Lemko: Resources for Endangered Minority Language Revitalization

Natural Language Processing (NLP) can be used to help revive endangered minority languages like Lemko. Computer aids help meet the special needs of the disabled and new speakers. Below are resources for building such tools.

Resources

Non-Governmental Organizations (NGOs)

The Endangered Languages Project

Living Tongues Institute for Endangered Langauges

Foundation for Endangered Languages

Committee on Endangered Languages and Their Preservation (CELP)

Long list of organizations and resources at the Endangered Language Fund

Machine Translation

Users will input English and receive rough-draft Lemko output. This will allow new speakers to revive the endangered language by creating content in it before having achieved fluency.

• Statistical Machine Translation (SMT): Moses statistical machine translation system, Microsoft Translator Hub
• Neural Machine Translation (NMT): Neural machine translation for low-resource languages

Convert to/from Cyrillic script (OPERATIONAL)

Online app that will convert to and from Cyrillic can be used as a crutch until new speakers master the alphabet.

Spell check

An online spellchecker based on the most authoritative handbook will save time and help new speakers learn. Spell check will be for Open Office, Libre Office, and possibly Microsoft Office.

• How to Write a Spelling Corrector

Grammar check

An online grammar check for punctuation and case endings will save time and help new speakers learn grammar.

• Ґраматыка лемківского языка (Grammar of the Lemko Language)

Transcribe audio or video

Automatic closed captioning (CC) for the hearing impaired and automatic rough draft subtitles.

• Google Cloud Speech API
• The Holy Grail: Automatic Speech Recognition for Low-Resource Languages.

Text to Speech (TTS) Voice synthesis

Computer will read out text for those who don't know the alphabet, the blind, etc.

• Project Unison - Crowdsourcing a Text-to-Speech voice for low resource languages

Optical character recognition (OCR)

Will digitalize handwritten and printed text for subsequent publication online and/or translation

Text aligner

Would automatically align Polish or English translations of Lemko texts and vice versa.

• The hunalign sentence aligner

Bilingual online dictionary

Would be the first ever online Lemko-English dictionary/glossary

• Лемківско-польскій/польско-лемківскій Словник (Lemko-Polish Dictionary)

Social network

Would be possible to network with other new speakers, find language exchange partners, bulletin boards, etc.

• Building a social network with Django
• Bootcamp - open source enterprise social network built with Python using the Django Web Framework.
• Cabin - build a feature-rich, scalable social network app using React and Redux.

AI Chatbot with voice synthesizer

Just for fun - would be an artificial intelligence talking Lemko chatbot, like Cortana or Siri.

• Ultimate Guide to Leveraging NLP & Machine Learning for your Chatbot
• How To Develop a Chatbot From Scratch