The h Bar

Secret

17:00

the moelcules are given by 2 coordinates, and it goes like:

0 0
1 0 1 1
2 0 2 1 2 2
3 0 3 1 3 2 3 3 etc.

danielunderwood

Fill with 0s?

Secret

no, the labelling is literally a triangle sequence

The outcome is for each molecule and for each orientation, compute the energy difference with the 0 orientation of that molecule

so you get pairs of relationships and the systematic trend of that is the focus of my PhD

The python code is the most annoying part though, often have to think about how to code the loops

enumaris

o.o

danielunderwood

Refactor all of your loops into separate functions and that part should be a lot easier (and less bug-prone)

That's also boring, but can save a lot of work in the long run

David Z

There's usually some sort of vectorized way to deal with the loops, too

Secret

17:09

Yeah, in my code, most of the loops are only used in generating the numerous plots (because there isn't a nice way to vectorise it), but otherwise, column substraction, slicing and other things can be done in terms of arrays and lists

The above triangle shaped indices cannot be generated by list comprehension as far I know from python though, because the iterator itself changes (e.g. if i=0, then j runs from 0 to n, if i=1, then j runs from 0 to n-1 and so on)

thus it has to be generated using two for loops

David Z

Well list comprehensions are more or less equivalent to for loops anyway, at least in terms of efficiency

danielunderwood

You can be a heathen and nest comprehensions lol

David Z

You can do stuff like [[f(i, j) for j in range(n-i)] for i in range(n)] but it's not really going to make that much difference

Oh @danielunderwood I see we were thinking along the same lines

The real speedup comes from vectorization vs iteration though

Secret

Yeah, I really love thinking in terms of vectors (to the point I can compute matrix stuff manually very quickly because of that), thus I always get frustrated whenever pandas or python throws some error saying that I cannot use lists

David Z

I'd see if there's some way to rearrange your calculation so that it can be done in vectorized form. Possibly even if each value winds up being computed twice, i.e. if you reflect your triangle across the diagonal to create a full matrix

I have no idea if that's possible, of course, but if it is, it could still be faster than the iteration

Secret

17:21

Well the major issue here is the symmetry of the data makes that reflection very complicated. Specifically, the orientation is a rotation of the molecule around a central axis in increments of 20 degrees, thus if the molecule has coordinates (a,b,0) where the entries means (atom group 1, atom group 2, orientation in degrees), then we have the symmetry where (a,b,180) = (b,a,0) for all a,b

Semiclassical

@Secret hmm, it looks like I was wrong. the attracting fixed points (for large times) are at x=-1.255 and x=0.86

Secret

minus 1.255 seconds???

Semiclassical

fixed

by comparison, the classical minima are at x=pm 0.943

Secret

What do the new streamline diagram look like afte fixing that when it goes all the way to t=4?

vzn

@Semiclassical sounds exotic is that Method of Moving Asymptotes?

Semiclassical

17:26

also, the repelling fixed point (what the critical trajectory is heading to) is at x=0 classically and x=-0.442 according to bohm

Secret

hmm, so bohm is saying the classical maxima is displaced slightly to the 2nd well

Semiclassical

and in the first as well

which is goofy

not necessarily wrong, of course

@vzn well, in the simplest case, it's just the ode you'd have for constant velocity motion: dx/dt = x/t

Secret

Can you show an updated streamline plot with t running from 0 to 4?

Terry Bollinger

@bolbteppa for whatever it's worth, the unit vectors (ijk only) of H and their sums provide a surprisingly succinct and informative way of classifying the fermions and their properties. Glashow indirectly noted this (using just cubes, no mention of quaternions) way back in 1980.

Semiclassical

@Secret once I actually trust it, yes. right now I'm finding reason to distrust it

Secret

17:30

ok

Semiclassical

...oh, shoot. Now I remember

Secret

It will also be interesting to see what happens if the trajectory is started at x=0 and see if the two streamline plots agree qualitatively with each other

Semiclassical

the ode i had extracted was for the small-time asymptotics only

Secret

ah, so t > 1 will be way off

Semiclassical

ya

I have the full-time expression, I just was only looking at terms of order t^2 or less

let's see if mma can do the full time expression numerically

enumaris

17:34

mixed martial arts?

Semiclassical

sure, why not

vzn

@Semiclassical are you using MMA built into mathematica? found some posts on that but nothing else. is it built into NDsolve/ Dsolve?

