*The following contents are written in 2000.

Good research is not something that can be done with ideas alone.

• Literature review: find relevant studies and read papers
• Hands-on: actually program and try to produce results.
• Discussion: Discuss what you are doing and the results, and listen to others' opinions.

It is important to repeat those steps above. Conversely, if you can repeat this stage many times, you can do good research even within a month. If you cannot repeat this cycle, you will not be able to make progress in your research even if it takes you a year. Of course, the main thing in research is to think by yourself about the content of your research, but you need to do the steps in a balanced manner. If you get stuck, you can usually figure out what you need to do next by searching the literature or having a discussion.

When you are doing research, you tend to think, "Other people won't understand it anyway," and you tend to barricade yourself. However, it is important to have other people understand the content of your research, and getting opinions from different perspectives is a shortcut to advancing your research. Matsuo Gumi (Matsuo group)'s meetings are held in a frank atmosphere, so please take the initiative to make a presentation and take advantage of it!

Additionally, we often think, "If we do it this way, surely we can do it this way. But there are some problems too." and we often end up just thinking about it without implementing it. However, there are so many things that we realize only after seeing the results of the implementation. Since many intelligent researchers are working in the same field, they all know what they can find out after thinking about it for a while. In order to understand a subject well, it is necessary to try various things in order to do good research. Don't be lazy, just do it. This is important.

Also, we tend to neglect literature research, thinking, "My theme is this new topic, so there is no other research I am doing," or "The only relevant research is this one that my seniors are drawing on”. However, many researchers around the world have done a lot of research over a long period of time. If you search for it, you will find relevant studies. But it is not so easy to find them. It is only when you understand various other studies that you can understand how your research fits in with other studies. Being able to do a proper literature search is also one of the research skills.

Research involves assumptions. These assumptions can be abstractions of phenomena, fixation of objectives, and so on. We then propose various methods, but it is important to understand the valid scope and limitations of the assumptions. For some fields, certain assumptions are almost implicit, or assumptions made in the research of a senior colleague are carried over as given. It is important to always think to yourself, "Is this objective and this method really the right one?" and to do your research with understanding. Dr. Kanade's book, "Think Like an Amateur, Perform Like an Expert (素人のように考え、玄人として実行する)" is very helpful.

*The following contents are written in 2000.

First, look up the topic you want to research on Google. If you cannot find the main body of the paper on the Web, look for it in a library (we recommend the JSAP library or the Library of Information Science). It is also necessary to read textbook-like books. In addition, there are always related English literature, so after you have a general grasp of the contents, survey mainly English literature. In particular, NEC's CITESEER has a large number of papers, so you can get a lot of information from related papers. If you know a major international conference, you may search for related papers mainly on the new papers from that conference. As you collect many papers and read some of them, you will get a general idea of the issues in your research field, the flow of research, and the papers you should cite. Rather than spending a lot of time reading a single paper, it is more efficient to read through the papers with an emphasis on understanding what the paper is trying to say, and then read the ones that you think are really important, especially the ones that you think are important.

In any case, in the end, it is necessary to 1) have an overview of what kind of research is being done in your field of study whether domestic or international, 2) to know and read some of the major papers in the field, and 3) to read accurately some papers that are close to your methods.

*The following contents are written in 2000.

Deciding on a research topic is in itself a very difficult task. In a thesis, you are given a theme. For a master's degree, you are given a general field of study. For a doctoral degree, you basically have to come up with your own theme. I think you should consult with your professors and seniors while deciding. I am not going to go into this too deeply here. A good research theme is one that makes an important contribution to the flow of research in academic societies and society, but if you follow it, it is actually influenced by the policies of the national science and technology policy. There are also strategies for the direction of the laboratory as a whole. There are various kinds of research I would like to do, so let’s conduct collaborative research if you are interested in the same research theme.