Academy notes
Articles and updates from SoloQuant Academy.
Tag: rl
Liu et al. (2020) introduce FinRL: layered gym stock envs, fine-tuned DQN/DDPG/PPO/SAC/A2C/TD3, and Jupyter demos for one name, many names, and allocation. Plain-language intro. Not investment advice.
Liu, Yang, Gao and Wang (2021, ICAIF) present FinRL as a full-stack pipeline: gym + live APIs, plug-in SB3/RLlib/ElegantRL, train-test-trade, ensemble and crypto notebooks. Plain-language intro. Not investment advice.
Park, Sim and Choi (2019) train a DQN on discrete buy/hold/sell lots with a mapping for infeasible trades. LSTM encoder + all-feasible-action replay. Plain-language intro. Not investment advice.
Liu et al. (2022) ship FinRL-Meta: DataOps from vendor dumps to gym-style markets, paper reproductions, and a curriculum of notebooks. Plain-language intro. Not investment advice.
Zhang, Zohren and Roberts (2019) map TSMOM features to positions with DQN, PG and A2C, plus volatility scaling, on 50 liquid futures 2011–2019. Plain-language intro. Not investment advice.
Park, Sim and Choi (2019) train a DQN on discrete buy/hold/sell lots with a mapping for infeasible trades. LSTM encoder + all-feasible-action replay. Plain-language intro. Not investment advice.
