I know that it sounds weird, but I read somewhere that VGGnet has roughly 14.5k neurons and 138/144M parameters depending on the exact architecture as it's one of the biggest.

Also I read here, that the biggest neural networks trained on super computers have roughly 80M neurons, which is similar to frog's brain.

If we had to compare the complexity of VGGNet to brain of some kind of animal (even simple one) what would it be?

  • $\begingroup$ Number of neurons in neural tissue doesn't matter much, as we have no idea how single neuron works, how neurons are connected, or how learning works there. Or whether there is learning in the first place, maybe it's hardcoded in DNA for simple animals. $\endgroup$
    – tomash
    May 24, 2019 at 9:34
  • $\begingroup$ @tomash Uh... no. We actually have a very good idea how some single neurons work. People in computational neuroscience and biophysics have been working on this since even before Hodgkin & Huxley got the Nobel prize 6 decades ago for breakthroughs in it. As for how neurons are connected, we have literally complete mappings of the entire nervous systems and connectomes down to sub-histological levels of some animals, like C. Elegans. As for whether "there is learning", Kandel got the Nobel prize for showing mechanisms of learning and memory in sea slugs so...... these are well-known..... $\endgroup$ Jul 19, 2019 at 2:14

1 Answer 1


According to this wiki: List of animals by number of neurons

The closest ones you got are:

  • Pond snail with 11,000 neurons.
  • Sea slug with 18,000 neurons.

I think this was more of a biology question since you covered the data science part.

  • 1
    $\begingroup$ This list is perfect as I don't need very precise answer. This was rather question out of curiosity. Thanks! $\endgroup$ May 19, 2019 at 16:12

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.