Closed vadimnazarov closed 7 years ago
One question to consider here: l
th Calpha position is influenced by both l-1,l-2,...
and l+1,l+2,...
. We need to somehow apply symmetry here. Note that for CDR3 (but not CDR1 and AG) the x[0]==x[L-1] + fixed_value
, where L
is the length. This is known as Omega loop.
Reinforcement learning: https://www.nervanasys.com/demystifying-deep-reinforcement-learning/ https://github.com/matthiasplappert/keras-rl
meta-optimization ?