Language

It is not a bunch of words strung together that we exchange in order to get our needs met. It is not just a medium for communication. It is the medium of the mind. It shapes how we think. It is what helped us out smart other animals, by carrying knowledge of medicine, astronomy and etc over to us from the past. It is also the realm, where poetry lives. It is the time capsule of our history. It is where our ancestors are still alive. It is what comes close to what we might call heaven.

Unity in diversity

Linguistic diversity is a testament to the richness and complexity of human mind and our history. Every language is unique and has different forms depending on its geographical location and culture of its indigeous people. Language is the conduit that carries culture and heritage across generations.

How can we preserve languages?

The easiest way to keep a language alive is to speak it. Using our mother tongue in day to day life will keep it alive. But is that enough? No. We want the indigenous language not just to survive but thrive. Technology has become an essential part of our day to day life. We need to be able to read, write and converse with computers in our mother tongue. That is where AI comes in.

Aritificial Intelligence

Artificial Intelligence is already doing lot of amazing things. From playing Chess and Poker to driving cars and spotting celestial bodies. AI will accelerate the pace of innovation and creativity, like never before. In simple words, you can just describe the picture on your mind to the machine and it will paint it, bringing out the creativity in people who cannot handle a paint brush and even people without arms. We believe Artificial Intelligence will play a key role in shaping our future as a whole.

What should we do?

But we need to be ready. We need to train the machines to understand our language and that is where we need your help. We need your help to teach AI your language. AI can already understand many languages like English and French, because there are huge corpora from which the AI can learn. Indigenous languages do not have that. Join us in our quest to make machines learn what it means be unified under rich diversity.

What have we done so far?

We have built language models for Tamil, Malayalam and Bangla.

Language model is the ability assess whether a given sentence is valid one. For example, a native english speaker can easily say that the phrase 'how are you?' is valid but 'are how you?' is not.

Language modelling can also be thought of as: given sentence with few words removed, one can easily fill in those blanks. Everyday, I ____ up at 6 am and drink ____ before going to the gym. The first blank can be filled with wake and second one with anyone word among tea, coffee, juice and water
You can play with the model here

We have ambitious goal of building a series of tools for indigenous languages, such

Android/iOS App for Data collection, Tagging
speech-to-text and text-to-speech to help converse with the computer
translation system across Indic and other languages.
Auto complete keyboard utility for Indic languages
document analysis and recognition system

But one step at a time. Our first task to build a parallel corpus for training the translation model. Please join the group indicnlp@googlegroups.com to help.

For more information, suggestions or questions, you can contact us at this email id.

IndicNLP

AI for Indian languages