Hard Fun

NTU Weekly Progress Report 20200217

Byron, 09 March 2020

Completion (2020/02/17 - 2020/03/08)

• Courses Assigment
  • implement the A2C on the Atari Game (Pong and Boxing), and find some difference
• Other Material read:
  • Pathfinding
  • A* search algorithm
  • Introduction to A*

Ideas

Questions

Basic

• Q: Why the same (like A2C) algorithm can converage on certain game (Pong) but can not converage on another (Boxing). When we meet this problem, how can we analysis the reason for the problem and how can we improve this? From hyperparameter, or model structure, or the property relationship of the env and the algorithm?

Project related

Next Step

• Finish the course project for the AI Introduction (RL relatied)
• Finish the course project for multi-agent project
• Research about the Gym-soccer environment