EasyDeepLearn

Population-Based Training (PBT) for RL.

hard

Answer

  • Jaderberg et al. 2017: run N agents in parallel with different hyperparameters.
  • Periodically: worst agents copy weights from top-N + perturb hyperparameters.
  • Simultaneously trains agents AND does hyperparameter optimization — no manual sweep.
  • Standard in complex RL (DeepMind's Capture-the-Flag, AlphaStar).
  • Modern alternative: Bayesian optimization + parallel runs.
Check yourself — multiple choice
  • Same as grid search
  • N parallel agents with different hyperparameters; periodically bottom agents copy top weights + perturb hyperparams; simultaneous training + HPO; used in AlphaStar / CTF
  • Random
  • Not real

PBT: parallel training + weight-copy + HPO; AlphaStar / CTF standard.

#engineering

Practise Reinforcement Learning

214 interview questions in this topic.

Related questions