. 2019 Nov 29;6:131. doi: 10.3389/frobt.2019.00131

Algorithm 1.

Learning algorithm of the integrated cognitive model. Please refer to Appendix for Equations A-1 and A-11.

observe w^O,w^M,w^R,w^w

set initial random value to z,z^O,z^M,z^R,s,λ

set w^wC←w^w

for i←0 to n−1 do

update parameter of lower MLDA z^O,z^M,z^R (See Equation 10)

send P_1C to top MLDA

send P_3C to BHMM

update parameter of top MLDA z (See Equation A-1)

send P_2C to lower MLDA

update parameter of BHMM s by using P_3C as an initial value (See Equation A-11)

send P_4C to lower MLDA

update word information w^wC (See Equation 11)

end for

send P₅ to REINFORCE

update parameter of REINFORCE λ (See Equations 6 and 7)