[D] Weird trend in machine learning: papers tackle easy problems are well cited

I don't know if anyone else have encountered this, but I have seen a lot of ML papers with extremely ideal assumptions (sometimes hardly relevant to machine learning) and then a group of reseachers, sometimes even from very well known universities, come together to "solve" this problem.

Despite this, these papers will be quite well-cited as compared to the mean, which makes me really confused.

I am not sure if the researchers are just not aware of the weakness of …

easy learning machine machine learning machinelearning trend