3 ms·

Is it really a feeling? What I find weird id how it's supposed to be defined in the genome.

by narag 7y ago

magicalhippo 7y ago
You can get pretty complex behavior almost entirely due to the structure of the neural net[1], and judging by the rest of our bodies, building non-trivial structures is something the genome is quite capable of. [1]: https://ai.googleblog.com/2019/08/exploring-weight-agnostic-neural.html https://ai.googleblog.com/2019/08/exploring-weight-agnostic-...
- narag 7y ago
  You're arguing that it's possible, I'm saying that I haven't seen any data showing genes defining behaviour in a way that can be called "prewired to perceive the world". Not really a conflicting statement, but neither to contradict my skepticism.
  - monocasa 7y ago
    The coordinate system(s) used by cell differentiation can collectively figure out body location down to a single cell. Using those inputs to define how the cell differentiates from a stem cell line is how nearly all the body differentiates, why would neurons be different? Beyond that, it's pretty clear that there are very hardwired parts of the brain around bootstrapping. You're born with the ability to recognize faces, and mirror neurons reflect the actions of those around you in a very similar way to you taking those actions yourself.
  - magicalhippo 7y ago
    Random example, baby kangaroos when born[1] know a) that they need to climb into the mothers pouch b) know _how_ to climb into the mothers pouch and c) know _where_ to climb in order to get into the mothers pouch. They start climbing immediately after being born, before the mother can teach them anything. What else than genes could be behind that? https://youtu.be/UpsnREY-6no?t=25 https://youtu.be/UpsnREY-6no?t=25
  - stanfordkid 7y ago
    Genes don’t directly create complex structures in the way you are thinking. Rather these patterns arise via emergent effects of proteins interacting and resisting each other across populations of cells.[1] So really what you might have is a certain neuronal cell type that sends signaling molecules that cause it to divide in a very particular way that could roughly create a structure similar to a face detecting filter. If cells can grow into bones or coral reefs, why not face detecting neurons? [1]https://www.hindawi.com/journals/ddns/2020/2308074/ https://www.hindawi.com/journals/ddns/2020/2308074/

Is it really a feeling? What I find weird id how it's supposed to be defined in the genome.

by narag 7y ago

Is it really a feeling? What I find weird id how it's supposed to be defined in the genome.

You can get pretty complex behavior almost entirely due to the structure of the neural net[1], and judging by the rest of our bodies, building non-trivial structures is something the genome is quite capable of. [1]: https://ai.googleblog.com/2019/08/exploring-weight-agnostic-neural.html https://ai.googleblog.com/2019/08/exploring-weight-agnostic-...

narag 7y ago
You're arguing that it's possible, I'm saying that I haven't seen any data showing genes defining behaviour in a way that can be called "prewired to perceive the world". Not really a conflicting statement, but neither to contradict my skepticism.
- monocasa 7y ago
  The coordinate system(s) used by cell differentiation can collectively figure out body location down to a single cell. Using those inputs to define how the cell differentiates from a stem cell line is how nearly all the body differentiates, why would neurons be different? Beyond that, it's pretty clear that there are very hardwired parts of the brain around bootstrapping. You're born with the ability to recognize faces, and mirror neurons reflect the actions of those around you in a very similar way to you taking those actions yourself.
- magicalhippo 7y ago
  Random example, baby kangaroos when born[1] know a) that they need to climb into the mothers pouch b) know _how_ to climb into the mothers pouch and c) know _where_ to climb in order to get into the mothers pouch. They start climbing immediately after being born, before the mother can teach them anything. What else than genes could be behind that? https://youtu.be/UpsnREY-6no?t=25 https://youtu.be/UpsnREY-6no?t=25
- stanfordkid 7y ago
  Genes don’t directly create complex structures in the way you are thinking. Rather these patterns arise via emergent effects of proteins interacting and resisting each other across populations of cells.[1] So really what you might have is a certain neuronal cell type that sends signaling molecules that cause it to divide in a very particular way that could roughly create a structure similar to a face detecting filter. If cells can grow into bones or coral reefs, why not face detecting neurons? [1]https://www.hindawi.com/journals/ddns/2020/2308074/ https://www.hindawi.com/journals/ddns/2020/2308074/