A New Method For Language Revitalization Using
Language Teaching Machines
Rakesh Kumar Reddy Gowra
University of Windsor, ON, Canada
Language is one of the basic needs of the
humans after environment, water and food
Method 1: Standalone
How This Works
Language Teaching
Machine Learning
According to UNESCO endangered Languages
Atlas, India tops the list with 197 endangered
Languages and USA is second in the list with 191
Language Revitalization is reversing the decline
of a endangered Language by using available
Source: Sebastain stuker, IEEE 10.1109
Method 2:Network based (proposed Method)
Language Teaching Machines(ASR-SMT) use
the Training Data to teach a Language. The
proposed method effectively collects the training
data from the community with the help of a simple
Hybrid approaches like ANN/HMM, RBMT/SMT
are to be considered to increase the performance
of the system
The proposed method is very effective method
to aid language revitalization.
Processing speed is a main concern in real time
[1] Sakriani Saktia, Michael Paula et all. (2013). ASTAR: Toward translating Asian
languages. Computer Speech and Language. 27
(1), 510-525.
[2] Bowen Zhou, Xiaodong Cui et all . (2013). The
IBM speech-to-speech translation system for
smartphone: Improvements for resourceconstrained tasks . Computer Speech and
Language. 27 (2), 592-618.
Designing a mobile phone application which
collects data and teaches a language
Encouraging the communities to collaborate for
the research and learn a endangered Language
Designing ASR-SMT machine for Language
 Maintaining and collecting Language Database
The performance of the proposed method
depends on ASR and SMT Techniques
Performance of ASR by NIFT

Scientific poster example