11011110

Posted by David Eppstein

https://11011110.github.io/blog/2025/05/31/linkage.html

https://11011110.github.io/blog/2025/05/31/linkage

• Anti-inductive dice (\(\mathbb{M}\)). One player rolling \(n\) copies of one die is more likely to win than the other player rolling \(n\) copies of the other die, except when \(n=4\).

• Berlin Universities Publishing (\(\mathbb{M}\)), a new diamond open-access publishing venture of three German universities.

• Cut / fold templates for paper mechanisms (\(\mathbb{M}\)).

• Elsevier journal articles now feature seriously-erroneous AI-generated summaries (\(\mathbb{M}\), archived) that take weeks to be corrected, if ever.

• Korean sports were climbers blocked from an international competition in Salt Lake City (\(\mathbb{M}\)) by revoking their long-preapproved visa waivers days before the competition. After a fuss on social media the top star was able to get an emergency visa; I didn’t hear about the rest. Beyond casting doubt on the US-hosted soccer World Cup next year and Olympics in 2028 (noting that climbing is one of the Olympic sports), this can be added to all the existing reasons why now is not a good time to host other international events in the US, such as academic conferences.

• I recently found and corrected a three-year-old mistake in Wikipedia’s coverage of polyhedra (\(\mathbb{M}\)): since 2022, several articles including the ones on tetrahedra, convex polyhedra, and orthoschemes (the tetrahedra in which three right-angled edges form a path) have included the claim that every convex polyhedron is scissors-congruent to an orthoscheme (meaning it can be cut into polyhedral pieces that can be reassembled to an orthoscheme).

  It is not true. Scissors congruence is possible only for pairs of polyhedra with equal Dehn invariants. The Dehn invariant, as a tensor over \(\mathbb{Q}\), has a rank (like the rank of a matrix, but in an infinite-dimensional space where we haven’t specified a basis). The rank is at most equal to the number of edges of a polyhedron, because the Dehn invariant can be written as a sum of rank-one terms, one for each edge. Thus, any orthoscheme has rank at most three (the right-angled edges don’t count). But an arbitrary convex polyhedron can have a Dehn invariant with arbitrarily large rank (for instance cut off the corners of a regular prism by general-position planes), and when the rank is four or more it cannot be scissors-congruent to a single orthoscheme. For more on the tensor rank of Dehn invariants see my paper “Orthogonal dissection into few rectangles” and my posts here “Dissection into rectangles and tensor rank” and “Dehn rank revisited”.

• AI-generated origami art instructions (\(\mathbb{M}\)), complete with non-sequential step numbering and repetitive steps that just don’t work.

• Interlace: A journal of mathematics and fiber arts (\(\mathbb{M}\)). New, diamond open-access, edited by sarah-marie belcastro and Carolyn Yackel.

• “New study reveals that when used to summarize scientific research, generative AI is nearly five times LESS accurate than humans. Many haven’t realized, but Gen AI’s accuracy problem is worse than initially thought.” (\(\mathbb{M}\)).

• Papercraft 3d Penrose tilings by Debora Coombs Criddle and Duane Bailey, exhibited at Williams College in 2018 (\(\mathbb{M}\)). The 3d rhombic version of the tiling used in these pieces allows the rhombs to be all the same shape; they form a polyhedral surface that projects onto a plane to form the more familiar 2d tiling. The artists have used color to highlight different aspects of the tiling.

• Johanna Grönqvist draws icosahedra, dodecahedra, and stellated dodecahedra on isometry graph paper.

• Nature on ‘journal snatchers’, companies that buy previously-reputable academic journals and make them predatory (\(\mathbb{M}\)). The study on which this news article is based is “Invasion of the journal snatchers: How indexed journals are falling into questionable hands”, by Alberto Martín-Martín and Emilio Delgado López-Cózar.

• Tony Vladusich wonders: how do we perceive objects as having solid body colors when the actual pixel values can vary significantly? With a nice example from a photo of colored clothespins.

• zbMATH now indexes papers by their individual open-access status (\(\mathbb{M}\)), regardless of whether the journal they appear in is fully open-access or hybrid.

• Tracy Kimbrel receives the 2025 ACM SIGACT Distinguished Service Award (\(\mathbb{M}\)), for his long work at the National Science Foundation.

https://11011110.github.io/blog/2025/05/31/linkage.html

https://11011110.github.io/blog/2025/05/31/linkage