Evan Patterson

@ejpatters

PhD Candidate in Statistics at Stanford University

Vrijeme pridruživanja: rujan 2016.

Tweetovi

Blokirali ste korisnika/cu @ejpatters

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @ejpatters

  1. 1. velj

    The slides raise the question of the reproducibility of mathematical proofs. Everybody talks about scientific reproducibility, but hardly anybody talks mathematical reproducibility. For the proofs of hard theorems, this is a real issue.

    Prikaži ovu nit
    Poništi
  2. 1. velj

    Interesting MO question: among hard mathematical results, which can we trust and how can we know? Question inspired by a talk by Kevin Buzzard (), with slides online:

    Prikaži ovu nit
    Poništi
  3. 30. sij

    FYI, the old blog post by Fernando mentioned above is here: Back then, Jupyter was still called the IPython notebook but the idea of a generic messaging protocol for kernels and frontends was already established and implemented.

    Prikaži ovu nit
    Poništi
  4. 30. sij

    The same lesson can be drawn from some other highly successful software projects. I don't believe that Emacs or Vim would be still be in widespread use today if they were not so extensible. More recently, Atom and VS Code are benefiting from the same principle.

    Prikaži ovu nit
    Poništi
  5. 30. sij

    Today, there are Jupyter kernels for dozens, maybe hundreds, of languages. That is why everybody uses the Jupyter notebook, despite the fact that Mathematica and Maple introduced notebook interfaces 30 years ago. By being closed off, these systems missed a big opportunity.

    Prikaži ovu nit
    Poništi
  6. 30. sij

    Eight years after this blog post, the biggest lesson I take away from the success of Jupyter is the importance of open, transparent, extensible platforms. Fernando, Brian, Min and others built a platform that anyone could easily extend for their own purposes.

    Prikaži ovu nit
    Poništi
  7. proslijedio/la je Tweet
    28. sij
    Odgovor korisnicima i sljedećem broju korisnika:

    In SemanticModels.jl, we use Catlab WD interface, support decorated cospans, and use Diffeq.jl to solve the ODEs. Definitely still a prototype, but it’s getting there.

    Poništi
  8. 17. sij

    I really enjoyed talking about Catlab.jl (for the first time ever!) at the MIT Applied Categories Seminar. Slides are now online: Video is below (big thanks to ):

    Poništi
  9. proslijedio/la je Tweet
    15. sij

    For anyone doing computational* work involving measure theory, what capabilities would you look for in a "measures" software library? * Turing machines are cool, but here I mean "computed using an actual physical computer"

    Prikaži ovu nit
    Poništi
  10. 6. sij

    With the ability to convert wiring diagrams into expressions, introduced in v0.4, Catlab now has two very different algorithms to lay out wiring diagrams: 1. Rank-based layout, aka layered drawing, using 2. Series-parallel decomposition, using the above techniques

    Prikaži ovu nit
    Poništi
  11. 6. sij

    Support for Compose.jl is completely new. Catlab previously had TikZ support, but this code has been generalized and rewritten, with the result that the generated TikZ code is much nicer and the diagrams better looking in some cases. Improvements in this area will continue.

    Prikaži ovu nit
    Poništi
  12. 6. sij

    New in Catlab v0.5: a generic system for laying out wiring diagrams using their representations as symbolic expressions, with graphics backends for Compose.jl (declarative vector graphics) and TikZ (LaTeX).

    Prikaži ovu nit
    Poništi
  13. 17. pro 2019.

    Delighted that will become open access. , MathSciNet should follow suit!

    Poništi
  14. 8. pro 2019.

    The 2020 Adjoint School in applied category theory is shaping up with a nice list of topics. I recently compiled all the readings and blog posts associated with the 2018 and 2019 schools: Now if only I could find time to read it all!

    Poništi
  15. 5. pro 2019.

    This 2020 course on nonlinear algebra by Bernd Sturmfels is very enticing: For those of us who cannot it make into the lectures at UC Berkeley, a draft textbook (by Michalek and Sturmfels) is available:

    Poništi
  16. proslijedio/la je Tweet
    3. pro 2019.
    Poništi
  17. 2. pro 2019.

    New on the wiki: my favorite examples of monoid objects in monoidal categories. What good ones am I missing? I actually find this bit of abstract nonsense very helpful for remembering the definitions of rings, algebras, monads, and so on.

    Poništi
  18. 2. pro 2019.

    What I love best about using a Mac in late 2019 is clicking "remind me tomorrow" every single day to indefinitely postpone an OS upgrade that will permanently, and by design, hose all my old programs.

    Poništi
  19. 25. stu 2019.

    It goes without saying that functions should therefore be composed from left to right, not from right to left.

    Prikaži ovu nit
    Poništi
  20. 25. stu 2019.

    Functions should be applied on the right, not on the left. When taking introductory group theory as an undergrad, I could not fathom why Michael Aschbacher would subject us to this convention, but now I understand and wish I could get away with it.

    Prikaži ovu nit
    Poništi

Čini se da učitavanje traje već neko vrijeme.

Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.

    Možda bi vam se svidjelo i ovo:

    ·