The cold start problem refers to the lack of accumulation of public questions and discussions for a given research paper over time. A direct cause of this fact is as follows; any reader is in the position of the first person who actually reads the paper, whether the paper is 20 years old or just has been published a week ago. The cold start problem appears in the author’s perspective as well. More specifically, authors need to reply to similar questions as if the question has never been asked before.
All of this is due to the fact that the current question & answer mechanism in academia is running between the following private parties:
- The reader itself.
- The reader and its colleagues.
- The reader and the authors of the paper in question.
This is a problem that needs to be fixed urgently, because the past literature is not being discussed as actively as the current literature. Hence, we lose potential accumulation of questions and answers for the current literature by the time a fix to the cold start problem is not being implemented.
You can find my EUROCRYPT 2021 rump session talk regarding the topic (starts at 26:00): https://youtu.be/-Gz7j_WUOW8?t=1560