22
$\begingroup$

It's been over a year since his January 2017 retraction and correction. Is there news?
If not is this normal for validation to take this long? I would expect it would get plenty of attention. Has anyone of note spoken up to support/doubt the quasi-polynomial result?

$\endgroup$
  • 6
    $\begingroup$ The correction (and hence the quasi-polynomial result) was immediately supported by Harald Andrés Helfgott. His expository paper (arxiv.org/abs/1701.04372) and its translation (arxiv.org/abs/1710.04574) are all the support that is needed for the quasi-polynomial results (as long as no further holes or mistakes are found). There is no news to wait for. $\endgroup$ – Thomas Klimpel Mar 11 '18 at 7:06
  • $\begingroup$ have these been published somewhere?(not arxiv) $\endgroup$ – Meir Maor Mar 11 '18 at 7:18
  • 5
    $\begingroup$ It says "To appear in Astérisque. Fascicule no 1125 of the Bourbaki seminar (69th year, 2016-2017)" on arXiv. It has not appeared yet, (smf4.emath.fr/Publications/Asterisque/2017) lists "Séminaire Bourbaki, volume 2015/2016 ", and (smf4.emath.fr/Publications/Asterisque) does not yet list anything from 2018. $\endgroup$ – Thomas Klimpel Mar 11 '18 at 7:43
  • 10
    $\begingroup$ It's absolutely normal for a math paper to take more than a year, even more than two years to publish. $\endgroup$ – Sasho Nikolov Mar 11 '18 at 21:01
  • 3
    $\begingroup$ On the other hand, Publication of a paper does not mean that the results are correct. There are many cases of results withdrawn after getting published in journals. For instance, sciencedirect.com/science/article/pii/… $\endgroup$ – Mohammad Al-Turkistany Mar 13 '18 at 12:39
19
$\begingroup$

Aggregating comments by Thomas Klimpel, Sasho Nikolov and Mohammad Al-Turkistany into a community answer:

The correction (and hence the quasi-polynomial result) was immediately supported by Harald Andrés Helfgott. His expository paper (https://arxiv.org/abs/1701.04372) and its translation (https://arxiv.org/abs/1710.04574) are all the support that is needed for the quasi-polynomial results. Helfgott’s exposition appeared as Exposé 1125 in Astérisque 407 (2019), Séminaire Bourbaki 2016/2017, pp. 135–182.

There are no current known issues with Babai’s proof, and it has gone through extensive peer review. That said, even published peer-reviewed papers have later been withdrawn as incorrect.

The support by Harald Andrés Helfgott along with the lack of other issues should be sufficient for us to accept that graph isomorphism is indeed solvable in quasi-polynomial time.

$\endgroup$
  • 4
    $\begingroup$ In case it's unclear, the thing which is "to appear" in the Bourbaki seminar is Helfgott's exposition, not Babai's paper. $\endgroup$ – Sasho Nikolov Mar 14 '18 at 14:22
  • $\begingroup$ Note I deliberately made this answer a community wiki, not only to not take credit, but anyone including you can easily edit to make any improvements/clarifications. $\endgroup$ – Meir Maor Mar 14 '18 at 14:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.