MACE-RL is a novel reinforcement learning framework that learns to dynamically regulate its own exploration strategy. It integrates:

• A curiosity module informed by episodic memory
• A meta-learning network that adaptively regulates the curiosity bonus

This enables agents to balance exploration and exploitation dynamically.

pip install -r requirements.txt

python experiment_runner.py --quick-test

python experiment_runner.py --run-experiments

python experiment_runner.py --publication-mode

python visualize_agent.py --model-path path/to/model.pth --env CartPole-v1

mace_rl/
├── agents/         # RL agents (PPO)
├── modules/        # Core MACE-RL components
│   ├── curiosity.py        # Curiosity module
│   ├── episodic_memory.py  # Episodic memory
│   └── meta_adaptation.py  # Meta-learning network
├── envs/           # Environment wrappers
└── utils/          # Utilities and helpers

Metrics: Sample Efficiency / Convergence Steps (lower is better)