3 ms·
I tried training small RNN models using GA around 1990. I had lunch with John Koza (the genetic programming pioneer) and he suggested that it was an interesting
by mark_l_watson 8y ago
I tried training small RNN models using GA around 1990. I had lunch with John Koza (the genetic programming pioneer) and he suggested that it was an interesting idea but would not scale. The Uber team, my controlling mutation, got it to scale - good for them. I did a few years later use this as an example in my book "C++ Power Paradigms", McGraw-Hill 1994 (Genetic Algorithms, Neural Networks, and Constraint Programming).