Autonomous Learning of State Representations for Control: An Emerging Field Aims to Autonomously Learn State Representations for Reinforcement Learning Agents from Their Real-World Sensor Observations

By Wendelin Böhmer, Jost Tobias Springenberg, Joschka Boedecker, Martin A. Riedmiller, Klaus Obermayer. 2015

70 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Cognitive Neuroscience, Statistical and Nonlinear Physics, Reinforcement Learning, Robotics & Control

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

In Cognitive Neuroscience

In Statistical and Nonlinear Physics

In Reinforcement Learning

In Robotics & Control

Citations in the corpus

Embed to Control: A Locally Linear Latent Dynamics Model for Control from Raw Images (2015, neutral): “[17] for a review), deep autoencoders (ignoring state transitions) similar to our baseline models have been applied previously, e.”
Deep Reinforcement Learning for Autonomous Driving: A Survey (2020, background): “it should encode all necessary information to be able to select an action based on the current state only, and not any previous states or actions [60], [62].”
Deep spatial autoencoders for visuomotor learning (2015, background): “Prior work has proposed to learn representations using general priors from the physical world [2] or the ability to predict future observations [18], [19], [20].”
Deep-learning in Mobile Robotics - from Perception to Control Systems: A Survey on Why and Why not (2016, background): “The observations of DQN are the raw images from the screen display of Atari 2600 games.”
Learning state representation for deep actor-critic control (2016, background): “However, end-toend learning requires a large amount of data [13], which is not always available.”
Learning Visual Feature Spaces for Robotic Manipulation with Deep Spatial Autoencoders (2015, background): “Prior work has proposed to learn representations using general priors from the physical world [2] or the ability to predict future observations [18], [19], [20].”
Decoupling feature extraction from policy learning: assessing benefits of state representation learning in goal based robotics (2018, methodology): “Hence, this process can be automated using methods that are able to learn these features (also called representations) (Böhmer et al., 2015; Singh et al., 2012; Boots et al., 2011).”
An Overview of Natural Language State Representation for Reinforcement Learning (2020, background): “Although it remains a common practice, feature engineering has evident drawbacks: it needs an experienced designer, is tedious and time-consuming and does not generalize (B¨ohmer et al., 2015).”
DREAM Architecture: a Developmental Approach to Open-Ended Learning in Robotics (2020, background): “This issue is addressed in a variety of ways, such as learning a lower dimensionality representation [111, 53], or using learned models of the environment [112].”
Learning Real-World Robot Policies by Dreaming (2018, background): “See [15] and [16] for a more detailed review.”
Transfer Learning for User Adaptation in Spoken Dialogue Systems (2016, background): “A wide range of algorithms for learning a state representation exists [3, 9].”