A 50-minute five-moment lesson plan maps each teaching moment to a specific interface affordance¶

Strategy · All strategies
Evidence · 1 claim (1 for) · 1 study (1 causal), q2 · 0 of 1 report an effect size · 1 claim rests on one study

Description¶

The tool ships with a 50-minute lesson plan structured around five moments — stepping through ignorance, the first goal, convergence, hyperparameter experiments, and reward design — 'each mapped to a specific interface affordance.' Moments use single-step mode, ×100 training, and editable rewards; the reward-design moment has students check pit-seeking behavior against the optimal policy to distinguish under-exploration from a broken objective.

Design Implications¶

Context¶

Requirements¶

• Requires the tool's single-step mode, ×100 training mode, editable reward values, and the Bellman panel.

Constraints¶

• Learning gains have not been measured; the authors deliberately deferred a classroom study with pre/post concept inventory and usability instruments.

Target Learners¶

• undergraduate machine-learning students

Target Learning Goals¶

• understanding why exploration is needed
• observing backward value propagation from the goal
• designing and diagnosing reward functions

Claims¶

• Hyperparameter sweeps reproduce the lesson's claims: γ = 0 never solves the task while ε = 1 still converges [+W]

Related Strategies¶

• Structuring Lessons
• Time Management In Instruction

Examples¶

-

Key Sources¶

• Jueng, Ekkachai. (2026). Q-Learning Lab: Teaching Reinforcement Learning Through Learner-Generated Trace Analysis. https://arxiv.org/abs/2607.10802. doi:10.48550/arxiv.2607.10802