Recently I am working on projects that deal with datasets directly. What I found is that there is so little research on data such as data annotation, though there is a lot of work on semi-supervised learning, and more recently, self-supervised learning.

In my opinion, the research community will gradually move from a model-centered research to data-centered research. I wrote an article to discuss this on Towards Data Science. https://towardsdatascience.com/data-the-predicament-and-opportunity-in-the-deep-learning-era-256f4b4fef I’d really like to hear you guys’ opinions~