Course materials· Reinforcement Learning for Electrical Engineering

Reinforcement Learning for Electrical Engineering

English translation

Graduate course · Assessed through a design project without exams

  1. Week 1

    Course overview and experiment setup

    CPU baseline path, public environments, and experiment records

  2. Week 2

    Public environments and MDP/reward design

    Observations, actions, termination, and random/rule-based baseline policies

    Next activity
    Introduction to Reinforcement Learning
  3. Week 3

    Value-Based Reinforcement Learning

    Calculate targets and updates from transitions and compare policies

  4. Week 4

    DQN and public implementations

    Function approximation, replay, and comparison in the same environment

  5. Week 5

    DQN Variants and Replay Buffer Sampling

    Roles of Double, Dueling, and PER, and state visitation coverage

  6. Week 6

    Policy Gradient and Actor–Critic

    Return, baseline, and baseline control

  7. Week 7

    PPO and public benchmarks

    Comparison with matched evaluation groups and training budgets / design topic introduction

    Reference exercisePPO and public benchmarks
  8. Week 8

    SAC, Continuous Control, and Design Topic Confirmation

    Finalize the problem, baseline, and evaluation plan for an open topic or an open topic within a specified field

    Reference activitySAC and Continuous Control
  9. Week 9

    Sim-to-Real: Model Uncertainty and Action Delay

    Evaluate input delays and model changes

  10. Week 10

    Isaac Lab and Robot Simulation

    Isaac Lab path or CPU transition analysis

  11. Week 11

    Robot dataset analysis and VLA inputs/outputs

    Episode-level splits, inputs/outputs, and baseline model

  12. Week 12

    Robot Dataset Analysis and VLA Inputs/Outputs

    Complete course content / Check data, comparison baselines, and evaluation metrics

    Reference activitiesSim-to-Real: Model Uncertainty and Action Delay · Isaac Lab and Robot Simulation · Robot Dataset Analysis and VLA Inputs/Outputs
  13. Week 13

    Sim-to-Real: Model Uncertainty and Action Delay

    Baseline and reproduction records using the selected public environment and data

    Reference exercisesSAC and continuous control · Sim-to-Real: Model Uncertainty and Action Delay
  14. Week 14

    Robot Dataset Analysis and VLA Inputs/Outputs

    Validate the effects and limitations of one improvement

    Reference activitiesSim-to-Real: Model Uncertainty and Action Delay · Robot Dataset Analysis and VLA Inputs/Outputs
  15. Week 15

    Design project: present and organize results

    Runnable code, results, contributions, and limitations