Applying Many-to-Many Alignments and Hidden Markov Models to Letter-to-Phoneme Conversion

229 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Sequence-to-sequence neural net models for grapheme-to-phoneme conversion (2015, background): “…input and target-side output sequences and an alignment A, the posterior probability of output sequence given the input sequence is p(φT1 |A, lT1 ) = T∏ t=1 p(φt|φt−11 , l t 1) (4) where the current phoneme prediction φt depends both on its past prediction φt−1 and the input letter sequence lt.”
Combining Word-Level and Character-Level Models for Machine Translation Between Closely-Related Languages (2012, methodology): “The alignment models used in the transliteration literature have the same problem as they are usually based on edit distance operations and ﬁnite-state automata without contextual history (Jiampojamarn et al., 2007; Damper et al., 2005; Ristad and Yiani-los, 1998).”
Character-Based Pivot Translation for Under-Resourced Languages and Domains (2012, methodology): “Weighted finite state transducers (WFST’s) can be trained on unaligned pairs of character sequences and have been shown to be very effective for transliteration tasks or letter-to-phoneme conversions (Jiampojamarn et al., 2007).”