« first day (3183 days earlier)      last day (29 days later) » 
00:00 - 23:0023:00 - 00:00

12:03 AM
@RyanUnger That's not special. We don't tolerate name-calling in general.
12:21 AM
@rob Random question. Why would a user have a desire to make few minor edits to many posts, some of which are years old (talking about xray0 if you have noticed their activity recently). I have been approving their edits. They aren't bad edits, but I would say a lot of them aren't too important either.
@rob I never said it was special
but in certain cases the users are indeed crackpots
@AaronStevens Those frustrate me to no end. Some of them actually add mistakes, or try to correct British English spelling with American English.
@JMac Right, yeah I agree on some of those, at which point I don't accept it. It's just on some of them I am tempted to reject as "no improvement whatsoever", but there are ones that are technically improvements, even if they are minor improvements
Yeah, a lot of those ones bug me because they're almost pointless in my opinion. Or they find an obvious edit, and then also add punctuation that doesn't always make sense; so you have to look at all the little things done instead of one obvious fix.
@JMac Yeah I agree. Do you choose to accept them if they are technically correct but pretty much pointless? Or do you just skip? I must admit I do like getting my suggested edits bar up though haha
12:47 AM
@AaronStevens Depends on how pointless, and if they introduce more mistakes. I rejected this one physics.stackexchange.com/review/suggested-edits/268484 and just got rid of "years years" because all his comma business was useless, possibly harmful. I approved a couple of his, but I had to add edits because he missed things or part of the edit wasn't good.
I assume he has good intentions, looking through questions trying to clean them up. But he's very over-zealous, and it makes the edit queue fill up with one person.
@JMac Yeah, I think I agree with you on this. The actions themselves aren't wrong, they just seem kind of pointless. And the additional errors don't help. I guess I will just continue to take it case-by-case. Thanks for the input
1:06 AM
lol! namecalling in general is not allowed, but there are specific exceptions leading to no enforcement. got that? :P
5 hours later…
5:52 AM
@EmilioPisanty I have a tone problem :(
@RyanUnger hard to invert
6:16 AM
"they hit the singularity at different times" - This may be misleading, because "time" in the coordinates of your diagram does not refer to the physical time of any real observer. The Schwarzschild singularity is spacelike, so everything hits it at the same physical time. — safesphere 16 hours ago
We're going to have to do something about this sort of persistent nonsense.
6:41 AM
@JohnRennie he was saying weird stuff here: physics.stackexchange.com/questions/492477/… as well
oh nevermind that's the one you linked
We have a few members who think they understand GR :-)
Incidentally you can do links using [link text](URL)
oh thanks I was trying to figure that out
GR is not as hard as generally thought, but there are some aspects of it that are really non-intuitive and this trips up people who have a superficial knowledge of the subject. In particular the nature of time in GR confuses people, as indeed it has confused our beloved safesphere.
I can't actually find the older comments in his actions, I guess when ACM deleted them they get entirely moved to chat.
but I remember @PM2Ring was getting suspicious
@JohnRennie did you use MTW or what text did you use primarily to learn GR?
@kylecampbell Good heavens no, MTW is not a book for the beginner :-)
Unless you're a beginner weightlifter in which case trying to lift MTW could be a useful part of your training :-)
6:52 AM
hartle then?
I started with Bernard Schutz - A First Course in General Relativity.
However this is really most suitable for someone like me who wants to understand GR without necessarily wanting to do research in it.
If you were learning GR properly I'd probably start with Carroll's book.
interesting, I had never heard of Schutz
I can't remember who recommended Schutz's book to me. In any case I wouldn't recommend it for a student.
I wouldn't recommend MTW either. It's a wonderful book and I rarely dip into it without learning something fascinating, but I don't think it's a good book to get started with GR. It tends to assume too much of the reader. Well, it assumed too much of me! :-)
7:05 AM
yeah, I suppose you could use it more as a reference later on when you've learned more
7:24 AM
MTW is good to learn about stuff not commonly in GR books
@TaeNyFan hi, did you want to ask something?
About the question I just closed possibly?
@Slereah weird sh*t in other words :-)
@JohnRennie not necessarily
MTW has a very good part on observers
which is rarely done in GR books for some reason
7:58 AM
Hi, assume we're thinking about the particle-like properties of a photon. Is it possible to describe the shape of the photon?
For example, like we would describe the shape of an atom saying there is the fuzzy electron cloud and the nucleus and so on.
