so, after we are all done with writing, what are some recent nlp papers that you think are worth reading?— (((ل()(ل() 'yoav)))) (@yoavgo) December 11, 2019
The above twitter thread made me think.
Not about what we read but about why we read. I personally find it extremely hard to recommend papers to others because the papers I most cherish are not necessarily relevant to the world at large, and they might not even be what your standard reviewer would consider ‘a good paper’.
I cherish papers that connect the things I’m thinking about. They’re puzzle pieces that fit my personal puzzle and may well be totally useless to someone else’s puzzle. And sometimes it’s not about the whole paper. The missing puzle piece may have been in a single paragraph or a footnote.
Many people out there think current NLP is making groundbreaking advances. Many people think the field is utterly boring. I don’t think it is either. A ‘field’ is a collection of puzzle pieces that may or may not fit the big picture that each of us is trying to put together, for the questions they’re trying to answer.
Also, the pieces that would help your puzzle may not even be in your field. I often read things from other fields which I don’t fully understand but still twist my thoughts in a helpful way.
We often hear complaints that many NLP papers just look the same. This may just come from our reading lists looking the same – there is always so much pressure to read paper-X-that-everybody-has-read. But if paper X is not doing anything for your research question, you might want to read something else first.
How should we read? How about: more selfishly? More in line with what it is that we’re desperate to find out? If we let our reference sections bring together pieces of literature that truly reflect individual questions, we might all collectively be richer.