First-order-principles-based constructive network topologies: An application to robot inverse dynamics

Keyphrases

• First-order 100%
• Inverse Dynamics 100%
• Physical Systems 100%
• Generalization Ability 100%
• Principles-based 100%
• Network Topology 100%
• Constructive Neural Networks 100%
• Manipulator 50%
• Neural Network 50%
• Robot Dynamics 50%
• Model Accuracy 50%
• Order of Magnitude 50%
• Articulated Robot 50%
• Seven Degrees of Freedom 50%
• Building Elements 50%
• Parametrized 50%
• Feedforward Neural Network 50%
• Number of Parameters 50%
• Interdependency 50%
• Activation Function 50%
• Model Generalization 50%
• Node number 50%
• Deep Recurrent Neural Network 50%
• System Knowledge 50%
• Operational Network 50%
• Input-output Mapping 50%
• Learning Speed 50%
• Model Precision 50%
• Generalization Power 50%
• Node Function 50%
• Unconventional Feed 50%
• Accurately Model 50%
• Building Connection 50%

Computer Science

• Robotics 100%
• Network Topology 100%
• Physical System 100%
• Robot 50%
• Manipulator 50%
• Neural Network 50%
• Model Accuracy 50%
• Input/Output 50%
• Reported Result 50%
• Knowledge-Based System 50%
• Output Mapping 50%
• Layer Function 50%
• Feed Forward Neural Networks 50%
• Deep Recurrent Neural Network 50%
• Model 50%
• Interdependency 50%
• Network Operator 50%
• Activation Function 50%

Engineering

• Physical System 100%
• Network Topology 100%
• Robotics 100%
• Degree of Freedom 50%
• Feedforward 50%
• Nodes 50%
• Building Element 50%
• Robot 50%
• Manipulator 50%
• Model Input 50%
• Interdependency 50%
• Network Operator 50%
• Activation Function 50%
• Building Connection 50%
• Recurrent Neural Network 50%
• Model 50%