@undefined the problem is that a photon can be localised to different degrees so it doesn't have a shape.
A photon with a precise wavelength is completely delocalised i.e. an infinite plane wave and as such doesn't have any shape at all.
Real photons have some degree of localisation but this can vary from highly delocalised to very higly constrained in space.
hello @JohnRennie nice to see you again
@undefined hi :-)
so despite the the particle-wave-duality it is still more like a wave than anything?
Have a look at:
Q: Do photons truly exist in a physical sense or are they just a useful concept like $i = \sqrt{-1}$?

JensReading about photons I hear different explanations like "elementary particle", "probability cloud", "energy quanta" and so forth. Since probably no one has ever seen a photon (if "seen" it supposedly - and rather conveniently - ceases to exist) but many experiments seem to verify its properties ...

Particles are a surprisingly elusive concept when it comes to quantum field theory.
8:05 AM
They're not that mysterious but QFT people are shy about actually talking about them
same as in QM, just have a localized wavepacket
Although of course
Malament's theorem
It can't be very localized
@JohnRennie but very fascinating for me :)
why are they shy talking about them?
Probably WORK SHY
the lazy bums
although if you want a good discussion of actual photons in QFT I'd advise reading a quantum optics book
they're the ones that have to deal with real photons
I mean, how is it then possible to create a hologram of a single photon when they are like a wave? Or to combine them?
Photons are neither waves nor particles, alas
most optics books I found so far were mostly dealing about other stuff and about the particle thing. But I'm sure there are some
8:09 AM
although there's no particular reason why being waves would preclude them from that
you can have fairly localized waves
ya, they are sometimes this and sometimes that :)
I still have to
8:47 AM
I have 2 very very stupid question but imma ask anyways
in $4x(x^2 - 2) = 0$ we do $4x = 0$ and $x^2 - 2 = 0$. How does the latter turn into $x = + or - \sqrt 2$ ?
Why the plus or minus thing
The equation $x^2 = 2$ has two solutions.
Both $\sqrt2$ and $-\sqrt2$ are solutions. To see this try putting those values into the equation.
Got it
a very very very very stupid question. How $x^2 = 2$ turn into $x = \pm \sqrt 2$?
Use \pm for the plus or minus symbol
If $x = -\sqrt2 = -1 \times \sqrt2$ then $x^2 = -1 \times -1 \times \sqrt2 \times \sqrt2 = 1 \times 2 = 2$
So $-\sqrt2$ is a solution to the equation.
Sure, but how does the power component in $x^2$ turn into $\sqrt 2$ component?
The square root function is in general two valued i.e. for any argument $x$ it produces two values.
8:55 AM
I know you are an intellectual being, but please explain it to the poor urban very little above averages like me :D
Do we multiply both sides by something or?
got it
We multiply both sides by a square root. (very dull explanation but it satisfies)
A graph of $y = \sqrt{x}$ looks like this:
For all values of $x$ greater than zero there are two values of $y$.
Why are you overcomplicating the problem. Just multiplying both sides by a square root does the job.
@NovaliumCompany I'm telling you the reason why there are two square roots.
I got that part. It satisfies the = 0 part.
Meaning, putting $\sqrt 2$ or $-\sqrt 2$ works.
@JohnRennie I was talking about my second problem. Which I solved :P
Nevermind, everything is clear now. Thanks @JohnRennie
@NovaliumCompany cool :-)
9:56 AM
I hope @vzn will be happy that my list of dumb problems include solutions with Bohmian mechanics
It's gonna be hard because those grifters can't seem to decide what the equations for it are
Hopefully they're all equivalent
The fun part will be doing quantum experiments in the classical setting
DOUBLE SLIT EXPERIMENT : The particle crashes into it
10:32 AM
hey I got a dumb ass idea
Can you solve the least action of a point particle by finding the action of every possible path using Fourier series
Just express the path on $[t_1, t_2]$ as a Fourier series, compute the action of that path, show that if any $n \neq 0$ term added will make the action larger
it should be doable if I can show that any term $n \neq 0$ always adds a positive term to the action
Oh wait
10:49 AM
I'm Beff Jezos
$$S = \int_{t_i}^{t_f} dt \dot{\vec{x}}(t) \cdot \dot{\vec{x}}(t) = 2\pi T^{-1} \sum_{n = -\infty}^\infty \| n \vec{c}_n \|^2$$
How to show that it is minimal if we put every $c_n$, $n > 0$ to $0$
I can see how to do it for some cases but not all
Might want to make it a PSE question
In my DNA, how is it possible with relatively few nitrogenous bases to describe the position of each atom in my body, the form of my fingers, my overall structure, my organs, fingerprint, predetermined neural circuits...?
