The deepest problem with deep learning is its utter inability to tell correlations from causation. Gary Marcus's article (https://tinyurl.com/6vnpd8xu) is rather insightful, and comparing it with Francis Crick's similar effort (https://tinyurl.com/54y3975f) from 1989 immediately highlights the lack of academic and scientific integrity in corporate backed research. Notice that the two articles are separated by almost 30 years, yet the problems remain the same.They don't care about finding things out. They just want to sell you crap...
Add comment