Tips for reading scientific papers
Knowing why you want to read a paper is the most important
The most important thing is to know why you are reading the paper.
If you are uncertain, you should read the paper to figure out why you need to read it (or don't). You now think: what am I talking about?
There are many reasons to read a paper:
To figure out whether you should read it
To get a high-level overview of the results
To get a glimpse of the field
To get to know the details of the method
These require a different depth of reading. This is a mistake I made many times at the beginning: I wanted to read from beginning to end, even if that paper was not particularly useful for me at that point in time. The heuristic I developed is to
Stop at any point if you got what you came for.
This can be boiled down to approximately the following process:
Do you know why you want to read the paper?
If yes, go to the specific part
If not, skim for a high-level overview. What to check?
Figure 1
Contributions
Figures
Results
Discussion
+1: If unsure, you can ask NotebookLM to summarize
If the first pass makes you want to know more, you still don’t need to read the whole. You can (and probably should) skip:
The literature review (if you are familiar with the field)
The introduction (and only focus on contributions)
The methods/theory/experimental setup — depending on what you want
The second pass should give you a decent understanding of technical details, not the minutiae. If you want to know those, it's time for a third pass. This is mostly for:
Proofs
Experimental details
Code
And the appendix
Dig Deeper
Improving the clarity of scientific papers
Doing good Science requires precision. But precision must not obfuscate the message. You have seen this: a simple idea clothed into the most delicate fabrics with words no one ever uses.