Goal-oriented behavior sequence generation based on semantic commands using multiple timescales recurrent neural network with initial state correction.
Sungmoon JeongYunjung ParkRammohan MallipeddiJun TaniMinho LeePublished in: Neurocomputing (2014)
Keyphrases
- goal oriented
- recurrent neural networks
- initial state
- neural network
- complex valued
- feed forward
- requirements analysis
- recurrent networks
- situation calculus
- artificial intelligence
- goal state
- action sequences
- artificial neural networks
- expert systems
- decision making
- decision problems
- software engineering
- state space
- requirements engineering
- control system
- search space
- viterbi algorithm
- echo state networks