Assignment #4: Evaluating a DH project
The project evaluated in this text is called "Information Extraction from Spoken and Informal Language". It is funded by the German Institute for International Educational Research (abbr.: DIPF) which supports projects that are concerned with educational research that aims at being of benefit for society. The timeframe in which the project is conducted is set for 2013-2015, all responsible researchers are named and information about them can be accessed immediately. Given the fact that the project works with data that is based on spoken and informal language, electronic devices and the possibility to get technical support are a great benefit for the task. Even if the objects of investigation are transcriptions, recordings of natural language have to exist and can be integrated in the project reports, so that everybody who is interested can comprehend the results of the study. With electronic means and programs it is also much less effort to examine huge amounts of data in comparison to research conducted on paper. The specific issues and steps the researchers will take are described detailed and so that everyone who reads the description understands what exactly the examinations purpose is. Examination of spoken language is a very interesting and important task for research in the field of Natural Language Processing. The development of translation systems for instance has improved over the years and the results are satisfying considering the complexity of the task. Projects as the one presented here could contribute to a further improvement of for example automatic translation programs, even if especially oral speech and informal speech changes quickly and it is nearly impossible to create "the perfect dataset" which contains all changes that exist of these language-variations. There is no primary audience named on the Website, but as the project concerns itself with the collection of data, everybody who is in need of results of examinations in this topic will find the projects results important and will have a use for it. I personally am very interested in the outcome of the project because I wrote a short paper about the performance of free online translation programs and found the performance with Natural Language and especially informal language to be very disappointing. The research examines how existing Natural Language Processing methods can be applied to utterances which do not fulfill the requirements of the standard language. It will be interesting to see whether the established algorithms and research methods for non-standard language have to be revised. Further information can be found via: http://www.dipf.de/en/research/projects/information-extraction-from-spoken-and-informal-language. The static Website contains the most important information about the ongoing project, it is therefore easy to navigate on it.