2 papers
cs.LG2018
Better Safe than Sorry: Evidence Accumulation Allows for Safe Reinforcement Learning
Akshat Agarwal, Abhinau Kumar, Kyle Dunovan +3
In the real world, agents often have to operate in situations with incomplete information, limited sensing capabilities, and inherently stochastic environments, making individual o…
cs.AI2018
Combining imagination and heuristics to learn strategies that generalize
Erik J Peterson, Necati Alp Müyesser, Timothy Verstynen +1
Deep reinforcement learning can match or exceed human performance in stable contexts, but with minor changes to the environment artificial networks, unlike humans, often cannot ada…