HKUST VisLab在疫情期间，坚持科研。同学们分散各地，有居家留守的，有在科大隔离的，在面临前所未有的挑战的情况下，齐心协力，守望相助，完成了10多篇IEEE VIS2020投稿。 投稿结束后，同学们写了不少对投稿的回顾与反思。 下面是受到广泛好评的袁林萍同学写的反思文章。 袁同学是西安交大软件学院第一名，三届国家奖学金获得者，虽然才是博士一年级，已经有过多次投稿经历了。 过去一年多时间，她在深研院曾伟以及Waterloo大学赵健指导下，研究用机器学习做可视化以及infographics的颜色推荐。 袁同学聪明伶俐，基础扎实，编程能力很强，做事积极主动，勇于挑战自己，转益多师，展示了非常好的科研素质。 林萍同学的主页：https://yuanlinping.github.io/

I have participated in VIS submission twice as the first author. Sweating over two years and suffering from much pain, I have learned much from my supervisor and my two mentors, Wei and Jian.  Below I will summarize my thoughts and improvements during VIS 2019 and VIS 2020 submissions.

Mindset

·       Have faith in myself and my projects; be optimistic and positive.

Among all the things Wei taught me, I think this one is most important and useful. During a meeting three weeks before the VIS 2019 deadline, I complained again that our work had severe limitations. But Wei said very happily “I think our work is very good! It can solve xx, xx, and xx. Can previous work solve these problems?” Suddenly, something stroked me. Jian also said similar words when I was really worried and dissatisfied with our results. The different attitudes between two senior researchers and me reminded me of the two thinking styles revealed in “Is the glass half empty or half full”. I tried to mimic them to think of the full half (optimism) instead of empty half (pessimism) and found it useful to reduce stress and self-doubt. Of course, I also needed to be critical to our work and accept others’ negative comments. The lesson I learned is that, There are drawbacks and advantages of our work. Be optimistic by thinking of the advantages and meanwhile, try to solve drawbacks or defend the paper as much as possible.

·       There are always solutions to problems.

So how to solve the drawbacks? We can always come up with several solutions after exploration and discussion. Don’t get stuck in one tough solution if the target problem is not a contribution, but also don’t give up so quickly. (How to decide when to stop and explore other solutions is another story.) However, some drawbacks may not be solved during the submissions. In these cases, there are still ways (admitting them as limitations, changing my mind) to accept them and avoid getting stuck and being bothered. For example, I thought the recommended palettes were not appealing and felt very upset. Later, I alleviated myself from the negative feelings by thinking that judgment on color palettes was quite subjective, and someone might think they were (relatively) appealing. Actually, it proved to be true after we obtained the results of the survey study and interview study.

·       Don’t be afraid of demonstrating my projects.

Jian suggested conducting an expert interview study. I was reluctant/afraid to do that at the beginning because I thought the results were not good enough, the interactions were tedious, and balabala. However, I didn’t reject the suggestion and prepared for it. Gradually, I changed my mind and overcame my pursuit of perfection. I realized that I did not have to show them a perfect system because my purpose was not to surprise them. My purpose should be to get constructive feedback and the expert interview was a great opportunity to achieve that. A prototype was enough for getting feedback from experts, and getting to know whether the tool was useful and how to improve.

Teamwork

My two mentors have totally different styles. Wei was very powerful, and it was he who was actually in charge of the project, while Jian gave alternatives and asked me to make decisions. It took time to adapt to their styles and overcome the difficulties. I learned the project management from Wei in VIS 2019 and practiced in VIS 2020 (though I almost failed). My goal for the next submission is to take more responsibility as the first author and reduce my dependence on senior coauthors.

In VIS 2020, I had two junior coauthors who mainly helped me with coding. It was my first time collaborating with junior students and found some difficulties, such as effective communication, whether I should push them, whether I should give them a specific task or just finish it by myself. I also realized that I made a detailed instruction for each task to control everything and worried too much, which made me tired and easy to get angry. After observing that Jian just believed us and let us finish the tasks in our own ways, I tried to trust my teammates and let them be creative. It turned out that they were qualified.

Writing and Figures

Many senior students told me that logic was more important than the language in academic writing. But when I started to write, I still spent much time on the language and lowered the priority of logic, until Jian asked me to list the bullet points of the whole paper, otherwise he would not help me to revise it. At that moment, I finally accepted that logic was more important. Using the bullet points, I felt like writing a paper was similar to writing code. First, we design the structure from a high-level, and then we fill in specific content, and we “refactor” (revise) the content from time to time. In addition, Wei emphasized a lot on figures and asked me to beautify the figures again and again last year. He suggested using the figures to facilitate the process of thinking and writing. I found that both bullet points and figures were tools to summarize our work from high levels. However, I felt difficult to get started from high levels because it took time to figure out a clear logic, which required further practice.