NaturalLanguageProcessing4All: A Tool for Natural Language Processing in Social Studies Education for Non-Programmers

Publikation: KonferencebidragPosterForskningpeer review

In this poster I present Natural Language Processing 4 All (NLP4All), a web-based text analysis system designed for use in classrooms, specifically high schools. NLP4All is designed to let students use Machine Learning approaches to explore language and words in corpora of texts, and to facilitate an explorative process of investigating similarities and differences between texts. To do so, NLP4All draws on concepts from the emerging field of Explainable AI – AI that is grey- or white-boxed so that users of ML understand how AI is making the decisions or classifications that it does (Gunning 2017; Holzinger et al. 2018). I present a a use case of NLP4All in which students use the system to explore how different political parties use different language on Twitter.
Giving learners without programming skill the ability to use Machine Learning is gaining increasing attention (e.g. Kahn and Winters, 2018.; Hitron et al. 2019). The idea behind NLP4All is slightly different: to let leaners use Machine Learning to explore an existing school practice, namely text analysis, but using Machine Learning as an exploratory process.
OriginalsprogEngelsk
Udgivelsesår2020
StatusUdgivet - 2020
BegivenhedConstructionism 2020 - Trinity College, Dublin, Irland
Varighed: 26 maj 202029 maj 2020
http://www.constructionismconf.org/

Konference

KonferenceConstructionism 2020
LokationTrinity College
LandIrland
ByDublin
Periode26/05/202029/05/2020
Internetadresse

Se relationer på Aarhus Universitet Citationsformater

ID: 177504904