Semiclassical

MMA = mathematica

it's just NDSolve magic

alas, mathematica seems unwilling to touch the full thing

can't really blame it

I may still be able to do streamplot though

Secret

[Meanwhile in my PhD slicing my geometry data]: 5000 molecules, let's just hope this hotfix will get me through

Nick

In the 00s, Brazillian kikboxing became popular. In the 90s, muay thai was a thing. In the 80s, it was ninja schools, In the 70s, I guess it was kung fu. Martial arts sure are a lot like programming languages.

Terry Bollinger

17:39

Wow, why did my stream show the quaternion discussion as having just happened... and now it disappeared? Sorry for the out-of-date irrelevancy, folks.

Nick

sub atomics ftw

enumaris

what's Brazillian kickboxing?

vzn

← (lol) black belt in coding + ninja master :P

Terry Bollinger

@enumaris this is a trick question, right? I'm going to guess... kickboxing that takes place in Brazil??

enumaris

But is it somehow distinct from just "kickboxing"?

Semiclassical

17:43

ok, uh

this is what i get as a streamplot

enumaris

I thought BJJ was the big popular martial art from Brazil.

Semiclassical

vzn

← just saw the Rock in Rampage on new home HD bigscreen prj, oh geez

Semiclassical

if you're looking for me to explain wth is going on there...well, keep waiting

Secret

Semiclassical

17:46

actually, the 'problem' feature is not so strange I think. it's basically the same as what's going on near (0,-2)

vzn

youre working with bohmian trajectories? wouldnt one expect some symmetries? where they are missing? maybe a way to debug it...

Semiclassical

actually, there are some symmetries. they're not just so obvious in that picture

if i plot only on [0,pi], by contrast:

vzn

@Semiclassical the picture is not symmetric. horizontal/ vertical gradient wise vs the color coding. or rather (stating the obvious) the gradations are symmetric but the lines arent.

Semiclassical

there's no exact symmetry, I guess, but there is an approximate symmetry if you rotate the figure by 180 degrees (ignoring the direction of the arrows)

vzn

lol ACM is probably gonna get annoyed about now :P

Semiclassical

17:50

That may be a reflection of what the classical trajectories would be, though

Secret

Why will we expect the wavepacket to tunnel completely into the second well though, I thought only half of it will tunnel?

Semiclassical

I think the lesson goes like this. Suppose I localize a particle in the first well

If very soon after that I measure its position, I'm most likely to find it still in the first well

if I wait a bit, though, then apparently I'm more likely to find it in the second well

and if I wait a bit longer, more likely in the first well

and back and forth

Secret

I see

vzn

ok, sounds like a lorenz-like strange attractor setup

Semiclassical

it's a bit goofy

enumaris

17:52

Maybe if you made the second well deeper, you can find it more often in the second well :D

Semiclassical

probably

this is a symmetric well, so i"m not sure what would happen for an asymmetric one

Secret

I am still more interested on those features I labelled "hmm" though, I wonder what causes those blips?

Semiclassical

yeah, I'm not sure what to make of those either

Cows

going to demo code to clients . Show time

enumaris

@Cows gl :D

Terry Bollinger

17:55

@bolbteppa what the heck, in case you drop back by, here's Glashow's little cubic fermion mnemonic (it's all there was in his paper!) from 1980: goo.gl/AY2gKe, and here's an example (out of date but reasonably accurate) of how such cubes correspond to quaternion unit vectors (the three \xi coordinates; I've since switched back to simple ijk, with IJK for the flipped and superimposed left-handed quaternion space for left fermions: goo.gl/JxUVeh

Secret

My experience on going though many seminars in various subjects told me that: If for any vector plots you see anything that spirals around, whirls around or acts like a bump, it is either behaviour due to special values in the function or it is reflecting some strange physics predictions

I don't know how one can zoom into those "hmm" s though, maybe there's a saddle point somewhere in there

Semiclassical

that's what I get if I set up the classical problem and consider a few trajectories starting at x=1 with various initial velocities

some obvious things. one is that the trajectories can now cross a lot; this is allowed classically, whereas with the bohmian trajectories you usually don't get crossings

bolbteppa

@TerryBollinger cool will look into it

Semiclassical

(the only times when you can get crossings, I think, is when $\psi$ blows up. that's definitely true of the localized initial state, and evidently it's also true at $t=\pi$ etc)

so this is already an example where the classical trajectories are considerably different than the bohmian ones

