Buch, Englisch, 205 Seiten, PB, Format (B × H): 148 mm x 210 mm, Gewicht: 400 g
Reihe: Karlsruhe series on intelligent sensor-actuator-systems
Buch, Englisch, 205 Seiten, PB, Format (B × H): 148 mm x 210 mm, Gewicht: 400 g
Reihe: Karlsruhe series on intelligent sensor-actuator-systems
ISBN: 978-3-86644-569-7
Verlag: KIT Scientific Publishing
This book examines Gaussian processes in both model-based reinforcement learning (RL) and inference in nonlinear dynamic systems. First, we introduce PILCO, a fully Bayesian approach for efficient RL in continuous-valued state and action spaces when no expert knowledge is available. PILCO takes model uncertainties consistently into account during long-term planning to reduce model bias. Second, we propose principled algorithms for robust filtering and smoothing in GP dynamic systems.