This preview shows page 1. Sign up to view the full content.
Unformatted text preview: Q3: T he authors think that adv ance s in ne ural ne tworks prov ide e v ide nce against
re pre se ntational nativ ism. In one se nte nce , summariz e one of the le ssons that the y draw from
ne ural ne tworks. In a se cond se nte nce , e v aluate the ir claim. Do you agre e ? Why or why not?
Advanc es in neural networks have allowed for s ys tem s to be developed that c an learn c om plex artific ial
gram m ars in quite s im ilar ways to how c hildren learn language (e.g., rec overing from errors without
explic it negative feedbac k). I agree that advanc es in neural networks c an s hed light on how the gram m ar
that we us e c an be learned by a s ys tem that only utiliz es general learning princ iples and that this
repres ents a huge s tep forward our unders tanding of the m ec hanis m s of how gram m ar c ould aris e.
Publis hed by Google Drive – Report Abus e – Updated automatic ally every 5 minutes https://docs.g oog le.com/document/pub?id= 1G- 1P4QInsdHJhNsUnl0Gr i8XW87YimIwCr f3cKVQtyg 1/1...
View Full Document
This note was uploaded on 02/09/2014 for the course COGS 101c taught by Professor Staff during the Spring '08 term at UCSD.
- Spring '08