Facebook To Use Artificial Intelligence System DeepText To Better Understand Posts, Filter Spam

Facebook has built an artificial intelligence (AI) engine to better understand the multitude of text and information that users post on the social network.

The social media giant on June 1 unveiled DeepText, a deep learning-based text understanding engine that can analyze the content and context of text users post on Facebook with near human accuracy and at an impressive speed: thousands of posts per second.

With users posting about 400,000 new stories and 125,000 comments per minute on Facebook, the AI system can potentially help improve overall user experience by classifying content, improving search results and weeding out spam.

"We want DeepText to be used in categorizing content within Facebook to facilitate searching for it and also surfacing the right content to users," said Hussein Mehanna, an engineering director at Facebook's machine learning team.

What's more impressive with DeepText is it can interpret text not just in English but also in more than 20 languages.

"The community on Facebook is truly global, so it's important for DeepText to understand as many languages as possible," Facebook said in a blog post. "Using deep learning, we can reduce the reliance on language-dependent knowledge, as the system can learn from text with no or little preprocessing. This helps us span multiple languages quickly, with minimal engineering effort."

Facebook already tests DeepText on some applications. In the case of Facebook Messenger, the tool offers a better understanding on the intent behind a text.

If a user says he needs a taxi, for instance, DeepText could recognize that this person needs a ride and thus sends a prompt to contact a taxi.

Projects like DeepText could play a pivotal role in Facebook with its plans to use AI to improve its platform.

The better Facebook understands what its users talk about, the better it will be at curating the News Feed. DeepText may eventually help counter online hate speech as required by European regulations, which call for the review and removal of hateful speech posted online within 25 hours of notification.

DeepText can also help in hiding spam, which can help improve user experience and keep more people using Facebook and its products. Search engine giant Google also uses AI to ensure delivery of important emails and improve filtering amid proliferation of spam and phishing.

ⓒ 2024 TECHTIMES.com All rights reserved. Do not reproduce without permission.
Join the Discussion
Real Time Analytics