This work is part of AISC Reinforcement Learning workshop conducted in October 2019.
In the foregoing study, the performance of different deep learning architectures (RNN, Transformers, Attention LSTM, Weighted Attention) for Reinforcement Learning stock trading policies is compared. S&P500 stock data from 1 January 2014 to 31 August 2019 was used for training and validation of these models.