pglpm

joined 2 years ago
MODERATOR OF
[–] pglpm@lemmy.ca 12 points 2 months ago* (last edited 2 months ago) (15 children)

I don't understand why they keep saying "the Trump admin is doing this", "the Trump admin is doing that", and so on. It isn't the Trump admin: it's the majority of USA citizens that's doing this and that. They voted it. They're the first responsible and guilty. Each and every single person in that majority.

[–] pglpm@lemmy.ca 5 points 2 months ago

One big, sad problem in machine learning and AI is that many, hopefully not most, practitioners there are largely incompetent in statistics and probability. This is why they often incorrectly evaluate the performance before deployment.

[–] pglpm@lemmy.ca 12 points 2 months ago

Does it really? sad. I was reading the "invite" process and can't say I fully understand it.

Papers can always be uploaded to https://libgen.is/scimag/librarian/ Many thanks to all the anonymous users who do.

[–] pglpm@lemmy.ca 1 points 3 months ago

Yes, though I think I also read it from other sources. But I want to read more and find out further news about those claims.

[–] pglpm@lemmy.ca 1 points 3 months ago

No, I stand corrected: the summary/snippet wasn't showing some of the quoted search words, but the page had all of them. Well done SearXNG.

[–] pglpm@lemmy.ca 2 points 3 months ago

Trying it these past days and I'm impressed!

[–] pglpm@lemmy.ca 1 points 3 months ago* (last edited 3 months ago)

I didn't know about this – that may explain the problem.

[–] pglpm@lemmy.ca 1 points 3 months ago

Nice initiative besides the search service! Thanks for sharing.

[–] pglpm@lemmy.ca 1 points 3 months ago

True about the good old-Google feeling! I want to find how much about of the shadiness claims about Startpage have been substantiated or denied.

[–] pglpm@lemmy.ca 1 points 3 months ago

Really funny about page 😂 Cool that they share the git code!

1
submitted 2 years ago* (last edited 2 years ago) by pglpm@lemmy.ca to c/generalrelativity@mander.xyz
 

I was reading some works – true pearls! – by Synge: his conference contribution Tensorial integral conservation laws in general relativity (1959/1962) and his book Relativity: The General Theory (1960). In these works Synge introduces an extremely interesting definition of four-momentum and of rotational momentum, based on two-point tensors. The definition is interesting because (1) it involves the full Riemann tensor, not just the Einstein tensor, (2) it includes the (or rather, defines a) four-momentum and rotational momentum of the gravitational field, (3) it obeys a conservation law as opposed to a balance law (the equation ∇⋅T=0 expresses in general just balance, not conservation).

The definition for rotational momentum is also interesting because it appears as the natural generalization of the one in Newtonian mechanics, which is based on the affine structure of its 3D space. Roughly speaking, in Newtonian mechanics we have (r-a)∧p, where a is a fixed point, r the point of interest, and p the momentum (density) at the point r. Synge essentially replaces the difference "r-a", which relies on an affine structure, with the geodesic distance between two points R and A in spacetime, through his two-point "world function". In his book he explains that general relativity requires the appearance of a reference point (a or A) also in the definition of four-momentum, whereas such reference point is superfluous in Newtonian mechanics.

OK this was a very poor summary, just to pique your interest. For details see Synge's conference contribution, and chapter VI, especially §4, of his book (refs below).

Bryce DeWitt even commented "Je suis tout à fait de l'avis du professeur Synge qui insiste sur le fait que ces fonctions de deux points se montreront très importantes dans le futur développement de la théorie de la relativité générale" on the conference contribution. Two-point tensors were quite fashionable in the 1960s, they are used in interesting ways also in Truesdell & Toupin's The Classical Field Theories (see Part F and Appendix III there).

Yet, these definition venues seem to have been abandoned today. Here are my questions to you: why? just for unfathomable sociology-of-science reasons, or because of physical-mathematical ones? Are there works today which further explore these venues?

