/sci/ - Science & Math

[Anonymous No.11416783]

Where is the field of artificial intelligence going after machine learning runs it's course?

[Anonymous No.11416832]

Deep learning

[Anonymous No.11416853]

>>11416832
and then world domination

[OP No.11416896]

>>11416783
>>11416832
>>11416853
Machine learning -> deep learning -> worlddomination -> quantumcomputing machine learning -> quantum computing deep learning

[Anonymous No.11417374]

>>11416783
At whatever point Douglas Hofstadter is now

[Anonymous No.11417429]

>>11416783
another ai winter, and relatively soon.
the truth is that human brain has astronomical computational capacity and computers are orders of magnitude too weak. The current dl algorithms are a variation of algorithms from 1980s, except computers are about 1M times faster per watt now which is what made them work.
At some point artificial computing is going to be powerful enough for agi to be possible and another ai boom is going to happen (probably the last one), but I can't say what increases are needed.

[Anonymous No.11417475]

>>11417429
What nonsense. Many animals that perform human-equivalent tasks not matched by anns have neuron counts in the order of what we do in anns right now, and natural neural networks work on the few hz timescale whereas computers work on the ghz timescale.

[Anonymous No.11417481]

>>11416783
> runs it's course
I mean that's kind of asking "what's at the end of number theory?" It's not happening for the foreseeable future, and it keeps branching out, which is sort of insinuated in >>11416581. Either way, muh AI and muh singularity are memes - there are far more interesting results about exactly *what* learning is, aside from models.

[Anonymous No.11417482]

>>11417374
Don't you mean: Jurgen Schmidhuber, well-known to have invented deep learning in the year 3083 BC?

[Anonymous No.11417487]

>>11416783
Contextual AI is next on the horizon.

[Anonymous No.11417491]

>>11417475
>have neuron counts in the order of what we do in anns right now
>thinking that biological neuron is equivalent to "neurons" in current ml

[Anonymous No.11417511]

>>11417475
read this:
https://bigthink.com/philip-perry/our-memory-comes-from-an-ancient-virus-neuroscientists-say
and realize neurons are dna computers that transport data between themselves by rna encapsulated in a capsid and store long-term data as dna in itself

[Anonymous No.11417645]

>>11417511
>read a bullshit opinion by a literal inbred moron
OK kid. I'll stick to science but hey, you do you.
>>11417491
>he thinks a hertz is the same as a gigahertz
lol

[Ironymous No.11417723]

>>11417511
Wow anon, you're so clever.
Can you tell me more about how DNA stores long-term memory?

[Anonymous No.11417835]

>>11417429
Not really, machines definitely have more computational capacity, it's just that it's not all there is to it to be able to perform cognitive functions on human level.