egg|nomz|egg changed the topic of #kspacademia to: https://gist.github.com/pdn4kd/164b9b85435d87afbec0c3a7e69d3e6d | Dogs are cats. Spiders are cat interferometers. | Космизм сегодня! | Document well, for tomorrow you may get mauled by a ネコバス. | <UmbralRaptor> egg|nomz|egg: generally if your eyes are dewing over, that's not the weather. | <ferram4> I shall beat my problems to death with an engineer.
<bofh>
Egyptian is apparently a mess.
<bofh>
But yeah, someone should standardize variation sequences for Sumero-Akkadian variational forms.
<bofh>
So this becomes doable in a nice consistent manner.
<egg|zzz|egg>
bofh: seems it's just a question of a font for the script, like CJK? variations are more of an intra-script thing for place names there
<egg|zzz|egg>
s/script/language/g
<Qboid>
egg|zzz|egg meant to say: bofh: seems it's just a question of a font for the language, like CJK? variations are more of an intra-language thing for place names there
<egg|zzz|egg>
(with plans to encode that as VS, but still plans)
* egg|zzz|egg
ought to zzz
<egg|zzz|egg>
bofh: well egyptian really needs ~format chars, proposals are being refined
<egg|zzz|egg>
their text is not very plain
<egg|zzz|egg>
whereas cuneiform is
<egg|zzz|egg>
rather more than cjk probably
<bofh>
yeah, cuneiform & cjk are very similar in processing reqs, arguably
<kmath>
<volatile_void> Remember twitter 8 years ago when people would advertise “follow back” in their bios? ⏎ ⏎ Same, but for otters. ⏎ ⏎ If yo… https://t.co/PlLEJwN7HH
<egg|zzz|egg>
!wpn bofh
* Qboid
gives bofh a bow taikonaut
<egg|zzz|egg>
!wpn whitequark
* Qboid
gives whitequark a dessicated bootstrap-like condenser
egg|phone|egg has joined #kspacademia
egg|mobile|egg has joined #kspacademia
egg|phone|egg has quit [Ping timeout: 186 seconds]
egg|phone|egg has joined #kspacademia
egg|cell|egg has quit [Ping timeout: 383 seconds]
egg|mobile|egg has quit [Ping timeout: 198 seconds]
egg|cell|egg has joined #kspacademia
egg|phone|egg has quit [Read error: Connection reset by peer]
egg|phone|egg has joined #kspacademia
egg|mobile|egg has joined #kspacademia
egg|phone|egg has quit [Read error: Connection reset by peer]
egg|phone|egg has joined #kspacademia
egg|cell|egg has quit [Ping timeout: 186 seconds]
egg|mobile|egg has quit [Ping timeout: 182 seconds]
egg|cell|egg has joined #kspacademia
egg|phone|egg has quit [Ping timeout: 198 seconds]
egg|phone|egg has joined #kspacademia
egg|cell|egg has quit [Read error: -0x1: UNKNOWN ERROR CODE (0001)]
egg|cell|egg has joined #kspacademia
egg|phone|egg has quit [Ping timeout: 186 seconds]
egg|cell|egg has quit [Read error: Connection reset by peer]
egg|phone|egg has joined #kspacademia
egg|cell|egg has joined #kspacademia
egg|phone|egg has quit [Ping timeout: 198 seconds]
egg|phone|egg has joined #kspacademia
egg|cell|egg has quit [Ping timeout: 383 seconds]
egg|phone|egg has quit [Ping timeout: 186 seconds]
awang has joined #kspacademia
UmbralRaptop has quit [Ping timeout: 186 seconds]
UmbralRaptop has joined #kspacademia
APlayer has joined #kspacademia
<APlayer>
SnoopJeDi: Remember that best fit problem we were talking about?
<kmath>
<PlavchanPeter> This is the best; for the current generation of students, understanding the bit encoding of ascii characters and nu… https://t.co/CIXVcDrCgQ
egg|phone|egg has joined #kspacademia
<egg|phone|egg>
UmbralRaptor: went to Greenwich today
<UmbralRaptop>
egg|zzz|egg: Any longitude shenanigans?
egg|cell|egg has joined #kspacademia
egg|phone|egg has quit [Read error: Connection reset by peer]
egg|cell|egg has quit [Read error: Connection reset by peer]
<SnoopJeDi>
APlayer, sure
<SnoopJeDi>
bofh, Knuth's talk about his Fantasia Apocalyptica is one of the best talks I've been to in my life
<SnoopJeDi>
he more or less just walked through how he constructed the work, but it was a fascinating artistic process to see inside of
<SnoopJeDi>
particularly cute: after doing a concordance of the Book of Revelations, "blasphemy" was one of the 167 terms that became the raw material for assigning motifs/etc., and since he was assigning different compositional styles to some words, he chose to assign the style of Schoenberg's 12-tone rows ("I chose the most perfect one") to it
<SnoopJeDi>
Michael Duff's talk was also excellent but...BSM theories, feh!
tawny has quit [Ping timeout: 186 seconds]
<APlayer>
SnoopJeDi: So, I should set the partial derivatives = 0 and solve this as a linear system
<SnoopJeDi>
right
<APlayer>
Well, this one part is easier said that done
<SnoopJeDi>
pretty concise summary of optimization
<APlayer>
Because I have large sums (can't do this by hand, or at least not without filling up a whole notebook with it) and those are not linear equations
<APlayer>
I can't just use the Gauss method or something on them
<SnoopJeDi>
when you say "those are not linear equations" what are you referring to?
