Skip to main content
9 events
when toggle format what by license comment
May 1, 2021 at 23:57 history edited Joseph O'Rourke CC BY-SA 4.0
added 80 characters in body
May 1, 2021 at 21:38 comment added Will Sawin The third example that you have just added does actually involve neural networks, and thus seems to be a better fit for the question than your other two. It seems worth posting a new answer, or editing your answer to focus more on it.
May 1, 2021 at 21:35 comment added Will Sawin @MattF. And now Wagner's paper refutes a conjecture by AutoGraphiX (which was previously, independently, less efficiently refuted by a human mathematician, the authors point out). It's the circle of life! I would actually defend this - if we have automatic conjecture generators and other systems that can automatically refute them, then by combining them we should obtain automatic generation of highly plausible conjectures.
May 1, 2021 at 13:32 history edited Joseph O'Rourke CC BY-SA 4.0
added 364 characters in body
Apr 19, 2021 at 9:16 comment added user44143 It is disconcerting that their first praise for this automated system is the way it refutes conjectures from a different automated system.
Apr 15, 2021 at 8:55 history made wiki Post Made Community Wiki by Stefan Kohl
Apr 15, 2021 at 1:29 comment added David Roberts And not just this type of thing, here's a sample from a student in my department: paperswithcode.com/author/matthias-fresacher Learning Erdős-Rényi Random Graphs via Edge Detecting Queries
Apr 15, 2021 at 1:28 history edited David Roberts CC BY-SA 4.0
Formatting
Apr 15, 2021 at 1:05 history answered Joseph O'Rourke CC BY-SA 4.0