Open-source Software Contributions
For a summary of all my open-source contributions, please visit my Github page.
Japanese Writing Assistance: The Hinoki Project
The Hinoki Project currently comprises of four systems. The Hinoki Project is described in @ALAHodoscek2012. Many of these topics are also described in further detail in my PhD dissertation: @HodoscekPhD2013.
Collocation search and comparison system Natsume
Timeline: 2009–
Summary: Users can find the collocations of a word, compare the observed frequencies for collocations across various genres, and check correct uses by looking at example sentences. Assists learners of Japanese as a second language in writing collocations in their target writing genre.
Scope of contributions: data (collocation) extraction, search API (since 2013/9), and server maintenance.
Source: Available from my Github page.
Publications
- @CASTELJHodoscek2012
- @ALAHodoscek2012
- @HodoscekPhD2013
- @ALANishina2014
Writing error detection and feedback system Nutmeg
Timeline: 2012–
Summary: Provides basic feedback for learner writing using automatic error identification of stylistic and orthographic inconsistencies, and register mismatch, among others. Error detection is based on data from Natsume and the Natane learner corpus.
Scope of contributions: register misuse detection feature and server maintenance.
Publications
- @yagi2014KNW
- @yagi20149KNW
- @CASTELJAbekawa2015
- @ALANishina2017
Japanese Speech Rating System
The Japanese Speech Rating System (JSRS) is a Python (Django) server and client application that crowdsources preference ratings of Japanese learner speech. The basic rating scheme is that raters indicate their preference between two Japanese learners speaking the same sentence.
The system also includes a sophisticated analysis module powered by the mdpref R package developed by @Okubo2015.
The source of JSRS is available from my Github page.
Publications
- @hatasa2016ICJLEBaliE