The state of Indic Languages in NLP

CalmDownKarm commented 5 years ago

Abstract (2-3 lines) The Hindi speaking internet is exploding, Hindi language newspapers absolutely outrank english language papers in terms of circulation. While there are a lot of players in the NLP space in India, those focusing on Indian languages are few and far between. Let's talk about the situation, the environment, the tooling, the language models, the problems
Brief Description and Contents to be covered The biggest problems with the environment today is that the gap between state of the art results and tooling people use in production is extremely large. Organizations that do work on Indian languages in NLP almost never publish source code or open source tooling. I'm going to talk about these problems and more importantly what people can do to help bridge this gap.
Pre-requisites for the talk Basic fundamentals of NLP are nice, but not absolutely necessary
Time required for the talk 30 - 45 Minutes
Link to slides TBD
Will you be doing hands-on demo as well? No
Link to ipython notebook (if any)
About yourself I'm a developer at gramener and a research assistant at IIITDelhi working on deep learning based approaches to NLP
Are you comfortable if the talk is recorded and uploaded to PyData Delhi's YouTube channel ? Sure
Any query ?

shagunsodhani commented 5 years ago

Hey @CalmDownKarm Thank you for proposing the talk. It is an interesting and extremely relevant topic. I am looking forward to your slides.

MSanKeys963 commented 5 years ago

Hi, @CalmDownKarm. Can you update the slides, please? And would you like to deliver this in our upcoming meetup?

CalmDownKarm commented 5 years ago

Yes to deliver this on the 15th, but earliest I can do slides would be this weekend

MSanKeys963 commented 5 years ago

Yeah, sure. That works!

CalmDownKarm commented 5 years ago

https://docs.google.com/presentation/d/1moSbzjQYmBIOs7SUeb1nuEgUr8K_hBiu1IhaNG2Lg6U/edit?usp=sharing Here's an early draft of the slides.

MSanKeys963 commented 5 years ago

@shagunsodhani please have a look.

shagunsodhani commented 5 years ago

Hey @CalmDownKarm The slides look good but I am worried that they would not be sufficient for a 3-45 min talk. If you are planning to talk about more things (or in more detail than what is appears on the slides right now), I would recommend that you add them to the slide for the benefit of the audience.

Best of luck for the talk!

pydatadelhi / talks

The state of Indic Languages in NLP #100