2025.04.16 meeting notes

Recap

• plot: added action hist plotting during training
• infra: added stochastic hurst FBM environment
• experiment: tested $H\sim U[0.1,0.3]$ FBM environment
• infra: fixed up LSTM training code, now it works properly

TODO

• add forced linear liquidation
• pretrain model on $H=0.1$ then continue training on $H=0.2$

Remaining from last week

• make asimptotic plot without market bound and put the trained policy eval on it
• make asimptotic plot for a single hurst
• show side by side of H0.1 T1024 with and without clipped action space, that is [-10, 10] and [-1_000, 1_000]
• if the training works for small T then we can try to incrementally extend the time horizon
• when running parallell envs try to shuffle the starting point to cover more of the episode length
• train model on $T=128$ with $H=0.45,0.55$ with big num_steps_eval number (5_000)
• fix training for $H>0.5$
  • Liquidation test env continue
  • make a grid search where there are three axis: clip_coef, learning_rate, H = 0.1, 0.7
  • make an env that is a child of FBMEnv and has a price process of a linear function with random slope