Facebook this week pulled the curtains off Deep Text, its new deep-learning-based text-understanding engine, which it said can analyze the textual content in several-thousand posts per second.
The social network said in a blog post by engineer Ahmad Abdulkader, applied machine learning platform team technical program manager Aparna Lakshmiratan and research scientist Joy Zhang that the aim of Deep Text is to better understand text across Facebook and to reduce reliance on language-dependent knowledge, adding that the technology is already being tested on Messenger and offering the example that Deep Text can decide whether a user is looking for a taxi by distinguishing between “I just came out of the taxi” and “I need a ride.”
Facebook said in an email to SocialTimes that examples of uses for Deep Text include:
- Better understanding people’s interests: Facebook is testing the ability to generate large data sets with semi-supervised labels. Using public Facebook pages, we train a general interest classifier we call PageSpace, which uses Deep Text as its underlying technology.
- Joint understanding of textual and visual content: Deep Text will work with the visual content understanding teams to build new deep-learning architectures that learn intent jointly from textual and visual inputs.
- New deep neural network architectures: Bidirectional recurrent neural nets show promising results, as they aim to capture both contextual dependencies between words through recurrence and position-invariant semantics through convolution.
Abdulkader, Lakshmiratan and Zhang wrote in the blog post:
Text understanding includes multiple tasks, such as general classification to determine what a post is about—basketball, for example—and recognition of entities, like the names of players, stats from a game and other meaningful information. But to get closer to how humans understand text, we need to teach the computer to understand things like slang and word-sense disambiguation. As an example, if someone says, “I like blackberry,” does that mean the fruit or the device?
Text understanding on Facebook requires solving tricky scaling and language challenges where traditional NLP (natural language processing) techniques are not effective. Using deep learning, we are able to understand text better across multiple languages and use labeled data much more efficiently than traditional NLP techniques. Deep Text has built on and extended ideas in deep learning that were originally developed in papers by Ronan Collobert and Yann LeCun from Facebook AI Research.
While applying deep-learning techniques to text understanding will continue to enhance Facebook products and experiences, the reverse is also true. The unstructured data on Facebook presents a unique opportunity for text understanding systems to learn automatically on language as it is naturally used by people across multiple languages, which will further advance the state of the art in NLP.
Readers: What are your initial impressions of Deep Text?