Reinforcement Learning for Finance
summary
Covers applying Deep Q-Learning to finance with runnable notebooks (TensorFlow originals plus PyTorch notebooks and Google Colab links), trading and hedging examples using ~10 years of end‑of‑day data (≈252 days/year), a plain‑vanilla DQN (memory 200, batch 32, small gamma, epsilon decay) that reaches ~52.5% directional accuracy on a demo, and a simple GAN (two hidden layers generator/discriminator) that produces 25 synthetic series passing KS tests (p>0.05) used to augment training; code and reproducibility notes are provided.
AI-written summary · about
chapters (6)
- 0:00Reinforcement learning for finance
- 11:48Dynamic programming in finance
- 26:36RL fundamentals and environments
- 48:00Learning with finance examples
- 1:08:00Synthetic data for RL
- 1:22:17Testing synthetic data & Q&A
related talks
01h 31mWorkshop on developing Reinforcement Learning
01h 17mTraining intelligent game agents using deep reinforcement learning
23mComplex AI Forecasting Methods for Investments
49mReinforcement Learning: Demystifying the hype
33mBreakthroughs and the Future of (Deep) Reinforcement Learning
28mDeep Reinforcement Learning: theory, intuition, code
44mReinforcement learning: A gentle introduction
59mAI in Finance: Scope and Examples