Like, are basically x, y, z position coordinates encoded for each atom?
Oh man, this is gonna be like
The Fourier series for a sawtooth function
11:20 AM
Q: Solving free particles with Fourier series

SlereahHere's a silly idea : take the action of a free particle, $$S = \int_{t_1}^{t_2} \dot{x}^2 dt$$ Our configuration space is the space of $C^1$ functions over $[t_1, t_2]$, which is spanned by the Fourier series $$x^a(t) = \sum c^a_n e^{i\frac{2\pi n t}{T}}$$ So that every path is represented ...

11:34 AM
Hm, maybe I can show that any difference between a random action and the straight line is $> 0$
although slightly cheating because it's just me knowing the extremal path in advance
12:15 PM
@NovaliumCompany DNA does not describe the position of each atom in your body. It mostly just encodes how to make your body's proteins. A huge amount of the information required to build your body comes from the context that the cells find themselves in, it doesn't need to be encoded into the DNA. But you should ask biologists about this, not physicists.
1 hour later…
1:40 PM
A: Delayed erasure experiment

Árpád SzendreiIn the original setup, a laser generates individual photons, 351.1nm, that passes through a double slit. After the slits, there is Spontenaeous parametric downconversion of the photon into two photons of 702.2nm. It prepares an entangled two photons state. This converts the photon into two ident...

Árpád hits a new record
2:08 PM
2:25 PM
there's the non-relativistic constrained point particle
@EmilioPisanty "This answer is amazingly wrong"
what happened to having the right tone
He said it was amazing
what more do you want
appropriate tone
imho hahaha specially since he then took the time to right an answer
@PM2Ring If I cut the top of my finger (where the fingerprint is), the same fingerprint will grow again. Doesn't that mean that the information about how the atoms on my fingers should be placed is stored somewhere?
The third (final) step of the solution is a bit confusing here: tutorial.math.lamar.edu/Solutions/CalcIII/DirectionalDeriv/… Help :P
I'm a bit confused on the dot product part
@NovaliumCompany I'm pretty sure that's just inaccurate, depending on what you mean. You can cut off the epidermis without issues, and have your fingerprint grow back; but if you damage the dermis it also can change your fingerprint AFAIK. It's not something purely determined by genes; development plays a role.
2:38 PM
@JMac Oh ok. But still, what kinda info does DNA store? Maybe what proteins to produce? If so, how does proteins come to determine so many things in us?
@NovaliumCompany DNA is more like a recipe than a plan.
If you are making a cake the recipe doesn't tell you the positions of all the molecules, it just tells you what chemicals (flour, eggs, etc) are needed and how to mix them.
@NovaliumCompany Really really complex processes. I'm not sure how much you've studied biology, but the mechanisms are usually explained in introductory bio courses. How it all results in an actual person or changes in people is a huge field of research.
DNA basically specifies what proteins to mix together so the reaction creates an organism.
The proteins are the cooks and are guided by the recipe?
Initially your mother is the cooker.
Then by the time you're born your body cooks itself.
2:43 PM
AFAIK it is still the proteins that are the "cookers", regardless if it's your mothers kitchen or your own.
The point is that your exact form isn't specified by the DNA.
Physical processes like diffusion are important as well.
Got it. What specific examples are there of DNA completely (or almost completely) determining something?
I guess I'll google that :P
@Gyromagnetic cute bear
@NovaliumCompany That might be better suited to biologists. As far as I know, it would be hard to draw that line; because environmental factors (and even inherited factors?) can influence how the DNA gets expressed; even if the DNA sequence is identical.
@JMac epigenetics?
@NovaliumCompany Yeah, that's why I said "and even inherited factors?" It's been awhile since I studied bio, and I was never that invested in it, so I don't know a whole lot about epigenetics. I know that protein/DNA interactions are quite complicated though. The DNA is itself manipulated by proteins that get formed from DNA instructions; so it can be hard to nail down what caused what.
2:52 PM
@JMac Alright, thanks
Hope CRISPR gets better in the future so we genetically stack another neocortex on top of our current one and see what happens xDDD (my stupid ideas again)
@NovaliumCompany sup m8 ;)
@NovaliumCompany I honestly would expect us to be able to augment our minds with computers before we would get to the point where we understand genetics that well.
