Spoken
Voice practice that levels with you.
Answer out loud. Reward-based Q-learning raises or eases difficulty as you go.
How the reward loop works
Each round picks easy, medium, or hard with epsilon-greedy Q-learning. Correct answers pay more at harder levels; misses cost more too. The table learns which difficulty fits you next.