0

They claim equation 5 is restatement of convexity. What am I missing? $\lambda = 0, \lambda'=1$ seems wrong no?

https://www.scihive.org/paper/1702.04877

enter image description here

mathtick
  • 147
  • 2
    When I encounter a statement like that in a paper (when it not an obvious typo.) I have the hardest time continuing. – copper.hat Jun 18 '20 at 20:02
  • 3
    Note that $(2)$ is not what we usually call "convex". So it seems those authors are all mixed up. – GEdgar Jun 18 '20 at 20:03
  • 1
    They have many errors. The convex function $1_{{0}}$ is not continuous on $X=[0,1]$. – copper.hat Jun 18 '20 at 20:04
  • Thanks, I'm usually reading this stuff on my own and never know if I'm just tired or going crazy. – mathtick Jun 18 '20 at 20:05
  • 1
    I understand. One issue with easy publishing is the amount of drivel that comes off the press. – copper.hat Jun 18 '20 at 20:06
  • I honestly wouldn't mind if it were a bit more interactive and feedback landed somewhere, but I'm always the only one posting comments on scihive or whatever pdf commenter. – mathtick Jun 18 '20 at 20:11
  • 1
    The issue is not just with easy publishing. The two authors are both 'professors', so it seems like the vetting process was not very thorough. Imagine paying tuition and getting these clowns. – LinAlg Jun 18 '20 at 20:17
  • I think this is just a posted note to arxiv, but it has ended up in arxiv. It's the same problem as all these kids writing medium articles. Poaches attention. The uncareful reader might assume paper volume = impressive. Might be useful, but the feedback for bugs/errors is too slow or costly so it doesn't improve. Sites like stackexchange etc are some of the few that really capture negative feedback and actually lead to organic learning and improvement. – mathtick Jun 19 '20 at 09:31

1 Answers1

3

Take $F(x) = x$ on the reals, $p=0,q=1, \lambda = 1, \lambda' = 1$ then the above says $F(q)=1 \le F(p) = 0$.

copper.hat
  • 172,524