Stabilizing Transformer-Based Action Sequence Generation For Q-Learning.

Gideon Stein Andrey Filchenkov Arip Asadulaev

Published in: CoRR (2020)

Keyphrases

action selection
action sequences
reinforcement learning
cooperative
state space
nonlinear systems
fuzzy logic
input data
generation process
state action
learning algorithm
function approximation
dynamic programming
stochastic approximation
multi agent reinforcement learning