Curiosity driven reinforcement learning for motion planning on humanoids

@article{Frank2014CuriosityDR,
  title={Curiosity driven reinforcement learning for motion planning on humanoids},
  author={Mikhail Frank and J. Leitner and Marijn F. Stollenga and Alexander F{\"o}rster and J{\"u}rgen Schmidhuber},
  journal={Frontiers in Neurorobotics},
  year={2014},
  volume={7}
}
Most previous work on artificial curiosity (AC) and intrinsic motivation focuses on basic concepts and theory. [...] Key Method Our novel reinforcement learning (RL) framework consists of a state-of-the-art, low-level, reactive control layer, which controls the iCub while respecting constraints, and a high-level curious agent, which explores the iCub's state-action space through information gain maximization, learning a world model from experience, controlling the actual iCub hardware in real-time. To the best…Expand
Curiosity Driven Deep Reinforcement Learning for Motion Planning in Multi-agent Environment
  • Gojko Perovic, Ning Li
  • Computer Science
  • 2019 IEEE International Conference on Robotics and Biomimetics (ROBIO)
  • 2019
TLDR
A Deep Reinforcement Learning approach for a continuous multi-agent setting that is robust enough to handle high-level task accomplishment and low-level collision avoidance control is proposed and effectiveness is validated when compared with the state-of-the-art Reinforcement learning algorithm. Expand
Curiosity based Self-Organization of Humanoid Robot
This thesis presents a novel approach to how a high dimensional humanoid robot of 18 dimensions can learn within a few hours to control its body so that it is able to perform simple tasks such asExpand
Fast online model learning for controlling complex real-world robots
TLDR
The main contribution of this thesis is to provide a framework for solving a control learning problem, previously largely unexplored with no obvious solutions, but with strong analogies to, for example, early learning of body orientation control in infants. Expand
From learning to new goal generation in a bioinspired robotic setup
TLDR
The Intentional Distributed Robotic Architecture (IDRA) here presented intends to allow the autonomous development of new goals in situated agents starting from some simple hard-coded instincts. Expand
Artificial Curiosity Driven Robots with Spatiotemporal Regularity Discovery Ability
TLDR
This paper presents a method to solve the problem of autonomous reinforcement learning by extending ACR and shows that the proposed algorithm can discover the regularity in its environment without any prior information on the task. Expand
Online-learning and planning in high dimensions with finite element goal babbling
  • Pontus Loviken, Nikolas J. Hemion
  • Computer Science
  • 2017 Joint IEEE International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob)
  • 2017
TLDR
A new method, Finite Element Goal Babbling (FEGB), that presents a natural extension to GB by partitioning the sensor-space into a disjoint set of finite elements where every element is seen as an independent GB problem, a planning module can be added by observing transitions between the different elements. Expand
Learning to Play with Intrinsically-Motivated Self-Aware Agents
TLDR
This work proposes a "world-model" network that learns to predict the dynamic consequences of the agent's actions, and demonstrates that this policy causes the agent to explore novel and informative interactions with its environment, leading to the generation of a spectrum of complex behaviors. Expand
A Fuzzy Curiosity-Driven Mechanism for Multi-Agent Reinforcement Learning
TLDR
Experimental results show that multi-agent learns joint policies more efficiently by the proposed fuzzy curiosity-driven mechanism, and it can also help agents find better policies in the training process. Expand
Surprise-based learning of state representations
TLDR
It is proved that deterministic environments and a useful subclass of POMDP environments can be perfectly represented with equivalent compactness by such models and provide an active algorithm for autonomously learning such models in unknown environments from experience based on the biologically-inspired notion of surprises. Expand
Survey of Model-Based Reinforcement Learning: Applications on Robotics
TLDR
It is argued that, by employing model-based reinforcement learning, the—now limited—adaptability characteristics of robotic systems can be expanded, and model- based reinforcement learning exhibits advantages that makes it more applicable to real life use-cases compared to model-free methods. Expand
...
1
2
3
4
5
...

References

SHOWING 1-10 OF 63 REFERENCES
Artificial curiosity with planning for autonomous perceptual and cognitive development
TLDR
A novel artificial curiosity system with planning is implemented, based on developmental or continual learning principles, to efficiently assign values for maximizing combined external and intrinsic reward. Expand
A possibility for implementing curiosity and boredom in model-building neural controllers
TLDR
It is described how the particular algorithm (as well as similar model-building algorithms) may be augmented by dynamic curiosity and boredom in a natural manner by introducing (delayed) reinforcement for actions that increase the model network's knowledge about the world. Expand
Reinforcement active learning hierarchical loops
TLDR
A model of hierarchical motor-sensory loops for an autonomous active learning agent, meaning a model that selects the appropriate action in order to optimize the agent's learning, is presented. Expand
Task-relevant roadmaps: A framework for humanoid motion planning
TLDR
Unlike Jacobian-based methods, this algorithm does not rely on calculation of gradients, making application of the algorithm much simpler, and it is shown how NGIK outperforms recent related sampling algorithms. Expand
The Modular Behavioral Environment for Humanoids and other Robots (MoBeE)
TLDR
MoBeE is developed, a novel behavioral framework for humanoids and other complex robots, which integrates elements from vision, planning, and control, facilitating the synthesis of autonomous, adaptive behaviors. Expand
Developmental robotics, optimal artificial curiosity, creativity, music, and the fine arts
TLDR
It is pointed out how the fine arts can be formally understood as a consequence of the basic principle: given some subjective observer, great works of art and music yield observation histories exhibiting more novel, previously unknown compressibility/regularity/predictability than lesser works, thus deepening the observer’s understanding of the world and what is possible in it. Expand
Learning Movement Primitives
TLDR
A novel reinforcement learning technique based on natural stochastic policy gradients allows a general approach of improving DMPs by trial and error learning with respect to almost arbitrary optimization criteria, and demonstrates the different ingredients of the DMP approach in various examples. Expand
Autonomous learning of abstractions using Curiosity-Driven Modular Incremental Slow Feature Analysis
TLDR
Experiments with a simulated iCub humanoid robot show how the proposed method effectively learns a set of abstractions from raw un-preprocessed video, to the knowledge the first curious learning agent to demonstrate this ability. Expand
Between MDPs and Semi-MDPs: A Framework for Temporal Abstraction in Reinforcement Learning
TLDR
It is shown that options enable temporally abstract knowledge and action to be included in the reinforcement learning frame- work in a natural and general way and may be used interchangeably with primitive actions in planning methods such as dynamic pro- gramming and in learning methodssuch as Q-learning. Expand
Learning tactile skills through curious exploration
TLDR
The approach demonstrates the feasibility of autonomous acquisition of tactile skills on physical robotic platforms through curiosity-driven reinforcement learning, overcomes typical difficulties of engineered solutions for active tactile exploration and underactuated control, and provides a basis for studying developmental learning through intrinsic motivation in robots. Expand
...
1
2
3
4
5
...