Learning to Poke by Poking: Experiential Learning of Intuitive Physics
Explore this paper's citation graph
Summary
A novel approach based on deep neural networks is proposed for modeling the dynamics of robot's interactions directly from images, by jointly estimating forward and inverse models of dynamics.
- Type
- preprint
- Published
- 2016-06-23
- Cited by
- 618
- References
- 36
- Access
- Open access
- OpenAlex
- https://openalex.org/W2473208550
- Semantic Scholar
- https://api.semanticscholar.org/CorpusID:6026836
Keywords
Artificial intelligence, Computer science, Feature (linguistics), Construct (python library), Robot
References
- Embed to Control: A Locally Linear Latent Dynamics Model for Control from Raw Images
- Planning algorithms
- The scientist in the crib : minds, brains, and how children learn
- Anticipating the future by watching unlabeled video
- From Pixels to Torques: Policy Learning with Deep Dynamical Models
- Autonomous reinforcement learning on raw visual input data in a real world application
- A Planning Framework for Non-Prehensile Manipulation under Clutter and Uncertainty
- THE PERCEPTION OF CAUSALITY
- An internal model for sensorimotor integration.
- Forward Models: Supervised Learning with a Distal Teacher
- Push-manipulation of complex passive mobile objects using experimentally acquired motion models
- The Neuro Slot Car Racer: Reinforcement Learning in a Real World Setting
- The Development of Embodied Cognition: Six Lessons from Babies
- Learning to predict how rigid objects behave under simple manipulation
- Action-Conditional Video Prediction using Deep Networks in Atari Games
- Automatic learning of pushing strategy for delivery of irregular-shaped objects
- Human-level control through deep reinforcement learning
- ImageNet classification with deep convolutional neural networks
- Galileo: Perceiving Physical Object Properties by Integrating a Physics Engine with Deep Learning
- Supersizing self-supervision: Learning to grasp from 50K tries and 700 robot hours
Cited by
- The developing infant creates a curriculum for statistical learning
- Interactive Perception: Leveraging Action in Perception and Perception in Action
- SE3-nets: Learning rigid body motion using deep neural networks
- Learning to push by grasping: Using multiple tasks for effective learning
- Deep visual foresight for planning robot motion
- Towards Lifelong Self-Supervision: A Deep Learning Direction for Robotics
- Learning to Perform Physics Experiments via Deep Reinforcement Learning
- Learning to Learn for Global Optimization of Black Box Functions
- Reinforcement learning with temporal logic rewards
- See the Glass Half Full: Reasoning About Liquid Containers, Their Volume and Content
- Deep Reinforcement Learning: An Overview
- Deep-learning in Mobile Robotics - from Perception to Control Systems: A Survey on Why and Why not
- Combining self-supervised learning and imitation for vision-based rope manipulation
- Learning A Physical Long-term Predictor
- Deep predictive policy training using reinforcement learning
- Prediction and Control with Temporal Segment Models
- Information-theoretic Model Identification and Policy Search using Physics Engines with Application to Robotic Manipulation
- A probabilistic data-driven model for planar pushing
- Learning to fly by crashing
- Curiosity-Driven Exploration by Self-Supervised Prediction
Related papers
- Review of journal of construction of modern and contemporary times in China
- S201014 Experiential Learning of Making Artifacts for Local Elementary, Middle and High School Students
- The Generic Cabling System of Construct
- The origins and early development of the psychological contract construct
- Construct Maps as a Foundation for Standard Setting
- Thoughts on the Colour Construct Lectures
- A local economy centre as experiential learning