Skip to content

Latest commit

 

History

History

curiosity

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
#Exploration A Study of Count-Based Exploration for Deep Reinforcement Learning.pdf
#Exploration A Study of Count-Based Exploration for Deep Reinforcement Learning.pdf
 
 
COBRA Data-Efficient Model-Based RL through Unsupervised Object Discovery and Curiosity-Driven Exploration.pdf
COBRA Data-Efficient Model-Based RL through Unsupervised Object Discovery and Curiosity-Driven Exploration.pdf
 
 
Curiosity-driven Exploration by Self-supervised Prediction.pdf
Curiosity-driven Exploration by Self-supervised Prediction.pdf
 
 
Dynamics-Aware Unsupervised Discovery of Skills.pdf
Dynamics-Aware Unsupervised Discovery of Skills.pdf
 
 
EPISODIC CURIOSITY THROUGH REACHABILITY.pdf
EPISODIC CURIOSITY THROUGH REACHABILITY.pdf
 
 
EXPLORATION BY RANDOM NETWORK DISTILLATION.pdf
EXPLORATION BY RANDOM NETWORK DISTILLATION.pdf
 
 
First return, then explore.pdf
First return, then explore.pdf
 
 
Go-Explore a New Approach for Hard-Exploration.pdf
Go-Explore a New Approach for Hard-Exploration.pdf
 
 
Large-Scale Study of Curiosity-Driven Learning.pdf
Large-Scale Study of Curiosity-Driven Learning.pdf
 
 
META-LEARNING CURIOSITY ALGORITHMS.pdf
META-LEARNING CURIOSITY ALGORITHMS.pdf
 
 
Noisy Networks for Exploration.pdf
Noisy Networks for Exploration.pdf
 
 
Self-Supervised Exploration via Disagreement.pdf
Self-Supervised Exploration via Disagreement.pdf
 
 
Unifying Count-Based Exploration and Intrinsic Motivation.pdf
Unifying Count-Based Exploration and Intrinsic Motivation.pdf
 
 
VIME Variational Information Maximizing Exploration.pdf
VIME Variational Information Maximizing Exploration.pdf
 
 
readme.md
readme.md
 
 

readme.md

Curiosity

Curiosity-driven Exploration by Self-supervised Prediction

ICML'17

Citation 1007

PDF HIGHLIGHT

The paper proposes ICM. The pioneer of curiosity.

EXPLORATION BY RANDOM NETWORK DISTILLATION

ICLR'19

Citation 235

PDF HIGHLIGHT

Go-Explore: a New Approach for Hard-Exploration Problems

Citation 155

PDF HIGHLIGHT

The paper is aim to solving hard-exploration problem. The agent would go to uncertaint state and explore directly.

First return then explore

PDF HIGHLIGHT

Github Code

Large-Scale Study of Curiosity-Driven Learning

ICLR'19

Citation 283

PDF HIGHLIGHT

The paper find out an big issue of intrinsic curiosity: the noisy TV problem.

NOISY NETWORKS FOR EXPLORATION

ICLR'18

Citation 406

PDF HIGHLIGHT

DYNAMICS-AWARE UNSUPERVISED DISCOVERY OF SKILLS

ICLR'20

PDF HIGHLIGHT

Episodic Curiosity through Reachability

ICLR'19

Citation 96

PDF HIGHLIGHT

Self-Supervised Exploration via Disagreement

ICML'19

PDF HIGHLIGHT

Github Code

Reference