Google KELM: Will it Reward Or Penalize Your Content?

Here’s some proof that good guys don’t always finish last…

Google is working on a way to reduce bias and toxic content in search. 

Google’s AI team recently announced KELM, which is an acronym for Knowledge-Enhanced Language Model. Google is developing this model to address some of the issues with natural language processing. 

Pre-trained natural language processing (NLP) models like BERT take natural language that comes from the web and other task-specific documents. 

Here’s the problem: non-factual information and toxic content can eventually cause biases in the resulting NLP models.

✨Enter KELM ✨

KELM will be adding trustworthy factual content to the language model pre-training in order to improve accuracy and reduce biases and toxic content. 

But why should you care? 🤔

With the KELM model, the factual accuracy of your content could affect how your site is ranked… 

If you’re committed to getting your facts straight, you should be in the clear. 👍

But…If you’ve been known to bend the truth, you could see a drop in your rankings.👎

When KELM rolls out, it could have a major impact on any sites that have factually incorrect statements or ideas while sites with acurrate information have the chance to get ahead.

Do you want more news stories like this sent straight to your inbox? Be sure to sign up for our newsletter

Posted on Instagram:

Learn something new? Share it with a friend