Introduction to policy search in Reinforcement Learning Djalel Benbouzid — Data:Lab Munich, Volkswagen Group June 27 th 2018
introduction and some reminders introduce a di ff erent way to RL , wrt. first lecture quick outline gradients-based methods gradients-free methods
“traditional” function output machine programming data machine learning data function machine output (supervision)
Machine Learning paradigms 3 2 1 0 −1 −2 • Supervised Learning −3 2.5 2.0 20 1.5 5.0 1.0 2bservations 0.5 −1.5 −1.0 −0.5 0.0 0.0 0.5 1.0 1.5 −0.5 Prediction 15 4.5 4.0 • Unsupervised Learning 10 3.5 f ( x ) 5 3.0 0 2.5 • Reinforcement Learning −5 2.0 1.5 −10 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 0 2 4 6 8 10 x
what about deep learning? source: Bengio et. al
Recommend
More recommend