@JMac The best thing about the future is that you can never be sure what'll happen ;)
@Gyromagnetic sup, how r ya
@NovaliumCompany fine, urself?
@Gyromagnetic fine, doin' som partial derivs
Paul's Online Notes is pretty damn good
3:02 PM
You know what's actually a good book on constrained systems
Quantization of fields with constraints, by Gitman and Tyutin
3:34 PM
Is The Terminator worth watching?
I'm kinda between Automata and The Terminator. I think I'll watch Automata first and leave The Terminator for next time.
Time to get a doner kebab then
cya guys
4:09 PM
Is there a nice simple dupe target for physics.stackexchange.com/questions/493004/… ? I had a look at several questions about the relativity of simultaneity, but they were mostly concerned with time dilation, so may not be so suitable for this OP.
Puk's answer there looks mostly ok to me, I don't think it deserves the downvotes.
4:59 PM
@JohnRennie Maybe I'm missing something, but I don't see how this particular comment is nonsense. The distance between the two events of the particles hitting the event horizon is space-like, isn't it? (time-like curves in Kruskal coordinates have to slope towards the y-axis more than the light-like 45° lines, like the trajectories of the particles do) So there really is a frame in which the events happen simultaneously.
@ACuriousMind The Schwarzschild singularity is spacelike, so everything hits it at the same physical time makes sense? Oh well.
@JohnRennie What is "at the same physical time" supposed to mean there? That's a frame-dependent statement!
The standard lore is that there is a frame in which two events happen simultaneously if they are space-like separated, and a frame where they happen at the same place if they are time-like separated.
I'm not sure how one could argue this doesn't hold within a black hole other than "that would be weird".
@ACuriousMind that's something you need to ask safesphere. At a guess I would hazard that he shares Duffield's conviction that the Schwarzschild time coordinate is the real time and that is what he means by physical time. But then who knows?
@JohnRennie But you also say: "They cross the horizon at different times and they hit the singularity at different times."
For the horizon, this statement is defensible since the distance between the events is light-like there, but I don't see how you can argue it for the singularity
I guess I'm not saying that safesphere is right, I'm saying your (and Andrew's in the other answer) claim is equally questionable
I'll no doubt find the enthusiasm to pursue this later.
5:13 PM
::turns on decoder for British:: Ah, drat. ;)
@ACuriousMind you'll have to forgive @JohnRennie -- doubtless he is out celebrating the belated arrival of the Winds of Change to good old Blighty
Boris Johnson is now our prime minister. The end of days is here.
@EmilioPisanty The Scorpions are playing in Britain?
@ACuriousMind I've never heard Boris described as a scorpion tbh
I like that song. It tends to be sneered at by the cognoscenti because it was such a big hit, but it's a good song.
@EmilioPisanty the adjective venomous fits though.
5:25 PM
@JohnRennie agreed
@John commiserations =|.
It's tempting to say it can't get much worse, but on October 31st it will do.
5:48 PM
the EU must fall
6:10 PM
Is this answer ok, or is it a bit spammy? physics.stackexchange.com/a/493207/123208
@PM2Ring The last paragraph is superfluous, so yes, it's spammy.
@ACuriousMind Thanks. It's an interesting concept, though I don't know if using a mechanical device to "play in" guitars will appeal to a lot of musicians. I'd rather have a guitar that was played in by one or more good players... but it'd be interesting to hear what the results are like.
2 hours later…
8:37 PM
@Slereah I have a French paper that I need help with
8:54 PM
Vito Volterra Bito Bolbteppa
In mathematics, the Volterra integral equations are a special type of integral equations. They are divided into two groups referred to as the first and the second kind. A linear Volterra equation of the first kind is f ( t ) = ∫ a t K ( t , s ) x ( s ) d s {\displaystyle f(t)=\int _{a}^{t}K(t,s)\,x(s)\,ds} where ƒ is...
This integral equation stuff gets nuts pretty fast
@Slereah why besmirch with the Bohm stuff on there
9:37 PM
The big companies seem to have occupied all the good ideas. Like, they are investing billions into space, ai... How are people supposed to start good tech companies when it seems like everything is already taken by the big ones?
@NovaliumCompany Big companies are actually rigid, and also psychotic structures. You can try to utilize their inflexibility
@NovaliumCompany Stand on the shoulders of giants just like they do (and probably pick something less obviously profitable and make it profitable; because obviously the people who already have money can invest more than you in ventures that are very likely to give profit)
...Why would people be "supposed to start good tech companies"?
