Learning to imitate stochastic time series in a compositional way by chaos (非線形問題)
スポンサーリンク
概要
- 論文の詳細を見る
This study shows that a mixture of RNN experts model can acquire the ability to generate sequences that are combination of multiple primitive patterns by means of self-organizing chaos. By training of the model, each expert learns a primitive sequence pattern, and a gating network learns to imitate stochastic switching of the multiple primitives via a chaotic dynamics, utilizing a sensitive dependence on initial conditions. As a demonstration, we present a numerical simulation in which the model learns Markov chain switching among some Lissajous curves by a chaotic dynamics. Our analysis shows that a self-organized chaotic system can reconstruct the probability of primitive switching as observed in the training data.
- 社団法人電子情報通信学会の論文
- 2009-07-06
著者
-
TANI Jun
Brain Science Institute, RIKEN
-
Tani Jun
Brain Science Institute Riken
-
NAMIKAWA Jun
Brain Science Institute, RIKEN
-
Tani Jun
Brain Science Institute (riken)
-
Namikawa Jun
Brain Science Institute Riken
関連論文
- Predicting Object Dynamics From Visual Images Through Active Sensing Experiences
- Experience-based imitation using RNNPB
- Acquisition of Motion Primitives of Robot in Human-Navigation Task : Towards Human-Robot Interaction based on ``Quasi-Symbols
- Open-end human-robot interaction from the dynamical systems perspective : mutual adaptation and incremental learning
- Reinforcement learning of a continuous motor sequence with hidden states
- Dynamic and interactive generation of object handling behaviors by a small humanoid robot using a dynamic neural network model
- Self-organization of Dynamic Object Features Based on Bidirectional Training
- Learning to imitate stochastic time series in a compositional way by chaos (ニューロコンピューティング)
- Learning to imitate stochastic time series in a compositional way by chaos (非線形問題)
- Learning to generate combinatorial action sequences utilizing the initial sensitivity of deterministic dynamical systems
- Self-organization of distributedly represented multiple behavior schemata in a mirror system : reviews of robot experiments using RNNPB
- Learning to generate articulated behavior through the bottom-up and the top-down interaction processes
- Towards Written Text Recognition Based on Handwriting Experiences Using a Recurrent Neural Network
- Acquisition of Motion Primitives of Robot in Human-Navigation Task