Multi-Task Deep Reinforcement Learning for Continuous Action Control
By Zhaoyang Yang, Kathryn Elizabeth Merrick, Hussein A. Abbass, Lianwen Jin. 2017
68 citations recorded by the source corpus.
In this paper, we propose a deep reinforcement learning algorithm to learn multiple tasks concurrently. A new network architecture is proposed in the algorithm which reduces the number of parameters needed by more than 75% per task compared to typical single-task deep reinforcement learning algorithms. The proposed algorithm and network fuse images with sensor data and were tested with up to 12 movement-based control tasks on a simulated Pioneer 3AT robot equipped with a camera and range sensors. Results show that the proposed algorithm and network can learn skills that are as good as the skills learned by a comparable single-task learning algorithm. Results also show that learning performance is consistent even when the number of tasks and the number of constraints on the tasks increased.
Subfields: Artificial Intelligence, Biomedical Engineering, Control and Systems Engineering
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Eric Eaton
- Sergey Levine
- Pieter Abbeel
- Sanja Fidler
- Jorge Armando Mendez Mendez
- Yoshua Bengio
- Mohit Bansal
- Chelsea Finn
In Biomedical Engineering
- Marcello Restelli
- Eric Eaton
- Jimmy Ba
- Sergey Levine
- Jorge Armando Mendez Mendez
- Jan Peters
- Tingwu Wang
- Sanja Fidler
In Control and Systems Engineering
- Marcello Restelli
- Sergey Levine
- Jimmy Ba
- Eric Eaton
- Pieter Abbeel
- Jorge Armando Mendez Mendez
- Carlo D'Eramo
- Jan Peters
Citations in the corpus
- Lifelong Policy Gradient Learning of Factored Policies for Faster Training Without Forgetting (2020, background): “Meta-RL [9, 11, 15, 8] and multi-task RL [27, 36, 40, 43] also seek to accelerate learning by reusing information from different tasks, but in those settings the agent does not handle tasks arriving sequentially and the consequent problem of catastrophic forgetting.”
- Sharing Knowledge in Multi-Task Deep Reinforcement Learning (2020, background): “Further developments, extend the analysis to policy search (Yang et al., 2017; Teh et al., 2017), and to multi-goal RL (Schaul et al., 2015; Andrychowicz et al., 2017).”
- Sharing Experience in Multitask Reinforcement Learning (2019, methodology): “The authors in [Yang et al., 2017] propose a multi-actor and single-critic architecture to implicitly exploit the shared visual features to tackle multi-task learning.”
- NerveNet: Learning Structured Policy with Graph Neural Networks (2018, background): “In Yang et al. (2017); Parisotto et al. (2015), the authors exploited shared visual features to tackle multi-task learning.”