https://github.com/mila-iqia/conscious-planning

Implementation for paper "A Consciousness-Inspired Planning Agent for Model-Based Reinforcement Learning".
https://github.com/mila-iqia/conscious-planning

Last synced: 7 days ago
JSON representation

Implementation for paper "A Consciousness-Inspired Planning Agent for Model-Based Reinforcement Learning".

• Host: GitHub
• URL: https://github.com/mila-iqia/conscious-planning
• Owner: mila-iqia
• Created: 2021-03-07T01:04:10.000Z (over 3 years ago)
• Default Branch: main
• Last Pushed: 2023-10-05T17:28:14.000Z (about 1 year ago)
• Last Synced: 2024-04-28T01:46:39.617Z (7 months ago)
• Language: Python
• Homepage:
• Size: 1.22 MB
• Stars: 56
• Watchers: 7
• Forks: 11
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
### [A Consciousness-Inspired Planning Agent for Model-Based Reinforcement Learning](https://pwnerharry.github.io/a-step-towards-conscious-planning/)

_By Mingde "Harry" Zhao, Zhen Liu, Sitao Luan, Shuyuan Zhang, Doina Precup and Yoshua Bengio_

![](CP_Poster.png)

#### ([BLOGPOST](https://pwnerharry.github.io/a-step-towards-conscious-planning/))

#### **Install Dependencies**

```

pip install -r requirements.txt

```

#### **Reproducing Results**

CP

```

python run_distshift_randomized_mp.py --method DQN_CP --num_explorers 8 --ignore_model 0 --disable_bottleneck 0 --size_bottleneck 8

```

UP

```

python run_distshift_randomized_mp.py --method DQN_CP --num_explorers 8 --ignore_model 0 --disable_bottleneck 1

```

WM

```

python run_distshift_randomized_mp.py --method DQN_WM --num_explorers 8 --ignore_model 0 --disable_bottleneck 0 --size_bottleneck 8 --period_warmup 1000000

```

Dyna

```

python run_distshift_randomized_mp.py --prioritized_replay 0 --method DQN_Dyna --num_explorers 8 --ignore_model 0 --disable_bottleneck 0 --size_bottleneck 8 --learn_dyna_model 1

```

*Special thanks to my colleague and friend Safa Alver [@alversafa](https://github.com/alversafa) for pointing out that Dyna should not use prioritized buffer as it shouldn't prioritize on the errors generated by potentially inaccurate imagined transitions, as well as the runtime bugs surrounding this matter!*

Dyna*

```

python run_distshift_randomized_mp.py --method DQN_Dyna --num_explorers 8 --ignore_model 0 --disable_bottleneck 0 --size_bottleneck 8 --learn_dyna_model 0

```

NOSET

```

python run_distshift_randomized_mp.py --method DQN_NOSET --num_explorers 8 --ignore_model 0 --layers_model 2 --len_hidden 256

```

#### **Changing Settings**

Read run_distshift_randomized_mp.py!