References:

• Synge: Tensorial integral conservation laws in general relativity, in Lichnerowicz,Tonnelat: Les théories relativistes de la gravitation (CNRS 1962), pp. 75–83. https://libgen.is/book/index.php?md5=74345AB69DDF9EE233FA55F55FDCB057

• Synge: Relativity: The General Theory (North-Holland 1960). https://libgen.is/book/index.php?md5=7AE08880CF8086FED4D3BCF732BE8E54

• Truesdell, Toupin: The Classical Field Theories, in Flügge: Handbuch der Physik: III/1 (Springer 1960), pp. I–VII, 226–902. https://doi.org/10.1007/978-3-642-45943-6_2 https://libgen.is/book/index.php?md5=728F54156B632C44EAC2C559F120DDAB

10
submitted 2 years ago* (last edited 2 years ago) by pglpm@lemmy.ca to c/bayes@mander.xyz
 

A little advertisement for a new free online course about the foundations of data science, machine learning, and – just a little – artificial intelligence. It's been designed for students in computer science and data science, who could be uncomfortable with a head-on probability-theory or statistics approach, and who might have a lighter background in maths. The main point of view of the course is how to build an artificial-intelligence agent who must draw inferences and make decisions. As a course, it's still a sort of experiment.

https://pglpm.github.io/ADA511/

In more technical terms, the course is actually about so-called "Bayesian nonparametric density inference" and Bayesian decision theory.

 

Can't help imagining Saitama putting a definite end, without so much back-and-forth, to Mahito's hateful smirk. One punch is all that's needed.

 

What are the comparative and superlative of the adjective "fun"? I'd say "more fun" and "most fun"...

But I'm somehow slightly tempted by "funnier" and "funniest", which should be for "funny" though, not "fun"...

I didn't find anything about this in the main dictionaries.

 

...and thought of randomly posting it here.

19
submitted 2 years ago* (last edited 2 years ago) by pglpm@lemmy.ca to c/sdfpubnix@lemmy.sdf.org
 

I wanted to tag SDF today. A #sdf came up, but it seems to refer to something(s) different. I also saw a #sdfdotorg.

Is there a tag that's sort of "standard" to refer to SDF? Standard in the sense that it's typically used by ~~SDF members~~ [Edit:] Mastodon users interested in SDF.

 

Personal websites often give an email address for contact, as a mailto:blah@blah.blah link. And the address is often obfuscated in a variety of ways to avoid its harvesting by spam bots.

If one wants to give one's Matrix address in a website, what's the correct way of writing it as link? is it recognized as any kind of MIME (like mailto:)?

And is Matrix-address spamming something possible and common? In this case, how should one obfuscate a Matrix address given in a website?

Lots of questions from a noob :) Thank you for your explanations!

Edit for others with the same question: as per @QuazarOmega@lemmy.world's explanation in the comments, the Matrix address can be given as the link

https://matrix.to/#/@[yourusername]:[your.server]
8
submitted 2 years ago* (last edited 2 years ago) by pglpm@lemmy.ca to c/firefox@lemmy.world
 

cross-posted from: https://lemmy.ca/post/2217942

In my desktop Firefox I use Cookie Autodelete to keep a whitelist of sites whose cookies won't be deleted. All other cookies are deleted as soon as all tabs for a particular site are closed.

Android's Firefox, from what I gather, only give you two choices: delete all cookies upon quitting (not tab closing), or save them across sessions.

Unfortunately the extension above does not work on Firefox Android, and I haven't found any other alternatives.

Do you know of any alternatives or other solutions, to get a behaviour similar to the desktop one? (And also: how come that extension is not supported on Firefox on Android?)

Cheers!

35
submitted 2 years ago* (last edited 2 years ago) by pglpm@lemmy.ca to c/firefox@lemmy.ml
 

In my desktop Firefox I use Cookie Autodelete to keep a whitelist of sites whose cookies won't be deleted. All other cookies are deleted as soon as all tabs for a particular site are closed.

Android's Firefox, from what I gather, only give you two choices: delete all cookies upon quitting (not tab closing), or save them across sessions.

Unfortunately the extension above does not work on Firefox Android, and I haven't found any other alternatives.

Do you know of any alternatives or other solutions, to get a behaviour similar to the desktop one? (And also: how come that extension is not supported on Firefox on Android?)

Cheers!

 

I wonder how many in this community resonate with this.

view more: ‹ prev next ›