Is the visual cortex of newborn babies right off the bat capable of making sense of raw visual data, for instance, converting the constant stream of raw RGB images perceived by the eyes into a meaningful higher-level representation of objects in motion in a 3D world? Yes? No? If not, then does it mean this skill has to be developed over time by means of some learning mechanism?

If the visual cortex needs time to learn advanced visual skills, how does the actual learning mechanism work? Does the visual cortex have to optimize the connections between neurons, in a way analogous to how artificial neural networks optimize their parameters through backpropagation algorithm (machine learning)? If this is the case, then where does the visual cortex get its error signals from? To make the last question clear, in machine learning the typical approach is to compute the gradient of a loss function which compares the model's prediction with the ground truth, and the model's parameters are updated by moving the parameters in the direction of the gradient. If the visual cortex is learning advanced visual skills by virtue of a similar learning mechanism, then what kind of loss function is the visual cortex optimizing?

  • $\begingroup$ This isn't on topic here (I am trying to find a better home for it) but what in the world gave you the idea that animals don't need this sort of training? That's precisely what games like this are for, for example. And that's one of the benefits of play in young mammals of all species. $\endgroup$ – terdon Apr 5 '19 at 14:54
  • 1
    $\begingroup$ I'm voting to close this question as off-topic because it does not appear to be about bioinformatics. Maybe biology.SE? $\endgroup$ – Kamil S Jaron Apr 5 '19 at 15:11
  • $\begingroup$ The question relates to developing better deep learning (neural network [NN]) machinery based on human/animal perception for image recognition (convolution is a classic NN method). This idea has allegedly succeeded before in NN. However, the empirical principles of why a DL NN approach works are often poorly understood, the OP will be performing vast numbers of differential equations, in contrast biological approaches are better understood, so I'm not sure this question is going anywhere in any case. $\endgroup$ – Michael Apr 7 '19 at 13:42
  • $\begingroup$ Just to finally add that the modern rule of thumb in image recognition is that the deeper the NN the better the result. So just increase the depth of the NN by ten-fold. $\endgroup$ – Michael Apr 7 '19 at 15:08
  • 1
    $\begingroup$ Also, it's super broad. It would be a decent thesis topic. $\endgroup$ – Kamil S Jaron Apr 8 '19 at 8:50

Browse other questions tagged or ask your own question.