich in content, and reflect the principles and methods of the scientific research project. The title should tell the reader whether they should continue reading your paper. A good SCI paper title not only allows readers to easily know the results of scientific research, but even allows them to speculate on the research hypothesis. Therefore, all words in the title should be chosen carefully - a title that is too short will make it difficult for potential readers to grasp the main content of the article, while a title that is too long is sometimes even more meaningless. Remember, the title of a scientific paper is not a sentence, let alone an abstract.
- When writing the abstract of a thesis, the number of words does not need to be too many. We must grasp a principle: be concise, and do not use 2 or 3 sentences to describe things that can be finished in one sentence. Very verbose. In addition, if it is a Chinese abstract, generally within 300 words is enough, and if it is an English abstract, it should be within 250 content words. It should also be noted that no figures, tables, chemical structural formulas, etc. can appear in the abstract. The most commonly used method may be the jack-of-all-trades method. This method is also called the outline method, which is to sort out and summarize the main research content of the paper according to the outline/table of contents of the paper. Basically any type or major of thesis is applicable. Basic sentence patterns: This article analyzes XX, adopts XX research methods, and finally obtains XX results.
- How to write keywords when writing a thesis? This is a very detailed issue, so detailed that we will ignore this issue, and never even think about how to write keywords. The number of keywords is 3-5. According to the call for papers for excellent journals and conferences in the past, the summary of the journal articles I have read, and the standard writing requirements mentioned by the teacher in the thesis writing course I have attended, everyone basically reached a consensus on the number of keywords for small papers : 3-5 pcs. Why are there keywords? It is mainly to facilitate the retrieval and screening of later scholars. When we conduct literature search on HowNet or other databases, the first thing we see is the complete title, abstract and keywords of the paper, and it is free. The main reason is to facilitate your search and screening, and then you will decide Do you want to download this article and read it, refer to it, and quote it.
- How to determine keywords? Keywords are the emphasis and prominence of the core content of the article, so the general selection of keywords can have the following two small strategies: one is to select from the title, abstract, titles at all levels and repeatedly emphasized content of the text. One is to select from the research background, research object, research question, core research content, research purpose, research perspective/theory, research method, core research conclusion, etc. of the article.
- How to write a good thesis introduction? The main task of the introduction is to outline the basic content and outline of the full text to the readers. It can include all or some of the following five items:
- Introduce the background, significance, development status, current level, etc. of a research field;
review and summarize the literature in related fields, including previous research results, problems that have been solved, and make appropriate evaluations or comparisons;
point out that predecessors have not The problems solved and the technical gaps left behind can also raise new problems, new methods and new ideas to solve these new problems, thereby eliciting the motivation and significance of your own research topic; explain the purpose of your own research topic; summarize the main content of the
paper
, Or sketch its general outline. - How to reasonably arrange the above contents and describe them clearly to readers in an orderly manner is not an easy task. Experience tells us that the introduction is actually the hardest part of the paper to write. This is because the author's familiarity with the relevant subject areas, whether the author's knowledge is profound or poor, and the significance and value of the research are all fully reflected in the lines of the introduction. We can arrange the content of the introduction into three or four levels. The first layer is composed of research background, significance, development status, etc., including a literature review of a certain research field; the second layer proposes unsolved problems or problems that need to be solved urgently, thereby eliciting one's own research motivation and significance; The third layer explains the specific purpose and content of your own research; the last is the end of the introduction, where you can introduce the components of the paper. First of all, you can write about the topic background, theoretical significance, and application prospects, but it should not be too much, as long as there is a short paragraph. The key part is the introduction and review of previous work, that is, what important innovative work has been done by predecessors on this topic. It is a brief research history, such as when and by whom, and which scholars at home and abroad have made what innovative contributions. The scientific or technical contribution of each author and each reference should be reviewed and affirmed. The main contribution here means that some important scholars may have published many articles and patents, but their main contributions should include several representative documents, not all documents. Some scholars don't have many papers, and whether to cite them mainly depends on the innovation of their literature. If there is no innovation, there is no need to cite.
