Collect4Nlp

This interdisciplinary PhD project is situated on the crossroad of Natural Language Processing (NLP), Computer-assisted Language Learning (CALL) and Crowdsourcing. It aims at tackling through Crowdsourcing two issues of the NLP and CALL domains: the lack of Linguistic Resources (LRs) for NLP, and the lack of exercise content for CALL. In order to do so, this PhD researches a recent subject with high-impact potential: the combination of implicit crowdsourcing and language learning through mechanisms that generate language learning exercises from LRs, crowdsource the answers of learners to these exercises and use the crowdsourced answers to correct and extend the LRs. With regards to languages, this PhD focuses on two languages of relevance to the Italian context: Italian and German as spoken in South Tyrol.  

This PhD aims at using crowdsourcing as an approach to treat the two major aforementioned NLP and CALL issues in combination. It focuses on a type of crowdsourcing called implicit crowdsourcing in which a crowd creates data as a side effect of using a tool created for a purpose that is not directly related to crowdsourcing data. This approach has been actively explored within the European Network for Combining Language Learning with Crowdsourcing Techniques (EnetCollect) COST Action. While the ideas and achievements reported so far on the subject are promising there are still many aspects to consider, double-check and evaluate more precisely in order to fully understand the viability and potential of this approach. This PhD will explore these questions by extending the approach to Italian and German as spoken in South Tyrol and to new combinations of language learning exercises and LRs.