In this thesis we describe Duolingo, a language learning system where students learn a foreign language while simultaneously translating text. Because students both learn and translate, Duolingo is both a language learning system and a translation system. We explain in detail how our system is built and highlight all the key components. We show that students learn with Duolingo as well as with traditional learning methods. Additionally, we show that Duolingo produces high-quality translations in very large quantities.
Duolingo is an example of human computation, where computer systems enable massive collaborations between humans and computers for the benefit of humanity. With Duolingo, we establish that education can serve as a vehicle for human computation with millions of collaborators.
Luis von Ahn (Chair)
Kenneth Koedinger (HCII)
Oren Etzioni (University of Washington)