Facebook will set up a consortium to tackle the challenges of Natural Language Processing (NLP). Realistic interaction between speech and computers creates quite a few obstacles, which Facebook wants to solve with this research collaboration.

Three months ago, Facebook distributed a number of calls for research proposals in three sub-areas of NLP. More precisely, Facebook asked for proposals for deep learning for NLP, NLP through efficient compute and neural machine translation. The latter technique is intended for dialects for which very little (written) data is available.

Facebook has selected eleven winners out of 115 proposals submitted from 35 countries. These proposals will be part of the research group, which has been renamed the AI Language Research Consortium. There is a whole range of research areas that need to be further assisted by the consortium. These are representation learning, content understanding, dialog systems, information extraction, sentiment analysis, summarization, data collection and cleaning, and speech translation.

Cooperation with academia

In other words, there is hardly any part of NLP in which the group does not want to be active. In addition to collaborating with Facebook researchers, the eleven winners also receive funding for their projects. The winners will also participate in an annual workshop where results will be presented. They will also play a role in large NLP conferences.

“These research awards in NLP and machine translation were launched as a follow-up to our long-term goal of supporting open research within the NLP community and strengthening the collaboration between Facebook and academia,” wrote Facebook in the blog post. “Facebook strongly believes in open science and we hope that the consortium, as well as these research awards in NLP and machine translation, will help accelerate research in the NLP community.

According to VentureBeat, the project could be seen as an answer to Amazon’s Alexa Fund and Google’s AI for Social Good project. Whether Facebook’s project is more framed as creating more innovation (like Amazon) or solving problems in the world (Google), the future must tell.

This news article was automatically translated from Dutch to give Techzine.eu a head start. All news articles after September 1, 2019 are written in native English and NOT translated. All our background stories are written in native English as well. For more information read our launch article.