On this page we would like to suggest and discuss possible UIMA applications and UIMA usage scenarios.
If you have a great idea for an UIMA application or usage scenario feel free to add it to the list below. We are also interested if you just want to integrate UIMA into an existing application.
This is the right place for all discussions and suggestions around UIMA applications.
The list below contains the current available suggestions and discussions.
Suggested UIMA applications
UIMA Firefox plugin
We would like to show-case the benefit of UIMA with a short simple application. This application should run as a Mozilla Firefox plugin to analyze the currently loaded web page in the browser and highlight the detected annotations. It should be easy to configure so that users can plug in their own analysis components to highlight the entities of their choice.
The flow of the application is:
1) Web page is loaded in the browser
2) With the browser page load event the UIMA plugin application is triggered to analyze the content of the web page and to return the analysis results.
3) The analysis results are transformed and used to highlight the corresponding entities in the web content.
The UIMA analysis engines in the plugin can be configured with PEAR package or UIMA service references that should be used for the analysis. The returned analysis result is transformed and prepared that it can be used for highlighting. Maybe the result can run through additional custom filers before it is used for the highlighting to have an additional customization on the highlighting.
The Java UIMA runtime in the plugin can maybe realized with the "Java in Firefox extension" plugin (http://developer.mozilla.org/en/docs/Java_in_Firefox_Extensions).
We currently have not much experience with Firefox plugin development so we are looking for volunteers to help us develop this application. We especially need some support in the area of "Firefox extensions with Java" and "highlighting with JavaScript". So everybody who is interested in such an application is welcome to help.
On Firefox plugin for UIMA...
Hi, at the University of Tor Vergata, the ART Group (http://art.uniroma2.it) is developing a Firefox extension for building/populating Semantic Web ontologies by getting content from traditional Web pages. The interaction with java is made possibile through the SIMILE java-firefox-extension project (http://simile.mit.edu/wiki/Java_Firefox_Extension). At present time (2008/06) the extension will be released before or just after summer, and will be compatible with the recent Firefox 3 release (we're now in test phase and correcting some minor compatibility issues with FF3).
Semantic Turkey (this is the name of the Firefox extension), uses OSGi standard (through Apache Felix) so that its java layer can be extended with further functionalities. The client side can be extended through standard Mozilla extension technology, by declaring your extension as a Firefox extension with a dependence on Semantic Turkey.
An here it comes UIMA: one of the first Semantic Turkey extensions we're working in, is an UIMA based extension, allowing users to extract relevant content from web pages and use it to populate an existing ontology. Much of the needed technology is already up, we have just realized an extension for taking pointwise annotations (the original Semantic Turkey does something more similar to bookmarking) and we're now in the direction of including UIMA into the whole stuff...
We have also been awarded with the IBM UIMA inovation award for proposing a project which was initially aimed to realize an web based annotation tool following UIMA standards for annotation, but we're moving beyond this to get full UIMA integration.
If you are interested in reading any documentation on Semantic Turkey, before its official site will be up, you can download the following papers, from my web page (http://art.uniroma2.it/stellato).
- ST now: Francesca Fallucchi, Maria Teresa Pazienza, Noemi Scarpato, Armando Stellato Semantic Turkey: A new web experience in between Ontology Editing and Semantic Annotation In proceedings of the Fourth International Conference on Web Information Systems and Technologies (WEBIST 2008) Funchal, Madeira - Portugal, 4-7 May, 2008
- ST and ontology population: Maria Teresa Pazienza, Marco Pennacchiotti, Armando Stellato A Web Browser Extension for growing-up Ontological Knowledge from Traditional Web Content In Proceedings of the Sixth International Conference on Language Resources and Evaluation (LREC 2008) Marrakech (Morocco), 28-30 May, 2008
- first paper on ST prototype: Donato Griesi, Maria Teresa Pazienza, Armando Stellato ESWC2007 - Semantic Turkey - a Semantic Bookmarking tool (System Description) 4th European Semantic Web Conference (ESWC 2007) Innsbruck, Austria, June 3-7, 2007
On Firefox plugin for UIMA...
We should also keep in mind to not only analyze websites through this extension, it should be also possible to analyze xml/text files since they can be opened by firefox.
On Firefox plugin for UIMA...
Semantic Turkey has been released. The official site is http://semanticturkey.uniroma2.it. The UIMA extension is under active development and will be released in the next months
On Firefox plugin for UIMA...
UIMAST, Semantic Turkey extension for integrating UIMA, has been released. This is the official site: http://semanticturkey.uniroma2.it/extensions/uimast/
At the moment we covered the target of the Award, that is to enable human annotators to take UIMA-compliant annotations from Firefox. Moreover, we also included a CAS editor.
UIMAST can be extended through OSGi to include several annotation types. The default one is xpointer.
We are now moving in the direction of including UIMA annotators for extracting info from Web Pages and use it to populate Semantic Turkey ontologies.