Min menu


Google uses artificial intelligence in search results

 Google is constantly working on improving its search engine, which has become more than a way to connect the user to a site that offers content. It is even faster to answer users' questions.

Google uses artificial intelligence in search results

Pandu Nayak, Vice President of Search at Google, puts forward an example with a sentence like, “Can you get someone’s medicine from the pharmacy?” And that question was a challenge to Google's search engine, as it was unable to understand it.

Therefore, the researcher looking for this sentence was receiving results from one of the sites that clarify the answer to this question. But now the search engine is getting smarter. As he is able to understand questions of this type and provide answers to them.

The key, Nayak explained, is artificial intelligence. The search engine has been able to answer these questions long ago after Google integrated artificial intelligence and machine learning models into the engine. Besides, much of the credit goes to natural language understanding systems, such as GPT-3.

And during 2021, Google will update its engine again. But this time it is based on a new generation of artificial intelligence and machine learning systems, which is called MUM. Google revealed it for the first time last May and this model can be described as a thousand times larger than the previous generation, which was named BERT.

Google uses artificial intelligence in search results

The new MUM model consists of 175 billion parameters, a unit of measurement for the size and complexity of AI models. Besides, this model understands more than 75 languages ​​after being trained on it. Besides, it is able to avoid racist or sexist contents. But do not forget that Google has fired two of its most important employees after they published a research paper on the problems of this technology in terms of discrimination and bias.

In addition, Google offers additional updates to the search engine, such as new search boxes called Things to Know, which provide the user with content related to what he is looking for for more information. In addition to focusing on visual search using Google Lens, all of these technologies will undoubtedly benefit from artificial intelligence.

The behavior of the search engine will change to provide the best possible user experience. For example, when you search for “acrylic drawing”, the engine may display different techniques for this type of drawing on the one hand, or display different types of drawing. This will depend on artificial intelligence.

In addition, the company introduces a new feature based on the MUM model, which is the feature of understanding video content. As the search engine will be able to provide search results for visual clips related to the content of the search, but based on understanding the clips and not relying on their titles only.