What allergy to innateness? Isn't basically all of computer vision now based on pre-trained ResNets, with *innate* low-level feature detectors?
until AI gets over its allergy to innateness, it’s not likely to get very far.https://twitter.com/TonyZador/status/1193176924497096706 …
-
-
-
read my january 2018 arxiv on innateness
- 10 more replies
New conversation -
-
-
artificial life / evolutionary computing has no allergy to innateness
-
I only half agree; even there, there is for the most part an allergy to build anything that looks much like cognitive level representation or algorithm.
- 15 more replies
New conversation -
-
-
I don’t think we should use the term of innateness in opposition to learning (in machine learning). ML algorithm do not « learn » like we learn. In some sense ML is just a way to adjust the weights of some function that could typically be hard wired in humans. ML ain’t learning
-
Different systems have different learning mechanisms. Bricks don't have any innate learning mechanisms, and learn nothing... ML is a field with many different learning mechanisms, none of which capture the full complement of what humans do when they learn.
End of conversation
New conversation -
-
-
The problem is, even innateness in an animal includes a great deal of plasticity. It is very wibbly.
-
i think this is a category error: the proper distinction, as i argued in the birth of the mind, is between prewiring and rewiring. those two processes complement each other.
- 1 more reply
New conversation -
-
-
innateness = inductive biases and computational principles in the learning architecture before any fitting?
-
sure, but that can encompass some stuff far from what he have been discussing, eg the algorithms behind Microsoft Excel's flash fill.
- 1 more reply
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.