search

danaugrs / huskarl

Deep Reinforcement Learning Framework + Algorithms
https://medium.com/@tensorflow/introducing-huskarl-the-modular-deep-reinforcement-learning-framework-e47d4b228dd3
MIT License
414 stars 51 forks source link

Curiosity-Driven Exploration #3

Open Oktai15 opened 5 years ago

Oktai15 commented 5 years ago

What is SOTA in Curiosity-Driven Exploration currently? Will you plan implement that or simple version?

danaugrs commented 5 years ago

The plan is to implement the version described in Curiosity-driven Exploration by Self-supervised Prediction. Not sure how much further the SOTA is regarding curiosity - let me know if you find anything interesting!

Oktai15 commented 5 years ago

https://arxiv.org/abs/1810.12894 And maybe this can be interesting too: https://eng.uber.com/go-explore/