Hacker News Reader

The papers published at top conferences are not the papers that is coming from this "industry" as the paper calls it.

These fraudulent papers are identified like this:

> For instance, of the 79 papers that one editor had handled at PLOS ONE, 49 have been retracted.

That's not what's happen at top conferences.

It is well-known that top-conferences had and still have many problems. Some examples: There used to be the problem with authors adding new co-authors after acceptance, aka "selling seats". There is a debate about how many papers one should be allowed to submit, as some people with money and influence are heavily franchising. It is unclear to what extent there is implicit and explicit reviewer collusion. Even double-blind reviews don't really solve the problem.

If we don't admit that there are fundamental problems that affect all of us (academics) and instead pretend it is only the lesser people who f things up, we'll all be screwed sooner or later.

> There used to be the problem with authors adding new co-authors after acceptance, aka "selling seats".

Top AI conferences allow that? That's insane.

They do not allow this any more. So _that_ problem has actually been addressed.

Reply on news.ycombinator.com