Difference between revisions of "Pereira and Riley, 1997"
| Line 10: | Line 10: | ||
They began by creating a speech recognition system using the composition of three finite state automata: a transducer A that converted acoustic signals to phone sequences, a transducer D that converts phone sequences to word sequences, and a finite state acceptor M that models a 5-gram model. | They began by creating a speech recognition system using the composition of three finite state automata: a transducer A that converted acoustic signals to phone sequences, a transducer D that converts phone sequences to word sequences, and a finite state acceptor M that models a 5-gram model. | ||
| − | One of the immediate problems was that their context independent phone model A | + | The biggest advantage of organizing the system of transducers this way is that they could apply properties about FSAs to the resulting computation. The calculation they want to perform is (A o D o M). While most would compose these together and then evaluation the probabilities of a new model, the computation can be organized differently. If (A o D o M) is too big for the computer's memory, one could instead calculate (A) then (A o D) then (A o D o M), pruning along the way. |
| + | |||
| + | One of the immediate problems was that their model A was context independent. This means that it assigned an acoustic signal to a phone with the same probability regardless of context. This turns out to be a very important consideration. Their solution was to make A a triphone model: one that would take an acoustic signal and output three phones. A problem then arose with D: D needed as input real phones, but what if A's output disagreed? For this, they inserted a new transducer, C, which converts from the context dependent parsing of A to the context independent of D. | ||
| + | |||
| + | ==Related Work== | ||
| + | Most of the ideas in their paper had previously been done by [[RelatedWork::Bahl, Jelinek, et al, 1983]]. The key insight this paper was that they used transducers to reduce the size of the resulting FSA (the so called "transduction cascade approach"). | ||
Revision as of 00:44, 3 November 2011
Speech Recognition by Composition of Weighted Finite Automata is a paper by Fernando Pereira and Michael Riley available online.
Contents
Citation
Pereira and Riley. Speech Recognition by Composition of Weighted Finite Automata. In Finite-State Language Processing, pages: 431-453, 1997.
Summary
They created a speech recognition system by using four finite state automata to model the entire pipeline of sound to phonemes to words. Since each step was fairly well understood, they wanted to use their knowledge in each step to improve the efficacy of the entire process. Unlike previous attempts, they didn't want to have to prune too early and leave out potentially fruitful matches. By using the composition of the various automata, they were able to search without pruning for some easier tasks.
Work
They began by creating a speech recognition system using the composition of three finite state automata: a transducer A that converted acoustic signals to phone sequences, a transducer D that converts phone sequences to word sequences, and a finite state acceptor M that models a 5-gram model.
The biggest advantage of organizing the system of transducers this way is that they could apply properties about FSAs to the resulting computation. The calculation they want to perform is (A o D o M). While most would compose these together and then evaluation the probabilities of a new model, the computation can be organized differently. If (A o D o M) is too big for the computer's memory, one could instead calculate (A) then (A o D) then (A o D o M), pruning along the way.
One of the immediate problems was that their model A was context independent. This means that it assigned an acoustic signal to a phone with the same probability regardless of context. This turns out to be a very important consideration. Their solution was to make A a triphone model: one that would take an acoustic signal and output three phones. A problem then arose with D: D needed as input real phones, but what if A's output disagreed? For this, they inserted a new transducer, C, which converts from the context dependent parsing of A to the context independent of D.
Related Work
Most of the ideas in their paper had previously been done by Bahl, Jelinek, et al, 1983. The key insight this paper was that they used transducers to reduce the size of the resulting FSA (the so called "transduction cascade approach").
