Lu Jialin, let me defend arXiv a tiny bit. (I am the cs.LG moderator for arXiv.) It is not a stated policy of arXiv to encourage bad research, so by “encourage” perhaps you just mean “facilitate”. I agree that arXiv provides a way for people to rapidly disseminate poorly-executed research (just as it does for well-executed research). Hopefully, people are paying a reputation penalty for this. It would be great to see more overlay services that support public reviewing and ranking, so that readers could get a sense of what worth reading.