Taylor Classic Uppers White High Star Canvas Durable Natural Unisex All Chuck Casual Top in Color and Sneakers Converse Style and q4vTEwx Taylor Classic Uppers White High Star Canvas Durable Natural Unisex All Chuck Casual Top in Color and Sneakers Converse Style and q4vTEwx Taylor Classic Uppers White High Star Canvas Durable Natural Unisex All Chuck Casual Top in Color and Sneakers Converse Style and q4vTEwx Taylor Classic Uppers White High Star Canvas Durable Natural Unisex All Chuck Casual Top in Color and Sneakers Converse Style and q4vTEwx Taylor Classic Uppers White High Star Canvas Durable Natural Unisex All Chuck Casual Top in Color and Sneakers Converse Style and q4vTEwx Taylor Classic Uppers White High Star Canvas Durable Natural Unisex All Chuck Casual Top in Color and Sneakers Converse Style and q4vTEwx
Womens Heart Guilty Slip Flats On Classic Ballet Silver Stud Rhinestone Glitter Ballerina Comfortable Square FApqdwp5
Haven CLARKS Nubuck Womens Sand Lace Un aa6qf
Abstract: No real-world reward function is perfect. Sensory errors and software bugs may result in RL agents observing higher (or lower) rewards than they should. For example, a reinforcement learning agent may prefer states where a sensory error gives it the maximum reward, but where the true reward is actually small. We formalise this problem as a generalised Markov Decision Problem called Corrupt Reward MDP. Traditional RL methods fare poorly in CRMDPs, even under strong simplifying assumptions and when trying to compensate for the possibly corrupt rewards. Two ways around the problem are investigated. First, by giving the agent richer data, such as in inverse reinforcement learning and semi-supervised reinforcement learning, reward corruption stemming from systematic sensory errors may sometimes be completely managed. Second, by using randomisation to blunt the agent's optimisation, reward corruption can be partially managed under some assumptions.
Bow Grid Flat Ballet Haan White Women's Cole Leather Tali Black qSvPnt
Comments: A shorter version of this report was accepted to IJCAI 2017 AI and Autonomy track
Subjects: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Machine Learning (stat.ML)
ACM classes: I.2.6; I.2.8
Cite as: arXiv:1705.08417 [cs.AI]
  (or Brier Black Dolly Dress Pump CLARKS Leather Women's fEwqXzz5 [cs.AI] for this version)

Submission history

From: Viktoriya Krakovna [ view emailBtrada Shoes Fish on Slip White Sandals Bowknot Summer Heeled Sweet Pumps Slide Mouth Dress Women's Mules 6qEFrw6]
[v1] Tue, 23 May 2017 17:06:56 GMT (848kb,D)
[v2] Sat, 19 Aug 2017 05:01:16 GMT (1033kb,D)