Francis Tyers <fty...@prompsit.com> writes: > Hello all, > > We had 11 projects accepted this year. Of those, 2 failed. Here is a > rundown of the remaining projects I was keeping an eye on: > > * apertium-id-ms: >80% coverage over Wikipedia, error rate of < 10%. > This pair has been moved to trunk and will be released shortly. > Congratulations Raymond and Tina for the excellent work. This is a > "canonical Apertium" pair. Testvoc clean in both directions. Comes close > to Google level for this pair. Further information here: > http://wiki.apertium.org/wiki/Indonesian_and_Malaysian/Work_plan > > * apertium-mt-ar: >80% coverage over Wikipedia, error rate around > 10-20%. This pair has been moved to staging, and a preliminary/beta > version will be released shortly. Excellent work by Miri and Unhammer. > This pair uses lttoolbox/CG. Beats Google for this pair. Testvoc clean > in both directions. Further information here: > http://wiki.apertium.org/wiki/Maltese_and_Arabic/Work_plan > > * apertium-sh-sl: >80% coverage over Wikipedia, error rate around 20%. > This pair will be moved to trunk. Fantastic work by Aleš, Hrvoje and > Jernej. This pair uses lttoolbox/CG. Testvoc clean sh->sl, needs work > sl->sh. Further information here: > http://wiki.apertium.org/wiki/Serbo-Croatian_and_Slovenian > > * apertium-tat-kaz: >80% coverage over Wikipedia, error rate around 20%. > This pair will be moved to staging. Really great work by Ilnar and > Jonathan. This pair relies on HFST/CG. More or less clean, but > challenging to testvoc because of `agglutinative' morphology. Pair not > in Google. Further information here: > http://wiki.apertium.org/wiki/Kazakh_and_Tatar > http://wiki.apertium.org/wiki/Kazakh_and_Tatar/Work_plan > > * apertium-quz-spa: Coverage difficult to calculate because of weak > adherence to orthographic norm. Will be moved to nursery. Good work by > pato. Not clean. Pair not in Google. Further information here: > http://wiki.apertium.org/wiki/Quechua_cuzque%C3%B1o_y_castellano > http://wiki.apertium.org/wiki/Quechua_cuzque%C3% > B1o_y_castellano/Apertium-quz-spa/Ortograf%C3%ADa > > * Corpus based lexicalised feature selection: The project was a > reasonable success. We didn't achieve any improvements for definiteness, > but we did for preposition selection. Further information and links on > Filip's user page: > http://wiki.apertium.org/wiki/User:Fpetkovski > > * Finite-state disambiguation: It was descovered at midterm that this > was too much work for Hrvoje student in the available time. After the > midterm he continued to work on Slovenian--Serbo-Croatian. > > Thanks to all our students for taking part! :) > > Fran
+1, great work, apertiumers :-) -- Kevin Brubeck Unhammer GPG: 0x766AC60C ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Apertium-stuff mailing list Apertium-stuff@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/apertium-stuff