- Even an article should be cited if the innovation is obvious in the introduction. Among all the references, the literature in recent years, especially the references in the past three years, should be highlighted, preferably accounting for more than half. If there are few literatures on a topic in recent years, it may be that the research on this topic is not necessary, colleagues are not interested, or it may be too difficult and there is no progress, or you do not have comprehensive literature. In the above comments, the focus is to introduce the academic or technological innovation contributions of predecessors or others to the readers, so as to clarify which tasks have been solved and which tasks have not been solved. The last paragraph is to introduce the main purpose, main content and methods of this work . At the same time, it makes people understand the necessity and innovation of this work. When some people write introductions, they like to unreasonably deny other people's work and focus on their lack of work. This is inappropriate, because your work is based on other people's work, and other people's work is insufficient. Normal, otherwise you don't need to do this job. Some people like to quote other people's language, paragraphs, and diagrams intact, which is inappropriate. When reading other people's literature, you should pay attention to note down the main points of other people's work, summarize, summarize, and organize other people's work, form your own opinions, and re-express them in your own language. Simple citations often raise issues of plagiarism or incorrect citations. When some people write the introduction, they only write about how fast certain work develops, how widely it is applied, and what problems there are, and cite multiple documents in one sentence. This is generally not advocated, because readers can't figure out each article in it. What was written, and what innovations or contributions did each author make.
- What is a literature review? The literature review is mainly after the author determines the topic of the thesis, combined with the opinions and research of the predecessors, conducts further research and analysis, so as to put forward his own opinions and research ideas on the topic. It is very important and is the focus of the opening report. In terms of content, the literature review mainly includes these contents: some existing viewpoints; the research background and current situation at home and abroad; the basic summary of previous research, the content to be studied, and the development direction. There are certain requirements for the number of words in the literature review, and the number of words is generally around 1000 words.
- How to write a literature review? Literature review generally has these three structures: the first structure: research background and significance-current situation-review-references; the second structure: research status-mainstream views-controversial points and contradictions-references; the third Structure: background and evolution - mainstream views - research methods - contradictions and deficiencies in current research - references. To find a literature review, you must meet these two conditions: one is new, and the other is representative. The literature review does not need too many, generally about 10 articles are enough. However, every literature review requires time to study one carefully. When writing the research background and significance of the literature review, the number of words does not need to be too many, generally about 200-300 words are enough. When writing, focus on one point and cover these contents: background at home and abroad, why you chose this topic, what is its significance, and so on.
- When writing a review of the literature review, the number of words should be between 200-300 words. The review should not only point out the bright spots and shortcomings, but also further describe the ideas for the next research. When writing references for literature review, you should pay attention to the format. Many friends make mistakes in the format when writing references. So pay special attention to format issues. When writing a literature review, it is a very irresponsible behavior not to create something out of nothing and let your imagination make up. In addition, writing a literature review must be objective, fair, and impartial. Don’t deliberately discredit and magnify the shortcomings of previous research in order to magnify the highlights of your own papers. You must know that this behavior is very shameful!
- Some suggestions for writing a literature review: Don’t just summarize without commenting. If you only summarize the previous research in the literature review, and your own comments are just passing by, then you will fall into the embarrassing situation of reviewing purely for the sake of review. A good review should be able to put forward the author's own point of view on the basis of extensive reading and clear explanation of previous literature, that is, it should be able to make comments or conclusions. We think that the ratio of "statement" and "comment" in the review should be 7:3. Moreover, there must be an original point of view when reviewing, and it is not simply a descriptive statistical review of the existing literature. A new proposition or model should be proposed on the basis of the review. This kind of literature review paper that puts forward a new proposition or conceptual theoretical model has high value and can point out the direction and focus for follow-up research. For example, after a more comprehensive review of the existing literature on the relationship between supervisors, employees and customers in the service industry, Anderson not only put forward several propositions, but also proposed a conceptual theoretical model. When Zahra and George reviewed the literature on absorptive capacity, most of the article was devoted to proposing propositions and constructing models.