@ACuriousMind I mean, people who want to, entrepreneurs
@peterh What do you mean?
@ACuriousMind All the good tech companies of the world history were already founded?
@NovaliumCompany What I wrote.
9:43 PM
@peterh No, I question the idea that "start a tech company" is a goal one is "supposed" to have :P
@peterh How should I analyze theirni flexibilit? What is psychotic structure?
E.g. there's lots of awesome tech in free/open software that don't have a "tech company" behind them
@NovaliumCompany there's no universal rule saying you're supposed to be able to start tech companies. It's like me saying "How am I supposed to start a Google competitor when they have so much time spent developing theirs?" The obvious answer is generally "you don't unless you have significant investment capital"
@ACuriousMind I meant "want", not "supposed to" sorry
Another thing to point out is that if you think all the good ideas are already occupied, there is a distinct possibility you're not good at coming up with good ideas :P
9:47 PM
@NovaliumCompany Work by some of them, and you will see. Many people likes it. There are years and countries where they are the only well-paying employer, in other years and countries they pay the same, in exchange they are also toxic
@NovaliumCompany If you have a 10-man company, and a customer comes to you. You can start on that day. In a big company, it is at least weeks, but it can be even months.
@ACuriousMind I haven't given much time and thought into business ideas. I said "seem" to be occupied. Of course there are many many good ideas but it's hard to get to them.
@peterh Simply put, big companies have large momentum
@NovaliumCompany If the command chain is long, what the people on the lowest level is doing - who are doing the actual work - will differ far more what the top management really wanted.
@NovaliumCompany sometimes you could do what someone else is doing but better - if mark zuckerberg just figured "oh there's already myspace" facebook never would have happened
@NovaliumCompany ...if they were easy to get to people would have already gotten to them. It's the nature of progress. The most obvious, easiest things generally get done first.
But still, if big companies spot a nice investment or idea, everyone else is doomed to do it
9:53 PM
@NovaliumCompany And they try to do it better. It's not a bad method of promoting progress
@NovaliumCompany Yes. And they are also far lesser effective. The ratio of 1) the total paid work 2) to the work which is directly converted to something valuable for paying customers. That is far worser by big companies.
@JMac yes, but big companies won't let the small succeed.
@SirCumference Some rule would help, an 1yr long ban should mean imho a clean, or nearly-clean past.
@NovaliumCompany If you come up with something useful, there are generally intellectual property/patent laws that help to protect people who have good ideas. Most reasonable systems have checks and balances to mitigate the ability to force monopolies and steal ideas. The issue is that profitable companies can afford to invest in research; but that's a pretty tough problem to get around altogether
@NovaliumCompany Sure, why would they? Make no mistake, the objective of big companies is not to produce useful tech or to foster an enviroment where others can try to innovate and outcompete them - it's to make money.
9:58 PM
@NovaliumCompany There is also a big luck component in the system. You try to do your best in your whole life. Sometimes you end up as a $200000/yr programmer in the silicon walley. Another possibility is to end up for much smaller, on a much lesser paying job market in another country. With the same motivation and skills. Only because luck.
@ACuriousMind Got it. So to succeed, you have to work more, do things better, put in the effort and if you are lucky, you might join the big ones in the future.
That's...not what I said
@ACuriousMind I haven't said you've said it
@ACuriousMind On an ideal market, these two should mean the same.
But that is indeed what adherents of market liberalism want you to think
10:02 PM
Ok, is there anyway to win over those big companies, without them eating you up?
@NovaliumCompany What do you mean by "win over"?
@ACuriousMind "But that is indeed what adherents of market liberalism want you to think" <----this
@JMac make a big tech company for example, join them
@NovaliumCompany ... they do hire people. They even promote from within in some cases.
@NovaliumCompany They are a few, and we are lot. This obviously seriously worsens our chance. Another, maybe phylosophical question: why would it be so good, to have billions of \$. Our world has some inherent, wealth-independent constraints. We all will die. None of us can go back in the past and get the love of the girl we loved in the high school 20 year ago. No money can solve these problems.
10:05 PM
@JMac that was irrelevant?
@NovaliumCompany What do you mean by "join them" then? You mean make your own big tech company?
Because making your own tech company isn't likely to win over other big tech companies
@peterh I don't give a single damn about money
...then why do you want a big tech company?
... yeah this is all getting more confusing to me. I really don't understand the motivations or real goals here
@JMac What I mean is, how to make a big company without the big players immidietly creating a competition that will be close to impossible to beat
@ACuriousMind lol.....
