Democratizing NLP: considerations from resources to algorithms - Livestream

In recent times, there is an enormous increase in the use of digital media as a channel for communication. This, combined with the fact that natural language is a ubiquitous mode of communication, has increased the volume of natural language data available for processing. Also, this has resulted in opportunities for extending NLP based services/solutions in multiple languages. Although existing algorithms for NLP tasks have been shown to achieve reasonable performances for the high-resource languages such as English, French etc. across tasks, the same is not true for many other languages widely spoken in the world. To ensure that the benefit of NLP research reaches a wider audience across different regions, it is important to pay special focus to the low-resource languages, or consider support for low-resource languages in an intelligent manner in the algorithm design phase. The talk will touch upon a few such considerations in this talk.

Speaker: Dr. Maunendra Desarkar, IIT Hyderabad, India

Register at weblink to attend