I did another test with my Google Home. Watch this video:
Based on the first question, Google knows the thematic context or entity I am interested in and can at least partially create the context for the entity in subsequent questions. Unfortunately, this does not work in the classic search, but only via Google Home. But this shows which direction it is going and how important the entity concept is for voice search.
Google patents on the subject
I was able to find the following active Google patents on this topic:
Autocompletion using previously submitted query data
This Google patent is a bit older and was signed in 2009. It probably forms the basis for Google's autosuggest suggestions in general.
“The methods include receiving query information at a server system. The query information includes a portion of a query from a search requestor. The methods also include obtaining a set of predicted queries relevant to the portion of the search requestor query based upon the portion of the query from the search requestor and data indicative of new-zealand phone number data search requestor behavior relative to previously submitted queries. The methods also include providing the set of predicted queries to the search requestor. Other embodiments of this aspect include corresponding systems, apparatus and computer program products.”
According to the patent, Google accesses a set of additional search queries that are related to the currently entered search query. These are then ranked based on previous search queries made by the respective user and a few suggestions from this set are suggested as a supplement. In addition, click data, location-specific data, language-specific data or other similar types of data can be included in the ranking in order to better understand the respective context of the user.
This method can fulfill the following tasks:
A searcher receives further search suggestions before the requester indicates that the query has been completed.
Additionally, information related to previous searches of the user (or users) is collected (e.g. click data associated with the search results). From the obtained query information and the previous search information, a set of predicted suggestions is created and presented to the user.
Identifying teachable moments for contextual search
This patent was signed by Google in 2018 and granted in early 2020. The patent describes how Google recognizes search patterns based on a series of consecutive search queries. These search patterns can be based on identical entities and other semantically related terms. Based on the recognized pattern, Google can better identify the user's context or the meaning of the search query and therefore provide better search results. The search results can relate to both the current search query and previous search queries.
The patent lists some examples of possible patterns:
An example series of searches might include [Obama White House], [Obama White House speech], and [Obama White House speech today]. In this example series of searches, the entity terms might include "Obama" associated with the primary entity "Barack Obama" and the secondary entity "President of the United States," and "White House" associated with the secondary entity "White House." In this example series, non-entity terms include "speech" and "today." The pattern might be that the primary entity and secondary entities are consistent within the query series, e.g., associated with every search query in the series, and that the non-entity terms are inconsistent within the query series.
Another example series of queries might include [Obama speeches at the White House], [Bush speeches at the White House], and [Reagan speeches at the White House]. In this example series of queries, the entity terms might include "Obama," "Bush," and "Reagan," each associated with the primary entity "Barack Obama," the primary entity "George W. Bush," the primary entity "George HW Bush," the primary entity "Ronald Reagan," the secondary entity "President of the United States," and "White House" associated with the secondary entity "White House." In this example series of queries, a non-entity term might include "speeches." In this example, the query pattern might include the primary entity that is inconsistent within the query series and the non-entity term(s) that is consistent within the query series.