###########################

LatticeWordSegmentation

###########################

Software to apply unsupervised word segmentation on lattices or text sequences using a nested hierarchical Pitman Yor language model

###########

Contact

###########

In case of questions, suggestions, problems etc. please send an email or check the disussion group.

Oliver Walter: [email protected]

Discussion group: email: [email protected] google groups: https://groups.google.com/d/forum/latticewordsegmentation

##############

References

##############

Iterative Bayesian Word Segmentation for Unspuervised Vocabulary Discovery from Phoneme Lattices Jahn Heymann, Oliver Walter, Reinhold Haeb-Umbach, Bhiksha Raj In 39th International Conference on Acoustics, Speech and Signal Processing (ICASSP 2014)

Unsupervised Word Segmentation from Noisy Input Jahn Heymann, Oliver Walter, Reinhold Haeb-Umbach, Bhiksha Raj In Automatic Speech Recognition and Understanding Workshop (ASRU 2013)

Graham Neubig, Masato Mimura, Shinsuke Mori, Tatsuya Kawahara "Learning a Language Model from Continuous Speech" In proceedings for InterSpeech 2010

######################

Manual Instalation

######################

Import project into kdevelop (or other IDE) Set cmake build path to $GITROOT/build/ (next to src/ and test/ directories) Install openFST from http://www.openfst.org/twiki/bin/view/FST/FstDownload Required boost packages: boost_system, boost_filesystem

Note: For more performace use release (-O3 -DNDEBUG) build!

#########################

Automatic instalation

#########################

run install.sh, this will also install boost and openfst in the tools directory

############

Examples

############

For demonstations see the scripts in the test/ folder