Will Leela (Neural Network) play like human if she were trained like human?

Since the Era of Alpha Beta search engines in 20 years ago, people have been trying to play anticomputer chess.  Most popular strategy is blocking the centre and king side attack beyond to reach a position beyond the horizon of engine's search. Most popular opening strategies are

1. Stonewall 

2. Dutch

3. King Indian Attack/Defense

4. The Hippopotamus Defense

In this diagram, Stockfish and Leela have different strategies on this positon.

Stockfish will be fighting for space and mobility, so SF prefer 4. d4.

Leela will be fighting to chase the opponent king, so she dont care d4 and alllow black to push d4. 

So, Leela allowed black to push black pawn to d4, and so f4,g4 push. I cant tell which strategy is better at this stage but I am fascinated by the way she plays like human, she plays like me.( I personally played such kind of blocked strategies against stockfish multiple times)