Abstract:

In order to carry out innovative research on education, academic research groups often make use of educational apps. These educational apps can be either existing commercial apps of companies, or apps that were developed by the researchers themselves. While a clear advantage of existing commercial apps is that one doesn't have to develop its own app, there are many important disadvantages, e.g. that such an app to a large extent is a black box, and it is not a flexible research instrument, for which aspects of the app can be easily changed and studied. Therefore, academic researchers often develop their app. The problem with such research-oriented apps is that they require much time and effort to be developed, while their use is often limited to one or a couple of research projects, with limited possibilities for future exploitation and sustainability. This has to do with a number of factors. First, research funding agencies and universities are interested in research and fund what is necessary to conduct research. In spite of the recent, growing attention and sometimes accompanying funding for outreach, in general the money is too limited to ensure future exploitation. In addition, it is not the primary task of research funding agencies and universities to develop commercial products. Besides using an existing commercial app or an app developed by the researchers themselves, another option is to cooperate with a spin-off company. Although the primary task of a spin-off company has a commercial nature, spin-off companies and the research groups from which they emerged can establish very fruitful forms of cooperation that benefit both of them.

In this paper we focus on specific ways in which this cooperation can benefit research and we present three examples of innovative research projects in which Radboud University [RU] cooperates with the RU spin-off company Novolanguage:
[1] DART: Dutch ASR-based Reading Tutor [http://hstrik.ruhosting.nl/DART/]
[2] MyST: My Speech Trainer [http://hstrik.ruhosting.nl/MyST/]
[3] Indonesian EFL learners’ perceptions on web-based language learning

In these three Computer Assisted Language Learning (CALL) projects automatic speech recognition (ASR) was employed to provide instantaneous feedback on spoken utterances of language learners. To make this possible, use was made of artificial intelligence (AI) deep learning toolkits to develop deep neural nets that can detect errors in spoken utterances. Advantages of this cooperation are that for carrying out the research, use can be made of the resources of the spin-off Novolanguage, esp. components of the CALL systems developed by Novolanguage. In the DART project only the ASR backend [for Dutch] was used, while in the 2nd and 3rd project use was made of the ASR backend [for English], in combination with Novo Studio, a user-friendly editor developed by Novolanguage that makes it easy to create and change content. An additional advantage is, that when language learners interact with the Novolanguage system, information about these interactions is stored in log-files. These log-files are rich resources of information which can be analysed later for research purposes. An extra feature of the MyST project was that the goal was to improve spoken Academic English in an educational context, and that it was tested at educational institutes, i.e. the Universities of Nijmegen and Utrecht. We will describe experiences, and discuss the possibilities this offers.

Keywords:

University spin-off, university-industry cooperation, innovative research, apps, education, Computer Assisted Language Learning (CALL).