[Fis] What are the most important 20% of Informatics?

[Marcus Abundis]

Hey Jason – interesting question, but I am unsure how to respond.

I do not view 'your issue' in 80/20 terms, but I instead see one central
'information theory' matter that must be addressed with many
implications—where information theory is just incomplete. There have been
many attempts at addressing this issue over the years, named by Shannon and
Weaver in 1949 as a missing "real theory of meaning". The earliest attempt
was perhaps Yehoshua Bar-Hillel and Rudolf Carnap's theory of semantic
information in 1950(?) which was grossly inadequate, and then you have
Pedro's more recent offering which I believe was cobbled together from FIS
exchanges over recent years. And of course, there are others on FIS who
make their own attempts, but none that I believe really 'hold water'.

There is also my own view which was accepted for poster presentation at ETH
Zürich and ICLR 2025 (https://urldefense.com/v3/__https://iclr.cc/virtual/2025/33131__;!!D9dNQwwGXtA!R0eaIy-Gky5Q4EQAslQl8PZeSPMEBeIAJXYoOj8RdpZomsAvhciNh5hrRrhxP3yCCRyUqGN99Gr0E9x9$ ), with an
earlier draft accepted to NeurIPS 2024. If you click on the Abstract and
Project Page links from the ICLR page, you can read the reviewer comments,
the abstract, and download the paper—all of which suggests good progress,
but with more work still needed.

Marcus
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listas.unizar.es/pipermail/fis/attachments/20250622/43284b70/attachment.html>