I'll also note that the propagator I got from the initial paper had $K(x,t|x',t')$ being $2\pi$-periodic in $t-t'$

and the bohmian trajectories should inherit this

(I expect that's the mechanism behind all the trajectories passing through the same point at t=pi)

I'm sorta weirded out by it, if I'm honest

but I guess it's not so strange for a propagator in QM to have explicit poles with respect to time, at least when dealing with a bound system

as far as symmetries go, I think it is the case that the trajectories are symmetric under a combined x->-x, t->pi-t operation. (it's not entirely possible to judge that from the picture, since it's not sampling in a symmetric way)

Novalium Company

18:22

Guys, I was just wondering, every cell in our organism has the same DNA, except the sperm and egg, that's why humans are not exactly copies of the parents?

My question is is this why all of the humans are not all the same?

Semiclassical

yep, $v(x,t)=v(-x,\pi-t)$

Abcd

in Mathematics, 1 min ago, by Abcd

danielunderwood

\mathcal{E}?

Abcd

What is this symbol's mathjax code?

danielunderwood

$\mathcal{E}$

Abcd

18:25

$\mathcal{E}?$

danielunderwood

$\mathscr{E}$

\mathscr{E}

Abcd

$\mathscr{E}$

no its different

Semiclassical

$\xi$, maybe?

danielunderwood

detexify.kirelabs.org/classify.html?

Semiclassical

i guess that's not right

$\zeta$

nope, not that either

Abcd

18:27

@danielunderwood thats where the screenshot is from lol

I couldnt find it on detexify

danielunderwood

oh lol

Does it reference a certain quantity or something?

Abcd

@danielunderwood emf, generally induced emf

Semiclassical

that's definitely $\mathscr{E}$ in some particular font

Abcd

$\mathbcal$

:/

Semiclassical

good look figuring out which font they used for it, though

Abcd

18:30

$\mathcal$

5

Q: How to write "ℰ"

could anyone please tell me how I can write "ℰ" in latex in both text and math enviroment? I have already tried $\mathscr{E}$ and $\mathcal{E}$, but none of those are satisfying. What I want to have is but not any of these

Semiclassical

myself I tend to use \mathcal{E} : $\mathcal{E}$ for emf

because making a curly E is easy

Abcd

I dont know what they are talking about in the answers

is that (my) E possible or not?

Semiclassical

in latex, yes. in mathjax, probably not

Abcd

$$\documentclass{article}
\usepackage{amsmath}
\usepackage{mathrsfs}

\DeclareFontFamily{U}{calligra}{}
\DeclareFontShape{U}{calligra}{m}{n}{<->callig15}{}

\newcommand{\calE}{{\!\!\text{\usefont{U}{calligra}{m}{n}E}\,\,}}

\begin{document}
Here is a $\calE$ (script E)

Here is a $\mathscr{E}$ (script E)
\end{document}$$

Semiclassical

yeah

Abcd

18:32

Oh such a pain

Semiclassical

mathjax is good for some things, but it's not able to do packages as far as I know

Abcd

leave it

Anonymous

@Abcd MathJax is a watered down version of LaTeX for web-browsers. mathbcal won't work here.

Abcd

hmm

Semiclassical

it's the kind of thing you can worry about if you write a full paper/article

but if you're just typing stuff up here or for a question, just find an equivalent symbol in mathjax

Novalium Company

19:00

Guys, is mRNA made of the same stuff that makes up DNA?

Semiclassical

mRNA = messenger RNA is a special case of RNA, so like DNA it's a nucleic acid.

"Like DNA, RNA is assembled as a chain of nucleotides, but unlike DNA it is more often found in nature as a single-strand folded onto itself, rather than a paired double-strand. "

so the underlying substance is the same (nucleotides) but the structure tends to be different

Anonymous

@NovaliumCompany No, they differ in chemical composition

Novalium Company

Well, I thought mRNA would copy the DNA?

In structure?

Semiclassical

more details here: en.wikipedia.org/wiki/RNA#Comparison_with_DNA

Anonymous

@NovaliumCompany Transcription is probably not such a simple process

Anonymous

19:04

Check a biochemistry textbook for details maybe

Novalium Company

I'm already reading one.

