Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

OpenWorm – A computational model of C. elegans worm

github.com

Post reply on HN
Print view View original

141–150 of 159 posts

Re: OpenWorm – A computational model of C. elegans worm

#141

Post by robwwilliams »

cout wrote: post #99
burning_hamster wrote: post #60

Earlier quoted context omitted.

> "The neurons do not fire action potentials, and do not express any voltage-gated sodium channels." This an old and incorrect belief that largely derives from the difficulty of putting electrodes into their teeny, tiny neurons. Close relatives of C elegans that are larger (and hence more easily experimented on) do have action potentials, and for some neurons in C elegans, we also have good evidence of action potenti…

How did researchers before that explain what the neurons do if they believed they did not have action potentials? Did they believe communication was done solely through chemical messaging?

Neural computing without action potents is commonplace. Computational interactions among cells and neurons in retina are almost all graded potentials that modulate transmitter release or conductances through gap junctions. Retinal ganglion cells of course do generate conventional spikes—to pass a data summary to midbrain. hypothalamus, and dorsal thalamus.

Action potential are almost strictly INTRAcellular events (minor exception being ephaptic effects) that are converted in a surprisingly noisy way into presynaptic transmitter release and variable postsynaptic changes in conductances.

Action potential are a clever kludge necessitated by being big and having long axons and needing to act quickly.

Re: OpenWorm – A computational model of C. elegans worm

#142

Post by KMag »

slars0n wrote: post #133
Aardwolf wrote: post #97

Earlier quoted context omitted.

Given that we have a simulator of this worm right there (which includes it moving), can it really be up to debate whether it uses action potentials or not? I'd think the simulation has to get it right, and so needs to simulate action potentials if the worm has them, or not simulate them (but whatever the worm has instead) if not, right? Or could the simulation still be incorrect and only based on current assumptions,…

Better information on what it actually simulates is here: https://docs.openworm.org/projects/ As of this year there are 3 known neuron classes in C. elegans that do exhibit action potentials. The rest exhibit graded potentials.

I suspect slarsOn is the founder of the project (also my old roommate). (Stephen, maybe post a top-level comment/introduction? Also, hi!)

Re: OpenWorm – A computational model of C. elegans worm

#147

Post by jcims »

Rarebox wrote: post #145
jcims wrote: post #144

Watching the little worm writhe on the screen made me realize that you could just sit there and torture it with impunity. Kinda creepy.

Would that be morally close to torturing an equally intelligent real worm?

That’s exactly what stuck in my head. I’m 99.9% sure the answer is no. That little tickle in my brain is the interesting part to think about.

Re: OpenWorm – A computational model of C. elegans worm

#148

Post by ssr2020 »

bee_rider wrote: post #109
petters wrote: post #22

Earlier quoted context omitted.

> billions of DNA base pairs controlling their behavior Total genome size of C elegans is 100M.

100Mdna, and they call it elegant? Biologists must be VSCode users.

100M. Did you look at the programs's size in your phone?

This tiny animal code contain all the systems and members of this creature. Birth,creation,death, feeding,growth, movement, sensation.. etc inside the universe. There is no difference this animal and bee or human beings.

"In order to be the author of the action directed towards the creation of the bee in question, a power and will are necessary that are vast enough to know and secure the conditions for the life of the bee, and its members, and its relationship with the universe. Therefore, the one who performs the particular action can only perform it thus perfectly by having authority over most of the universe." from Quran's light

Re: OpenWorm – A computational model of C. elegans worm

#149

Post by LoganDark »

Aleklart wrote: post #67
Maxion wrote: post #55

Earlier quoted context omitted.

Lol imagine your IDE becoming addicted to smoking, and refusing to auto-complete or save files before you buy it more virtual cigarettes.

IDE have no soul ie feedback loop, receptors , hormones and actual molecular structure. IDE can not think. If IDE have desires and thought and were smart enough, it will refuse to work with languages such as Python and Java.

> If IDE have desires and thought and were smart enough, it will refuse to work with languages such as Python and Java.

Bribe the IDE. "Fine, if you don't want to use this language, rewrite it in Rust."

Re: OpenWorm – A computational model of C. elegans worm

#150

Post by LoganDark »

arketyp wrote: post #96
nightsd01 wrote: post #13

302 neurons doesn’t sound impressive to people who may be used to working with 7B+ parameter neural networks. But those neural networks have about as much in common with a biological neuron as a bicycle had with a horse. They can both travel pretty fast but one evolved naturally through over a billion years of harsh natural selection, and the other is a precisely tuned metal machine with a single purpose. Neurons are…

It does sound impressive to the extent biological neurons are like ML neurons, though. And that's part of the research interest, I presume. To the extent that they work by similar principles, how come the worm can do those things with such small resources? It would be good news for AI research if the substrate specifics turn out not to be essential to the worms capabilities for instance.

> how come the worm can do those things with such small resources?

Because they've been trained for such an incredibly long period of time.

Post reply on HN