I'd be better off in the lottery rather than business if it was for the money
10:09 PM
You can innovate without a company, if it's just making awesome tech that you're after! E.g. Linux and Git have arguably shaped the world of software for more than a decade, and neither belongs to any company at all.
Whatever, I have to go to bed. Thanks guys for the nice dicussion :p
git is too hard to use
after failing to be able to use that I knew I was destined for academia
@RyanUnger No! It has a logic! It is only not just well advertised. If you understand the logic, you can do everything in git. Well, some very complex operations (like retroactively, recursively manipulate the history) needs a little scripting, but practically never will you face them.
@RyanUnger It's only hard to use if you try to use it without understanding what it's doing :P A bit like math, really ;)
@RyanUnger Forget everything, start with a command line. Only the command line git can do everything what you want. Forget all the GUI tools.
10:13 PM
It's too late for me, peter
@ACuriousMind Interesting question, to whom do they really belong
Well, Torvalds owns the Linux "brand", but the software itself? No one and every one! You can't really "own" something that's licensed under the GPL
Someone must own it
@RyanUnger Never late! Stats show, as you grow older, the cognitive skills decrease only a little, you can still learn, well not so quickly than the young ones, but not far away from them.
My cognitive skills have degraded due to alcoholism, not old age
10:17 PM
@RyanUnger What would it even mean to "own" a software that is by its license free for everyone to use and modify?
@ACuriousMind Maybe they belong to same unwritten, collective entity. For example, to the collective conscience of the humanity.
anything licensed under GPL is free for people to do essentially whatever they want with the software, so I'd say that's effectively not "owned" by anyone
@RyanUnger You have also a lot of experience, what the youngs simply could not earn until now. What I see on the few old programmers I know, that they have somehow... given up. And I decided, that I won't, ever.
@RyanUnger But I had serious problems due to the idiotism and various background attacks. Instead drinking, time to download something and play with it. For example, a Linux would be a good start. Not in a VM, but on your bare metal.
I'm not drinking right now
@RyanUnger Start in the deep water. Debian.
10:22 PM
I think we should clarify that Ryan is neither an alcoholic nor a programmer :P
I liked Debian when I used it actually
I almost forgot about Debian... brings back memories
I'm using it right now and much prefer it over the Windows I have at work
@ACuriousMind uhhhhhh are you sure
@RyanUnger Very sure about the 'programmer' part!
na klar that part was never under contention
10:25 PM
@ACuriousMind Of course in your work environment, if I suspect, it is totally unimaginable to solve a problem by installing Linux to your workstation and solve with that? :-)
@ACuriousMind You can't even think on that? You can't even think on far lesser "irregularities"? :-) :-) :-)
@ACuriousMind :-) Oops :-)
10:41 PM
Hello there
how is $?_1=\Delta x\cos{\theta}$ a valid approximation for the change in the line's length?
u guys see this news about Intel's new Loihi Chip?
Neuromorphic computing...and spiking neural networks...
I am kind of convinced, but I still feel "something in my heart"
hard to read dem scribbles
10:53 PM
@enumaris I have to admit that I don't really understand what a neuromorphic chip does - it is basically hard-wiring some particular algorithm?
I can't say I "really" understand them either lol, but this Loihi chip seems to be some architecture which is well suited for running spiking neural networks by simulating "neurons" rather than standard 0-1 gates...maybe...so the "neuron" in these chips accumulate electrical impulses sent by other neurons and when that electrical impulse exceeds some threshold potential it discharges a pulse.
And that is cheaper than simulating the equivalent amount of neurons on a general-purpose computer?
they make a lot of claims
but since I'm no expert in the field I can't really say...
They say: "Spiking neural network models, on the other
hand, are exceedingly poorly served by conventional
architectures. Just as the value of ANNs was not fully appreciated
until the advent of sufficiently fast CPUs and GPUs, the same could
be the case for spiking models—except different computing
architectures will be required. "
ANNs = artificial neural nets - the vanilla ones we are all familiar with
I have no real clue what I'm reading; but it sounds like they basically made a chip full of artificial neurons? It sounds cool, but I'm not quite sure I understand the application or need
I'm not a hardware guy so...I...I don't really follow the hardware parts of the paper lol
I think I follow to some extent - at least roughly - the "software" side
the spiking neural net side
00:00 - 23:0023:00 - 00:00

« first day (3183 days earlier)      last day (29 days later) »