<APlayer>
To the derivatives of the equations I mentioned
<SnoopJeDi>
IIRC the sum of squared errors in your case was linear in all your coefficients, except for one that was inverted (just define a new symbol)
<egg|zzz|egg>
!wpn UmbralRaptop
* Qboid
gives UmbralRaptop a skew Колмогоров cabinet
<APlayer>
The coefficients are in linear functions which are squared, summed and rooted, and this whole thing is then put in a ratio that looks like (term - 1) / term and then it is multiplied by another polynomial ratio
<APlayer>
I guess if you maximally simplify it, you could get it down to some ultra ugly polynomial ratio, if it is not too big to fully fit on a single page
<SnoopJeDi>
why are you taking the root? you can minimize Σe_i^2 directly
<APlayer>
Because I am subtracting 1 from it
<SnoopJeDi>
hm, you've lost me
<APlayer>
The non-derivative term looks like sum from i = 0 to 8 (sqrt(((x_i - ax) / bx)^2 + ((y_i - ay) / by)^2 + ((z_i - az) / bz)^2) - 1)
<bofh>
SnoopJeDi: huh, that's neat. I never knew Knuth did that, even. wow.
<SnoopJeDi>
yea it's incredible and very much your shit
UmbralRaptop is now known as MagnetoRaptor
<egg|zzz|egg>
!u 𒂍𒀀𒈾𒍢𒅕
<Qboid>
U+1208D CUNEIFORM SIGN E2 (𒂍)
<Qboid>
U+12000 CUNEIFORM SIGN A (𒀀)
<Qboid>
U+1223E CUNEIFORM SIGN NA (𒈾)
<Qboid>
U+12362 CUNEIFORM SIGN ZE2 (𒍢)
<Qboid>
U+12155 CUNEIFORM SIGN IR (𒅕)
<APlayer>
That is, vector magnitude of the individual axes where each axis is adjusted with two parameters, then you subtract 1 from the vector magnitude (because it should be as close as possible to 1, actually), and then you have one error term. I want to minimize the sum of 9 squares of such terms
<SnoopJeDi>
bofh, he used a few bars from the Delphic hymn to Apollo to represent a set of church bells for each of the 7 churches mentioned in Revelations, with a slightly different harmonization for each such church
<SnoopJeDi>
APlayer, the -1 is completely incidental as a constant: if you minimize the other piece, you've minimized that expression as well
<APlayer>
The other piece?
<SnoopJeDi>
i.e. d/d(anything) of a constant is zero, it only translates the *value* of the minimum
<SnoopJeDi>
the one with your actual coefficients in it
<SnoopJeDi>
I'm saying the -1 is a distraction
<APlayer>
It's not just a -1
<APlayer>
Its an (x-1)^2
<SnoopJeDi>
oh, nvm, that's a completely different set of equations
<egg|zzz|egg>
bofh: hmm even with both the transcription and the photo (for the complaint from nanni) I'm having trouble identifying which style of cuneiform this is
<SnoopJeDi>
uhh, or is it, hang on, let me think about this.
<SnoopJeDi>
APlayer, the sum of the squared errors is 2nd-order in all your coefficients, right?
<APlayer>
I guess I could minimize x^2- 2x instead, where x^2 simplifies because of the root
<SnoopJeDi>
if I understood what you've written there, you have no terms higher in polynomial order than a^2
<APlayer>
Ah
<SnoopJeDi>
which means its derivative is no higher than 2-1, i.e. linear
<SnoopJeDi>
so no, that's not an issue
<APlayer>
The partial derivative has quadratic and inverse cubic terms in it
<APlayer>
(Ask Wolfram Alpha)
<SnoopJeDi>
APlayer, define a new constant cx = 1/bx etc.
<SnoopJeDi>
new set of constants*
<APlayer>
Ah, wait, no
<SnoopJeDi>
it will be quadratic in those
<SnoopJeDi>
and once you've solved for those, you just invert them to get the thing you *actually* want
<APlayer>
No quadratic term, at least no quadratic term with the differentiation variable in it
<SnoopJeDi>
this is seriously so much easier to understand if you manually work out how to fit a line for a bunch of points in the plane, btw
<SnoopJeDi>
it's hard to visualize until you've toyed with it some
<APlayer>
So, if V_mag_i is sqrt(((x_i - ax) / bx)^2 + ((y_i - ay) / by)^2 + ((z_i - az) / bz)^2), then, if I differentiate with respect to an a_x (offset), I get 2(x_i - a_x)(V_mag_i - 1) / (b_x^2 * V_mag_i). This is the same for any other axis' a-term, except that the x_i, a_x and b_x would be y_i (or z_i), a_y and b_y
<APlayer>
b_x partial derivatives look rather similar, but the polynomial order of some parts of it is higher
<egg|zzz|egg>
hm 𒅕 being pointy points towards the older style?
