ACL RD-TEC 1.0 Summarization of P04-1008
Paper Title:
STATISTICAL MODELING FOR UNIT SELECTION IN SPEECH SYNTHESIS
STATISTICAL MODELING FOR UNIT SELECTION IN SPEECH SYNTHESIS
Authors: Mehryar Mohri and Cyril Allauzen and Michael Riley
Primarily assigned technology terms:
- algorithm
- approximation
- automatic speech recognition
- automaton
- backoff smoothing
- backoff smoothing technique
- beam search
- caching
- computing
- decoder
- difference method
- final state
- finite-state transducer
- finite-state transducers
- katz backoff smoothing
- language modeling
- learning
- learning algorithms
- matching
- modeling
- n-gram backoff
- principled learning
- processor
- ranking
- recognition
- search
- shortest-distance
- shortest-distance algorithm
- smoothing
- smoothing technique
- speech recognition
- speech synthesis
- speech synthesis system
- speech synthesizer
- statistical language modeling
- statistical methods
- statistical modeling
- statistical modeling framework
- synthesis
- synthesis system
- synthesizer
- text analysis
- transducer
- transducers
- unit selection
- viterbi
- viterbi decoder
- viterbi search
- weighted automata
- weighted difference method
- weighted finite-state transducer
- weighted finite-state transducers
- weighted transducer
Other assigned terms:
- acoustic signal
- alphabet
- approach
- automata
- backoff
- backoff model
- beam
- bigram
- bigram model
- cache
- case
- composition
- corpora
- distribution
- experimental results
- feature
- feature vector
- feature vectors
- finite automaton
- finite set
- grammar
- heuristic
- heuristics
- input string
- joint distribution
- joint probability
- joint probability distribution
- language model
- language models
- large corpus
- log-likelihood
- mapping
- method
- n-gram
- n-gram language model
- n-gram model
- opinion
- probability
- probability distribution
- probability estimate
- projection
- schema
- semiring
- signal
- standard deviation
- statistical framework
- statistical language model
- statistical model
- symbols
- synthesized speech
- technique
- text
- topics
- training
- training corpora
- training corpus
- training data
- trigram
- trigram language model
- utterance
- words
