Page 3 of 9

I still need Windows for one piece of software that I use occasionally. So, since I've moved to Ubuntu, I've been using VMware for this (as it was the only solution at that time).

My VMware stopped working under Hardy. Luckily, I never needed my Windows-App for half a year. In Intrepid, I was able to just install vmware from their webpage and it restored my windows session from a year ago. Probably a record-breaking uptime for windows...

So, yesterday I played around with bootchart and found out, that the VMware-services took 4 precious seconds of my (and my battery's) lifetime on every boot. Not really worth it, as I'm unlikely to use my Windows-App anytime soon. So, here's what I did:

remove the links in /etc/rc*.d/*vmware

as the first command in /usr/bin/vmplayer add:

gksu -D "Need root priviledges to start vmware services." /etc/init.d/vmware restart

Works like a charm.

The title says it all: I am looking for a generic implementation that tells me the edit distance of two lists. The implementations on CPAN all seem to work on string-data. Which is OK for finding typos but makes WER calculation tedious.

So, I want a generic implementation that takes a comparator-function (as in sort {$a <=> $b} @list) and two lists and outputs the edit distance. Nice to have would be distance-weights and really nifty if the value of the comparator function (not only !=0 but how much lower or higher) was taken into account.

Luckily I don't need it now, so I don't have to write it. But it would be a great finger exercise for a Perl-in-NLP class.

vorläufige Terminvorschläge: 24. oder 28. Juli (Meinungen dazu bitte bis 21. Juli)

Literaturliste

1. Mangu, Brill, Stolcke (2000): Finding consensus in speech recognition: word error minimization and other applications of confusion networks. [Timo, +]
   
2. Wilson, Wilson (2006): An oscillator model of the timing of turn-taking. [David]
3. was zu SDS-Architektur?
   

vorläufige Terminvorschläge: 24. oder 28. Juli (Meinungen dazu bitte bis 21. Juli)

Literaturliste

1. Mangu, Brill, Stolcke (2000): Finding consensus in speech recognition: word error minimization and other applications of confusion networks [Timo, +]
   
2. Oszillator-im-Hirn-Kram [David]
3. was zu SDS-Architektur?
   

Higgins-Howto 

• Mozart über Synaptic installieren:
• mozart
• mozart-gtk
• mozart-stdlib
  
• Higgins aus svn+ssh://helios/projekte/inpro/INPRO_SVN/Code/Higgins auschecken.
  
• export HIGGINS_RUN=<pfad-zu-higgins> (am besten gleich in die .bashrc schreiben, dann weiß es jedes neu geöffnete terminal)
  
• ozengine launcher/Launcher.exe applications/Numbers/numbers.higgins
• von da aus die einzelnen Module starten (die Inspektoren gehen noch nicht)
• für die Sphinx-Anbindung:
• svn update in Eclipse (die Pakete heißen de.cocolab.inpro.higgins.*)
• die sind jetzt veraltet, oder was?
  
• TODO: Build-Pfad anpassen, noch irgendwas einbinden
  
• voilà

Higgins-Howto 

• Mozart über Synaptic installieren:
• mozart
• mozart-gtk
• mozart-stdlib
  
• Higgins aus svn+ssh://helios/projekte/inpro/INPRO_SVN/Code/Higgins auschecken.
  
• export HIGGINS_RUN=<pfad-zu-higgins> (am besten gleich in die .bashrc schreiben, dann weiß es jedes neu geöffnete terminal)
  
• ozengine launcher/Launcher.exe applications/Numbers/numbers.higgins
• von da aus die einzelnen Module starten (die Inspektoren gehen noch nicht)
• für die Sphinx-Anbindung:
• svn update in Eclipse (die Pakete heißen de.cocolab.inpro.higgins.*)
• TODO: Build-Pfad anpassen, noch irgendwas einbinden
  
• voilà

Klotz hin Gnubbel

This is what I get with the current acoustic model and a LM that was even trained including the correct sentence (und füge es ein in den Bauch des Elephanten).

Even using only just the correct sentence as a grammar returns und füge es,  instead of the complete sentence. The alignment shows, that es is supposedly spans the complete ein in den Bauch des.

I read that the current models are severely overtrained on one speaker, so I tried one of his utterances (de43-01, die Anwendung wird entwickelt) which is correctly understood if I use it as a grammar (effectively resulting in forced alignment) and which results in the beautiful phantasie wird entwickelt if I include this one sentence in the statistical LM as above. 

Thus, the bad results are probably due to the bad acoustic model. I've already uploaded the PentoNamingCorpus to Voxforge, thus hopefully, acoustic models will improve eventually. But if bad comes to worse, we'll have to train based on KCoRS and Verbmobil...

I am currently investigating the ton of classes that implement the Sphinx interface "SearchSpace", or one of the three sub-interfaces. There are 19 in total and I am likely to have to add another one for the feature that I have in mind.

Anyway, I decided that I need something, preferably an Eclipse-plugin to visualize class dependencies, and there are actually a few options:

• X-Ray would probably do the job, but it doesn't work. Maybe I just don't know how to install it correctly.
• Byecycle have a great screencast on their page and more friendly installation instructions. It shows a dependency graph between classes and automatically and incrementally optimizes the graph layout. Infinitely. Using 20% of your processor(s). It's quite slow and it seems to be limited to only show dependencies within the package, while the dependencies I'm interested in often cross dependency boundaries (classes from different packages implementing an interface).

This has been programmed before, but here it is for you to see (and use):

IPATextField, a simple descendant of JTextField that will only accept phonetic input (either SAMPA or IPA if you know your uni codes by heart) and show IPA symbols.

You can try it out directly, as a main routine is included. It's even useful as your tiny copy-and-paste-IPA-editor.

Higgins-Howto 

• Mozart über Synaptic installieren:
• mozart
• mozart-gtk
• mozart-stdlib
  
• Higgins aus svn+ssh://helios/projekte/inpro/INPRO_SVN/Code/Higgins auschecken.
  
• export HIGGINS_RUN=<pfad-zu-higgins> (am besten gleich in die .bashrc schreiben, dann weiß es jedes neu geöffnete terminal)
  
• ozengine launcher/Launcher.exe Potsdam/number.higgins
• von da aus die einzelnen Module starten (die Inspektoren gehen noch nicht)
• für die Sphinx-Anbindung:
• svn update in Eclipse (die Pakete heißen de.cocolab.inpro.higgins.*)
• TODO: Build-Pfad anpassen, noch irgendwas einbinden
  
• voilà

Page 3 of 9