Reading tips:
- Read the arxiv multiple times a week, to spot papers that 1) directly relate to your research, 2) are of general significance, such as a impactful science result, 3) appeal to your curiosity for any reason. The idea is to have both breadth and depth in terms of topics.
- Learn to skim through papers quickly, in order to assess the quality, significance, timeliness, and relevance of the results to your work.
- Discuss papers with colleagues before spending too much time on them, in order to confirm or change your initial assessment by comparing with theirs. This is a good way to improve your skills. Also, they might tell you about the history of a field or a research topic. Some papers appear brilliant to an outsider of the field, but would in fact be flawed if you knew about the topic in detail. It’s absolutely fine to make mistakes. We all do. Reading and assessment skills are honed with practice and precisely by making mistakes, for example missing an important paper relevant to your research, presenting an erroneous result at journal club, etc. This will never be held against you.
- There are only so many papers one can read per week. It’s impossible to absorb all the information. A popular strategy is again breadth and depth. There are probably only a handful papers per year that you will read extremely closely, to the point of understanding every detail and even sometimes reproduce the results of. Those are often the ones that directly relate to your current research, or to a new question you would like to get into. The former is often obvious. The latter is a matter of taste: what picked your curiosity, where you woulld like to take your research, etc. For all the others, you will only have time to skim the paper, gather some information about the main results and sometimes some interesting details.
- It is useful to keep notes about what you read, because you will obviously forget most of it. To zeroth order, this could be a list of papers you skimmed, and what your conclusions are. Direct annotation can also be good. For the papers you work through in detail, this could go in your research journal.