Anonymous

Transcription (biology)

Transcription is the first step of gene expression, in which a particular segment of DNA is copied into RNA (especially mRNA) by the enzyme RNA polymerase. Both DNA and RNA are nucleic acids, which use base pairs of nucleotides as a complementary language. During transcription, a DNA sequence is read by an RNA polymerase, which produces a complementary, antiparallel RNA strand called a primary transcript. Transcription proceeds in the following general steps: RNA polymerase, together with one or more general transcription factors, binds to promoter DNA. RNA polymerase creates a transcription...

Anonymous

I'm not feeling motivated enough at the moment to look up the details :P So you'd have to search for yourself

Terry Bollinger

@bolbteppa this more recent figure is easier to read: goo.gl/hcACav The strong force anticolors are the ijk unit axes, inverted to make positrons positive. Since this is an excerpt from a rather long draft paper (I like open approaches), a lot is unexplained. But geometrically, the positron in this figure is simply the sum of the two vectors (i+j+k) (e+ left handed) and (-I-J-K) (e+ right handed), superimposed on the same space. The epsilon and chi spaces are, well... something new.

Novalium Company

Ha, ok no problems. You guys are really gods of knowledge.

A quick question. In the polypeptide chain, made of up of amino acids, in each amino acid group, there is that R. Is whatever you put as R what determines the function of the protein?

Semiclassical

19:14

probably not, given how much people talk about protein-folding and the complexity of such

Novalium Company

So it's still not understood what determines the function of the protein?

Semiclassical

no clue

the fact that nearly identical amino chains don't always fold similarly (quoting from wikipedia) makes me think it's complicated

my impression is that the mapping from protein structure to protein function is really tough

Novalium Company

Maybe the protein has a soul :D

ACuriousMind

@NovaliumCompany I'm not sure what you mean by "determining the function". A protein is a complex molecule, what it can "do" is determined mainly by both its shape and what it can (easily) bond to. Do not commit the fallacy of thinking that such a molecule has a predetermind "function" inherent in its molecular structure - what it does in any given organism is as much a consequence of the environment and what it can interact with as of its molecular structure.

Novalium Company

So simply said, a protein is a very complex thing, with strongly varying functions and structures?

ACuriousMind

19:19

How the molecular structure determines the shape is in itself a very rich subfield of physical chemistry

Novalium Company

I always wonder, do these things really have that shape. I mean, the chromosome for example has a weird X shape, is that in reality? And also the DNA for example, it also has a specific shape but if you were to look at it (somehow) would it really have that shape?

I mean, what does even determine their shapes (if they even have)

ACuriousMind

@NovaliumCompany Chromosomes definitely have that shape in a straightforward sense - you can see them on a karyogram

Semiclassical

the passage from a microscopic picture to a macroscopic one (or even a mesoscopic picture) is really hard in biochemistry

Novalium Company

It's like the chromosomes have their own shape language, that's really cool.

Semiclassical

it's far easier to say "this is what we see" than it is to determine how the things that are seen come to be

ACuriousMind

19:24

For a proteine, or molecules in general, you have to take most of the common depictions with a grain of salt - the angle and positions of the nuclei are meaningful enough, but remember that all bonds are in the end just shared electron clouds, van der Waals forces and similar things, and not "of substance".

Novalium Company

Is it possible for example one part of the chromosome is negative and the other positive and they kinda attract? (lol, what am I talking about...)

ACuriousMind

I don't know either what you're talking about :P

Novalium Company

Well, it looks like the chromosomes are bent. Maybe that is cause if one side of the chromosome has more electrons than the other side, and they kinda of attract and bend the structure?

Semiclassical

I imagine some things are known about the mechanisms of chromosome formation

But it’s probably not so simple as electrostatic repulsion

Novalium Company

(or attraction)

Semiclassical

19:30

A chromosome is a rather complicated object, with a lot of internal structure

By contrast, a molecule is sufficiently simple that electrostatics + quantum mechanics are largely enough to understand their structure

Novalium Company

After the polypeptide chains are produced from the ribosomes, do they go to the Golgi to get sent where they should go? (For example outside of the cell, or stay inside)?

Actually, proteins and lipids that should stay in the cell, do they even go to the Golgi?

enumaris

hmmm

dun dun dunnnnnnnnn

4

Cows

20:49

Just made small changes after meeting with client.

Heading back to him in 30 mins hehe

enumaris

21:03

nice

Transcript for