- The scope of the literature review should be narrowed and the key points should be highlighted: Narrowing the scope means not doing "big topics" but "small topics" during the review. For example, in the field of research on service quality and customer satisfaction, it is not appropriate to choose "Summary of Service Quality Research" and "Summary of Customer Satisfaction Research" as the title of the article, but should choose "Summary of Research on Service Quality in Developing Countries", "Based on Different Cultural A survey of customer satisfaction research in the background” and other topics with relatively small coverage, so it is easy to obtain valuable results. Focusing means that when doing a review, you only need to review the literature directly related to the topic. We should read the literature carefully, and we must read the existing relevant literature one by one honestly. For each cited literature, at least the content abstract and conclusion should be read. Most of the literature should read the abstract, introduction, findings and conclusion. A few important literature must Read the full text. Do not cite unread literature in your own review. A common problem in many literature reviews nowadays is that the author has not really read the original text quoted, and directly regards other people's comments as his own point of view; Finally, while carefully reading the literature, we should also pay attention to find the latest published articles or results, to ensure that the review includes the latest literature and can reflect the most cutting-edge results of related research topics.
- Algorithm (pseudo-code) writing: q ⇒ Queue (queue), s ⇒ Set (collection); pq ⇒ PriorityQueue (priority queue); d ⇒ distance; initialization: variable, definition of data structure, and assignment of initial value; for example The famous Dijkstra (single-source) shortest path algorithm: array: (d[⋅]): d[s]=0, {d[v]=∞|v∈V−{s}}; data structure, S←ϕ , Q←V (priority queue, will contain all vertices in the start state)
- Experimental part? Engels once said: "It can be said that truly systematic experimental science became possible for the first time." (Selected Works of Marx and Engels, Vol. 3, People's Publishing House, 1972 edition, pp. 523-524). As the name suggests, compared with other parts of SCI papers, the experimental part of a paper mainly revolves around experiments, investigations, or analysis of results. The so-called experiment, of course, refers to the scientific experiment of observing and studying natural phenomena and their regularity under the condition of artificial control by using certain instruments, equipment and other material means according to a certain purpose. The experiment part is the core of supporting the research conclusions in the whole paper, but at the same time it is not easy to write. The author makes the experimental part of the SCI paper more readable, avoids the dry and monotonous language of the experimental part, and can more effectively support the conclusions in the paper, thus making it easier for the paper to pass the review of editors and reviewers. The experimental part is very important. It is the basic method of collecting scientific facts and obtaining perceptual materials. It is also the practical basis for testing scientific hypotheses and forming scientific theories.
- The general function of the experimental description is: to make readers believe that the experimental results are reliable and convincing; to enable readers to benefit from the introduction of experimental materials, instruments, methods, etc., to provide inspiration and inspiration to other professionals, and to facilitate readers to further of thinking and research; the reader is allowed to repeat the reality/replicate experiment, the experimental process and method should be reproducible, and the experiment described should be replicated by others under the same conditions. The description of the experiment usually includes the experimental process carried out, the technology involved, the materials or data used, the equipment introduced, and the conditions provided, etc. Therefore, the experimental process must be clearly described in the experimental part of the paper. In an English scientific paper, the author/authors must show readers the experiments and experimental results he/they did, as well as the specific details or important key points that he/they encountered during the experiments. Precisely, he/they must tell the reader about the relevant experimental methods, equipment and materials, his experimental procedures, and the specific research results obtained through the experiment. When researchers/authors introduce their experimental process to readers, the description should be as clear and brief as possible, without missing what the reader should really need to know. Of course, in any case, it is not necessary for the authors/authors to provide information on all points of the experimental procedure. In order to briefly and purposefully state his/their information about any experimental procedure, care must be taken to write clearly that a procedure consists of several steps and that these steps occur in a certain time sequence.
- When introducing the experiment, it should include the necessary description and comments on the whole experiment, including the theory on which the experiment is based, design ideas and evaluation methods of the experiment results. Because the description of the experimental procedure is to clearly present the study, what experimental methods or techniques have been used, or what preparations have been made before the experiment started, the interpretation of the experimental/survey data, etc. can be described. In short, as long as it is necessary to describe the whole test, it can be written in the beginning of the test. In the result part of the final experiment, our papers on science and technology are generally presented in tabular form.
