Amir-massoud Farahmand

This study reveals SGD's advantages in robustness over adaptive training methods.

Addressing value overestimation and primacy bias to enhance agent performance.

New methods enhance speed and stability in value iteration.