Neural-Guided Runtime Prediction of Planners for Improved Motion and Task Planning with Graph Neural Networks.
Simon OdenseKamal GuptaWilliam G. MacreadyPublished in: IROS (2022)
Keyphrases
- neural network
- planning problems
- ai planning
- network architecture
- motion planning
- planning domains
- heuristic search
- planning process
- prediction model
- automated planning
- planning systems
- space time
- motion prediction
- image sequences
- prediction accuracy
- classical planning
- optical flow
- motion analysis
- forward search
- neural computation
- neural model
- pattern recognition
- learning rules
- probabilistic planning
- associative memory
- multi layer perceptron
- prediction error
- optimal planning
- neural network ensemble
- human motion
- motion model
- domain independent
- international planning competition
- artificial neural networks
- multilayer perceptron
- motion estimation
- temporal planning
- plan generation
- camera motion
- back propagation
- neural learning
- derived predicates
- neural fuzzy
- genetic algorithm
- planning graph
- causal graph
- initial state
- graph theory
- neural network model
- random walk
- fuzzy logic
- spiking neural networks
- recurrent networks
- biologically inspired
- directed graph
- search space
- moving objects