A Spectral Learning Algorithm for Finite State Transducers thumbnail
Pause
Mute
Subtitles
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

A Spectral Learning Algorithm for Finite State Transducers

Published on Oct 03, 20112845 Views

Finite-State Transducers (FSTs) are a popular tool for modeling paired input-output sequences, and have numerous applications in real-world problems. Most training algorithms for learning FSTs rely

Related categories

Chapter list

A Spectral Learning Algorithm for Finite State Transducers00:00
Probabilistic Transducers00:13
Spectral Learning Probabilistic Transducers01:59
Outline03:14
Deriving Observable Operator Models03:26
Observable Operator Model Parameters05:05
A Learnable Set of Observable Operators (1)07:01
A Learnable Set of Observable Operators (2)08:00
Spectral Learning Algorithm09:33
PAC-Style Result10:46
Synthetic Experiments12:42
Transliteration Experiments14:41
Summary of Contributions17:08
Thank you17:41