Theory and Application of Reward Shaping in Reinforcement Learning

164 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Reinforcement learning in robotics: A survey (2013, background): “to guide the learning process to a reasonable solution, a process known as reward shaping (Laud, 2004).”
Automated Curriculum Learning by Rewarding Temporally Rare Events (2018, background): “rds) are sparse and/or delayed. A popular way to overcome this issue is to shape the reward function with prior knowledge such that the agent receives additional rewards to guide its learning process [22, 23, 31]. Another approach is to gradually increase the difﬁculty of the environment to ease lea”
A Survey of Deep Learning Applications to Autonomous Vehicle Control (2019, background): “of tasks, intermediate rewards can be used to guide the agent towards desired behaviour, this process in known as reward shaping [175], [176].”
Dynamic Reward Shaping: Training a Robot by Voice (2010, methodology): “Reward shaping has been used to accelerate reinforcement learning [15,9].”
ActionSpotter: Deep Reinforcement Learning Framework for Temporal Action Spotting in Videos (2020, methodology): “According to reward shaping theory [27], this final reward can be hybridized with a potential-based shaping to help reinforcement convergence without changing optimal policy.”