|
|
|
|
|
|
by Nomentatus
3017 days ago
|
|
|
Re 3, last resort: isn't it also true that any other method can likely be improved upon by taking whatever network results and using it as a starting point for genetic improvements? Sometimes the extra expense won't be worthwhile, of course, but if you want to reduce the number of edge cases for autonomous driving, say, it might well be worthwhile IMHO. (There's an earlier discussion about the myth of local optima here that might or might not answer my question.) |
|
|