<APlayer>
SnoopJeDi: The 1/bx constant does not help me because I am struggling with the fact that this whole term is not a linear or even quadratic equation
<APlayer>
I have, in the worst case, terms with parts that have the differentiation variable with the power of -4 (I think?) and 0 (I think?). Actually var^(-3) * (var^2, stuff)^(1/2) and (var^2, stuff)^(1/2)
<APlayer>
I can't just solve for "var", just like that
<APlayer>
1 / (var^3 * (var^2, stuff)^(1/2)), even
<egg|zzz|egg>
bofh: also I'm really bad at reading actual tablets
<bofh>
egg|zzz|egg: looks like standard late Sumerian cunieform that's been badly damaged and a bit weathered, particularly in the center, but I'm at best 75% sure on that assessment.
<kmath>
<ssontjens> @Chemjobber @ChemScrapes Let he who has never “activated” food cast the first stone.
<iximeow>
bofh: did you microwave it too much?
<bofh>
iximeow: supposed to be 25 minutes in oven, turned into I think around 2.5 hours or so in oven when the fire alarm went off
<bofh>
no, closer to 3.5 I think
<iximeow>
fully activated!
<bofh>
basically I forgot about it and hyperfocused on Subterranean Animism stage 5 Lunatic practise mode for way too long, lol.
<bofh>
wtf did Southwest really send up a plane with a cracked fanblade???
<egg|zzz|egg>
bofh: blarg not sure whether 𒍢 or 𒍣 in ea-nasir
<egg|zzz|egg>
aside from that the first line seems to be consistent with the photograph with 𒀀𒈾 𒂍𒀀𒈾{𒍢,𒍣}𒅕
<egg|zzz|egg>
in the santakku font
<bofh>
I *think* the first of those two but let me take a better look.
<MagnetoRaptor>
Random observation: my advisor apparently has trouble reading the emotional state of people on the spectrum.
<MagnetoRaptor>
>_>
egg|cell|egg has joined #kspacademia
egg|mobile|egg has joined #kspacademia
egg|phone|egg has quit [Ping timeout: 198 seconds]
egg|cell|egg has quit [Ping timeout: 186 seconds]
egg|mobile|egg has quit [Ping timeout: 198 seconds]
egg|phone|egg has joined #kspacademia
tawny has joined #kspacademia
<bofh>
egg|zzz|egg: zethar sez: "Interesting. Also, it's pretty clear that it's Ur-III era cuneiform to me (It's easier to tell from the back side IMO)"
<egg|phone|egg>
Bofh: why is lethargic not here
<egg|phone|egg>
Um auto correct
<egg|phone|egg>
Zethar
<egg|phone|egg>
!Wpn bofh
* Qboid
gives bofh an aberrant inexplicable popup that didn't used to be there and doesn't do what it used to
<bofh>
!wpn egg|phone|egg
* Qboid
gives egg|phone|egg a drunk [REDACTED]
<bofh>
I'll try to drag them in :P
egg|phone|egg has quit [Ping timeout: 186 seconds]
<APlayer>
!wpn bofh
* Qboid
gives bofh a spicy point
<iximeow>
bofh: is that what caused the engine failure?
<bofh>
iximeow: the engine failure appears to be confirmed as separation of a turbofan fanblade at high speed, compare https://www.ntsb.gov/investigations/AccidentReports/Reports/AAR9801.pdf which is an investigation of a similar incident (the mechanism in this one is still unconfirmed and it *may* have occurred for a different reason, but the actual incident is essentially the same except this one happened
<bofh>
well into cruise instead of at takeoff).
APlayer has quit [Ping timeout: 383 seconds]
<iximeow>
agh i really want to read that but i need to do other things today
awang has quit [Ping timeout: 182 seconds]
e_14159 has quit [Ping timeout: 198 seconds]
e_14159 has joined #kspacademia
<Ellied>
wow cool alright I guess some weird update really did break connection sharing on all of my computers
<SnoopJeDi>
Knuth showed two results of that analysis on the Fantasia Apocalyptica that looked quite nice (one of the views had a cell-mitosis looking thing in it), although Dr. Davis said that's not online yet but should be soon
<bofh>
NICE. I'll take a look at this later when I have a moment, thanks.
MagnetoRaptor is now known as UmbralRaptor
<egg|zzz|egg>
!wpn UmbralRaptor
* Qboid
gives UmbralRaptor a submersible catbus
<egg|zzz|egg>
!wpn котя
* Qboid
gives котя a wing
<egg|zzz|egg>
bofh: please pet the kitten than Qboid gave you
<egg|zzz|egg>
bofh: probably going to go back to the british museum tomorrow, might try making some sense of the standard inscription (it is very legible)
<kmath>
<tony873004> #TESS launches today onto the 1st of 3 phasing orbits. After a Moon flyby it will settle into a 2:1 resonance ensur… https://t.co/NXwU9WBB7A
<bofh>
egg|zzz|egg: nice! try to take a picture if that's allowed.