General
Reinforcement learning with prediction-based rewards
OpenAI Blog0 views
We’ve developed Random Network Distillation (RND), a prediction-based method for encouraging reinforcement learning agents to explore their environments through curiosity, which for the first time exceeds average human performance on Montezuma’s Revenge.
Source:https://openai.com/index/reinforcement-learning-with-prediction-based-rewardsPreviousPlan online, learn offline: Efficient learning and exploration via model-based controlNextLearning complex goals with iterated amplification
Related posts
How to make almost anything
0Hacker News
BeST: Beta splatted tomography for direct reconstruction from coherent imaging measurements
0arXiv eess.IV
Joint Optimization of Antenna Partitioning and Distortion-Aware Precoding for Massive MIMO-OFDM with Peak Cancellation and In-Band Distortion Compensation
0arXiv eess.SP
Smoothed estimation of Wasserstein barycenters
0arXiv stat.TH
On Sparsity and Sub-Gaussianity in the Johnson-Lindenstrauss Lemma
0arXiv stat.TH
Applications of QR-based Vector-Valued Rational Approximation
0arXiv cs.NA