Share your thoughts, 1 month free Claude Pro on usSee more

Reinforcement Learning on the Room (test)

128Average Total Reward per Episode

Episodic-only

Updated 15d ago

Evaluation Results

Method	Links
Episodic-only 2022.12		128
Both, RL, semantic-pretrained 2022.12		123.8
Both, RL, semantic-scratch 2022.12		117.9
Both, RL, semantic-pretrained 2022.12		116.5
Both, RL, semantic-scratch 2022.12		109.1
Semantic-only 2022.12		92.2
Semantic-only 2022.12		91.6
Episodic-only 2022.12		90.6
Both, RL, semantic-pretrained 2022.12		86.8
Both, RL, semantic-scratch 2022.12		85.3
Semantic-only 2022.12		84.2
Both, random 2022.12		82.6
Both, random 2022.12		76
Semantic-only 2022.12		61.6
Both, RL, semantic-scratch 2022.12		60.4
Both, RL, semantic-pretrained 2022.12		59.2
Episodic-only 2022.12		56.8
Both, random 2022.12		54.6
Both, random 2022.12		39.4
Both, RL, semantic-pretrained 2022.12		38.2
Both, RL, semantic-scratch 2022.12		37.6
Semantic-only 2022.12		36.8
Episodic-only 2022.12		35
Both, RL, semantic-pretrained 2022.12		25.5
Both, random 2022.12		25
Both, RL, semantic-scratch 2022.12		23.9
Episodic-only 2022.12		23.6
Semantic-only 2022.12		22.4
Episodic-only 2022.12		17.4
Both, random 2022.12		15.4