Re:Transformers Are Graph Neural Networks
by treaheraher on Monday March 09, 2020 @07:50AM
(#22524631)
Sorry but I hate it. Graph->graph transform can describe almost any operation, it's a superset of linear transforms. When all you know about your model is that 'it's a graph', you barely know anything at all. And transformers are just glorified AST parsers.
[ Parent
| on HN ]
> And transformers are just glorified AST parsers.
Can you explain this?
I am guessing he means that Transformers 'discover' parse trees in sentences, as they operate on fully-connected graphs but end up focusing on the most important connections through attention.