Credibility Through Non-native Language Varieties in Conversational Pedagogical Agents
The research project includes five major work packages: The baseline is established through the compilation of key linguistic cues, the implementation testing of existing agents, and the development of instruments to measure credibility. The next phase prepares the linguistic input of annotated texts, which is fed into the CPA, and tested empirically. The evaluation phase includes the linguistic evaluation of the text output, the computational training of the CPA, and the psychological pre-testing. The central phase is the necessary extension and refinement of input, CPA training, and statistical analyses. The final phase makes the project results available to the academic community through respective discipline-specific publication outlets.