Academic research is being diluted by low-effort, repetitive publications.

There's a huge, huge number of papers that do this but with LLMs. 'we prompted ChatGPT and here's what it said' is an entire genre of paper, and it's almost always low-effort trash.

My old PhD team had a professor who would essentially freeze/assume the weights of parts of neural networks, and then report faster training with better results with those weights frozen, he is still publishing and is getting 20-30 papers out yearly together with his students, department loves him because he increases the state funding single-handedly by relatively big amount. Short answer is that the incentives for research are wrong.

Not misconduct, no. There’s nothing inherently wrong with it, assuming he’s not salami slicing, which is the most obvious form of dishonesty that might be applicable. Of course, it’s probably not that useful. I would imagine this is reflected in the quality of journals most of the papers are published in. Publication count on its own doesn’t mean a great deal.

Are they lying about what they have done? If not, why would it be research misconduct? There are thousands and thousands of PHD students, not everyone will generate great papers. If you see a paper is garbage just delete it and move on.

I once rushed to do a course project the night before it’s due. I opened Kaggle notebook, got a Kaggle dataset related to blockchain frauds, spent 1-2hrs to implement simple fraud detection using out of the box tools from sklearn and xgboost. I also found a paper with pretty much the same result, but it has 15 pages and 4 authors, together with a few dozens citations. They add a bunch of other pre-processing steps and have the same result as me rushing a course project in 2hrs. That’s the quality of many research papers nowadays.

If it's cited and published it seems to be valuable research. Not everything needs to be novel. Sometimes having a reliable benchmark for yolo is what other people need.

I think this happens when colleges focus more on quantity than quality. I can think of so many colleges that actually do this. This is not misconduct, but rather just a flaw in the system and how ppl are using the flaw to their advantage and pushing out stuff like this.

Sharing a late-stage professor's perspective: There are lots of different kinds of people with the title "professor," and just because one person does this, does not mean that you should do this if you want to become a respected researcher. At the upper stages of academia, we are used to seeing all sorts of "games" people play to juice metrics, like salami-slicing papers, writing non-replicable results, overclaiming, staking territory with shallow studies, etc. Sometimes it works and can convince deans and university administrators that you are important and valuable. But when you get to the …

It's probably fine Someone has to do that kind of research. It's useful to record historical benchmarks of these things. Research isn't necessarily meant to be hard. It can be easy as long as it's useful. Maybe that prof. found a way to make easy contributions which fill a necessary niche. Those publications probably also have a low impact factor.