According to foreign media, the verge reported that Google has introduced a series of new features at its “search on” campaign. It will improve its basic Google search service in the coming weeks and months. These changes mainly focus on the use of new artificial intelligence and machine learning technology to provide users with better search results. < / P > < p > according to Prabhakar Raghavan, Google’s search director, says that 15% of Google’s search queries every day are queries that Google has never seen before, which means that the company must constantly strive to improve search results. < / P > < p > part of this is due to poorly spelled queries. According to Cathy Edwards, Google’s vice president of engineering, one out of every 10 search queries on Google is misspelled. Google has long tried to help with its “did you mean” feature, which recommends correct spelling. By the end of this month, it will be updating the feature on a large scale, using a new spelling algorithm supported by a neural network with 680 million parameters. It runs within three milliseconds of each search, and the company promises to provide better advice for misspelled words. Another new change is that Google search can now index a single paragraph from a web page, not just the entire page. For example, if users search for “how to determine whether my windows are UV glass”, the new algorithm can find a single paragraph of text on a DIY forum to find out. According to Edwards, when the algorithm is rolled out next month, it will improve 7% of queries in all languages. < / P > < p > Google is also using artificial intelligence to divide a broader search into sub topics to help provide better results (for example, helping to find home fitness equipment designed for small apartments, rather than just providing general fitness equipment information). < / P > < p > finally, the company has also begun to use computer vision and speech recognition to automatically tag and divide video into segments, an automated version of the existing chapter tools it has already provided. For example, a cooking video or sports game can be parsed and automatically divided into chapters (Google has already provided the creators with manual manipulation), which can then emerge in the search. This is similar to the company’s existing work of showing specific podcast episodes in search rather than just general push. Skip to content