ACL RD-TEC 1.0 Summarization of H05-1127
Paper Title:
LEARNING MIXED INITIATIVE DIALOG STRATEGIES BY USING REINFORCEMENT LEARNING ON BOTH CONVERSANTS
LEARNING MIXED INITIATIVE DIALOG STRATEGIES BY USING REINFORCEMENT LEARNING ON BOTH CONVERSANTS
Authors: Michael English and Peter Heeman
Primarily assigned technology terms:
- algorithm
- computational linguistics
- database
- dialog manager
- dialog system
- dialog systems
- greedy algorithm
- greedy approach
- human language
- human language technology
- language processing
- language technology
- learning
- learning algorithm
- learning approach
- learning approaches
- learning procedure
- learning process
- natural language processing
- on-policy monte carlo
- one learning
- parsing
- processing
- q-learning
- reading
- reasoning
- reinforcement learning
- training process
Other assigned terms:
- approach
- association for computational linguistics
- convergence
- corpora
- dialog
- evaluation function
- fact
- generation
- histogram
- human involvement
- human-human dialog
- knowledge
- language access
- learning problem
- learning strategy
- likelihood
- linear combination
- linguistics
- method
- natural language
- probabilities
- probability
- procedure
- process
- sentences
- signal
- solution quality
- spoken language
- style
- technique
- technology
- training
- training corpus
- transition probabilities
- user
- user satisfaction
