If the Terminator movies scared you, and you’re paranoid about Skynet, don’t read on.
Google have announced a huge jump forward in the way it understands queries and language on the internet. In a lovely reassuring marketing spin they are calling their new AI ‘MUM’ – Multitask Unified Model (maybe the mother of HAL).
This major reboot is not live yet but Google announced it in their blog here. They noted it is: “new technology Google is exploring internally to better understand language & make it easier to get helpful responses to complex Search needs.” It will most likely also help with targeted advertising.
Google said it is like BERT, which they launched in 2018 and shared in open source as a “neural network-based technique for natural language processing (NLP) pre-training called Bidirectional Encoder Representations from Transformers, or as we call it–BERT, for short in that it is built on a Transformer architecture”.
BERT is critical for Google Search as it is designed to help computers understand the meaning of ambiguous language in text by using surrounding text to establish context. But here’s the real kicker, Mum is 1,000 times more powerful than BERT. In just three years Google has taken something quite incredible, and made it 1,000 times better!
Google explain that MUM doesn’t just understand language, but it also generates language. They said MUM is “trained across 75 different languages and many different tasks at once, allowing it to develop a more comprehensive understanding of information and world knowledge than previous models. And MUM is multimodal, so it understands information across text and images and, in the future, can expand to more modalities like video and audio.”
Keep an eye out for Google updates of Mum as it will transform AI Search across all platforms, and if it’s shared on open source like Bert was, many other providers and technologies will reap the benefits. Some for good, and some perhaps, for evil…
