Modern Natural Language Processing and AI during COVID-19
🆕 In this episode: Daniel Whitenack, Modern Natural Language Processing and AI during COVID-19.
🚀 You could sponsor today’s episode. Learn about your ad-choices.
💙 Show your support for HumAIn with a monthly membership.
📰 Receive subscriber-only content with our newsletter.
🧪 Visit us online and learn about our trend reports on technology trends and how to bounce back from COVID-19 unemployment.
Episode Show Notes:
The evolution in natural language processing is happening with large-scale models increasingly used by encoders and decoders. With over 7000 spoken languages in the world, NLP is supporting speech-text processes with Google Translate supporting in this area. Languages including Bert, GPT-2 and Elmo are becoming part of text classification and translation.
NLP developments such as Question-Answer multilingual models, Stanford Question-Answer data set and Google Natural Question data set will shape trends in 2020. Accordingly, Tensor flow from Google is supporting multiple languages with others such as MLQA from Facebook being the first multilingual question-answer data set.
In this era of COVID-19, natural language processing will change as new tools emerge to communicate about safe practices to convey information to people. However, languages are dying such as Icelandic as millennials adopt new social trends and NLP is bridging the gap by supporting development of ‘Language First’ applications.