2 ms·
An early (pre-Goodfellow) inventor of the idea, not mentioned in the article but recognised in some histories, including a note in Wikipedia was a non-ML resear
by mellosouls 7y ago
An early (pre-Goodfellow) inventor of the idea, not mentioned in the article but recognised in some histories, including a note in Wikipedia was a non-ML researcher who modestly (he acknowledges he didn't develop it beyond the idea) described his prior claim here:
https://stats.stackexchange.com/a/301280 https://stats.stackexchange.com/a/301280
- cma 7y agoAnd from the 1990, Schmidhuber: https://www.reddit.com/r/MachineLearning/comments/djju8a/d_jurgen_schmidhuber_really_had_gans_in_1990/ https://www.reddit.com/r/MachineLearning/comments/djju8a/d_j...
- mellosouls 7y agoYes - but I didn't mention that one as it's well known (referenced in the OP article) and contentious; part of a history of arguments which - whatever the priority truth - seem as much about big name egos and power cliques (on all sides) than a furtherance of knowledge.
- cma 7y agoYeah the article makes mention but I just noticed it didn't mention how far earlier it was, it just says Goodfellow wasn't the only one to pursue, which could mean Schmidhuber came afterwards.