Building a fly brain in a computer

https://www.cifar.ca/cifarnews/2018/10/25/building-a-fly-brain-in-a-computer

by breck • 8 years ago

99 29 8 years ago

bastih• 8 years ago

It's interesting to see biological system and implementation next to each other. Are there any similar visualizations for other ML-implementations?

speps• 8 years ago

This is a similar project: http://openworm.org/

greenstonekid• 8 years ago

I wonder what the computational difference between the fruit flies visual system and the computer system used to run the model is?

jobigoud• 8 years ago

Can it feel pain? I know pain is kind of tricky to evaluate in biological insects as well but that's a discussion we will have to have at some point.

Hei1Fuya3• 8 years ago

The headline is misleading, they built a neural network based on fly visual neurology, not a whole fly brain.

And even if it were a full brain, effective societal consensus seems to be that insects don't have a right to humane treatment.

jessriedel2• 8 years ago

based on fly visual neurology

That seems generous, I think? They restricted the input data to visual acuity at the level of a fly. But it doesn't look like the neurology actually influenced the design much:

This, combined with the discovery that the structure of their visual system looks a lot like a Deep Convolutional Network (DCN), led the team to ask: “can we model a fly brain that can identify individuals?”

Hard to tell without reading the paper.

kpil• 8 years ago

I wouldn't worry about that until much much larger animal models. Besides, it's just the visual system.

I don't know where there's a moral limit. Several millions of neurons? A billion?

viraptor2• 8 years ago

I don't think simulated brains can feel pain any more than a float variable "current_pain_level" set to 100 can "feel pain".

andyjohnson02• 8 years ago

I don't think simulated brains can feel pain

I'm not sure why you say that. Does pain have some property that makes it impossible to simulate?

If an organic animal can feel pain and it is possible to create a 100% accurate computational simulation of that animal, then the simulated animal must also experience pain. To me the interesting questions are at what level of simulation fidelity <100% (a) does the simulated pain be said to be "real", and (b) does it make sense for the simulation to be said to give rise to a subjective experience (if at all)? And the follow-on questions for both are "how?" and the "why?".

This whole issue of simulated pain is part of the premise for Roko's Basilisk [0].

[0] https://rationalwiki.org/wiki/Roko%27s_basilisk

empath75• 8 years ago

What’s the difference between that and your own pain neurons triggering?

dragonwriter• 8 years ago

I don't think simulated brains can feel pain

I don't think we have the kind of understanding of what it means to feel anything for there to be any meaningful, informed opinions on that topic.

thomasfedb• 8 years ago

As you alluded, mostly likely insects don't have what we'd consider to be an emotional response to pain. Hence, likely not.

scotty79• 8 years ago

"Sometimes, if you were going to have any hope of getting useful answers, there really was no alternative modelling the individuals themselves, at the sort of scale and level of complexity that mean they each had to exhibit some kind of discrete personality, and that was where the Problem kicked in.

Once you’d created your population of realistically reacting and – in a necessary sense – cogitating individuals, you had – also in a sense – created life. The particular parts of whatever computational substrate you’d devoted to the problem now held beings; virtual beings capable of reacting so much like the back-in-reality beings they were modelling – because how else were they to do so convincingly without also hoping, suffering, rejoicing, caring, living and dreaming?

By this reasoning, then, you couldn’t just turn off your virtual environment and the living, thinking creatures it contained at the completion of a run or when a simulation had reached the end of its useful life; that amounted to genocide.” - Iain M Banks "Hydrogen sonata"

I love to watch Star Trek with its upbeat "always do good" general vibe when contrasted with completely oblivious cruelty they show towards simulated life.

taneq• 8 years ago

I love to watch Star Trek with its upbeat "always do good" general vibe when contrasted with completely oblivious cruelty they show towards simulated life.

Star Wars too. Although concerns about "owning sentient droids is slavery" are missing the point that Star Wars culture is fine with actually owning enslaved humans...

bwbw223• 8 years ago

Sounds buggy...

Phemist• 8 years ago

Very cool! Takes me back to the days when I first studied Laughlin's short paper on fly eyes [1]. Amazing how we've gone from this early experimental result to something as complex as simulating a full visual system!

[1] https://www.ncbi.nlm.nih.gov/pubmed/7303823

psk0• 8 years ago

I couldn't see whether they have ruled out that ML is clustering based on the image noise and lighting.

I wouldn't be surprised if the same clustering is achieved after masking big/essential chunks of the fly from all the images (for training and testing sets). If indeed the 'imaging features' are the driving weights in clustering, that means the flies should be able to recognize that they are in a different room.

anonlastname• 8 years ago

All they did is train a neural network to recognize flies. they haven't replicated the nervous system of a fly. To say they "built a fly brain" is wrong because that actually has been done for nematodes and it's different.

Building a fly brain in a computer

Related Stories