Goal: Build an automated poker player which learns via machine learning and is capable of playing on websites
Purpose: Learning on selenium, OCR, and ML
Steps:
1. Build automated poker player for a single site without ML
- This will use python and selenium to perform actions on the site
- Understanding the current context (e.g. what cards you have) will be done via reading the
HTML Canvas element as an image file, then performing OCR on it
- Decision engine will just be hardcoded, no ML at this point
2. Change decision engine from hardcoded rules to ML
- TensorFlow will be used. Further details TBD.