- Of course, if necessary, the results of the experiment will be described in the Results section, and the effects of the experiment will be discussed in the Discussion section. If the author thinks it is not necessary to write Results and Discussion in separate paragraphs, they can also be combined into this section, or they can be written together as Results and Discussion. However, I personally suggest that SCI papers do not need to do this kind of simplification. It is better to write more details honestly, at least there should be a separate Results and Discussion part. In short, although the experiments have been done, it is not easy to write the experiments and write them clearly, and it still takes some effort (such as reading more literature, writing more papers, etc.). The foothold of the experimental part of SCI papers is to allow readers to understand and repeat the realization. To paraphrase Yu Qiuyu's words: "It is also an attitude to life without considering readers at all, and sometimes it is an attitude to life that is worth looking up to. Hold on People with this attitude can do many things, but they are not suitable for writing articles, especially SCI articles.”
- As a part of the writing content, the conclusion is indispensable. The conclusion of the paper is a summary based on the preface, background and arguments in the paper. It can also predict future trends or look forward to the future based on the analysis of the current situation, existing countermeasures, and development trends in the paper. The conclusion is the conclusion and epilogue, which is a creative, instructive, and empirical description of the results, and must be based on a certain theoretical analysis and experimental verification. Do a good job in the conclusion part, the purpose is to facilitate readers to read and provide evidence for secondary literature authors. The conclusion can be a restatement of the central idea, a summary of the research results or main points of view, or some enlightening explanation or consideration, and a prediction based on the research results, etc.
- Requirements for the text format of the conclusion: the conclusion of the thesis should be written separately as the last chapter of the main text, without adding a chapter number; the number of words in the conclusion should be around 600-800; only words should be used in the conclusion, except for some mathematics or chemistry papers in the conclusion In addition to having to use data formulas or chemical reaction formulas, icons and formulas are generally not used.
- The conclusion should generally describe the following contents: First, what problems are explained by the results of this research, what regular things are drawn, and what theoretical or practical problems are solved; the conclusion must clearly indicate the viewpoints of this paper, and what theoretical background does it have? Support, what is the guiding significance for practice, etc., if it is explained by numbers, the effect is too good and the persuasion is the strongest. Don't be ambiguous or vague. Avoid making people feel specious and doubt the true value of the paper. The second is the deficiencies or remaining problems of this study. Forward-looking statement of unresolved issues Follow-up research (or research proposals) are areas that students have thought about while completing this dissertation, but have not yet thoroughly explored. Research suggestions are often closely related to inadequacies, including inadequacies in research assumptions, data collection, and research methods, which can point out the direction for further research in this field. Third, follow-up research (or research proposals) is an area that students have thought about in the process of completing the dissertation, but have not yet thoroughly studied. Research suggestions are often closely related to shortcomings, which can provide ideas and references for latecomers to continue in-depth research.
- With the process of journals being networked, the sales of paper editions are getting less and less, and the function of searching articles by table of contents page number is becoming less and less important, but proper use of pictures in the table of contents will make complex and abstract scientific content easier to understand and have The power of dissemination has become an excellent way for editors and reviewers to quickly understand the basic content of the paper. For example, in a text interview with Mr. Yu Shuhong from the University of Science and Technology of China in the German Journal of Applied Chemistry, Mr. Yu mentioned: What I look for first in a publication is ... its graphical abstract. Therefore, although the graphic abstract will not be displayed in the formally published paper, the author should still attach the graphic abstract to the end of the paper according to the submission template requirements. A graphic abstract is a single, concise, graphical and visual summary of the main findings of the article. This can be the concluding character of the article, or a character specially designed for this purpose, which captures the content of the article for the reader at a glance. See the example below. Graphical summaries will appear in online search result listings, online content listings, and online articles, but will not (yet) appear in article PDF files or in print. The writing requirements of the graphic abstract include at least two important information. One is that you can only write 50-60 words, which is 1/3 or even shorter than the general abstract, so it must be a high-level summary of the content and innovation points of the article. . The second is to refer to the chosen figure, that is, it needs to echo with the picture to help readers understand the content of the article.
Summarize
After taking teacher Min Fan's class, I really gained a lot. What impressed me the most was the way of learning between English grammar and Chinese. I used Google Translate to check my English sentence structure and grammar. Another time, Teacher Min Fan picked me up and asked me how to buy pork and how to buy chicken. In fact, at that time, Teacher Min Fan wanted to let us understand that writing a thesis should clarify where the focus is, and only with focus can we better attract eyeball. Happy ending!