DeepText for Facebook Messenger

June 2016 Facebook

Recognizing intents and extractings slots (arguments) from text messages or Facebook posts can improve people’s experiences when using Facebook products. It can also enable monetization of platforms where text is the prevalent form of communication.

I developed the initial DeepText models for recognizing ride intents for Uber, Lyft, and Taxi in Messenger. I also developed the models for recognizing Facebook posts with for-sale intent, and extracting object and price in for-sale posts.

The work was presented in more details in the following articles:

Facebook Article
TechCrunch Article