According to an article published on Tuesday, Google’s featured snippets on desktop will now be entirely rendered through artificial intelligence. These "sentence compression algorithms" have just gone live specifically for desktop searches.
These algorithms, enhanced by deep neural networks, are designed to sift through vast amounts of data and text to deliver the precise information you need. For instance, if you ask, "What is the best Christmas movie?", Google will provide an answer like the one illustrated below:
Despite personal preferences in films, what stands out is that the featured snippet efficiently condenses the entire article into a concise list focused on the specific search query. The original article is much more extensive than just the 8 movies listed.
The newly launched sentence compression algorithm is also adept at scanning large volumes of text and summarizing it into a short, precise answer to the searcher’s question. Although I haven’t encountered a live example yet, it would be insightful if anyone could share one they find.
This artificial intelligence operates with the aid of "supervised learning," requiring the support of a global team of 100 PhD linguists to help the algorithm comprehend human language.
This team currently covers between 20 and 30 languages, with plans to transition towards a more automated form of AI, also known as "unsupervised learning."
Research in the field of unsupervised learning is already being conducted by tech giants like Google and Facebook. However, the full realization of such technology is said to be far off; artificial intelligence still necessitates some level of human intervention.
