Using Mobile Devices to Rewrite Queries on Search Engines
People use search engines to find answers to their situational or informational needs. A Google patent that was recently granted describes how a search engine might provide rewritten queries for people searching on handheld mobile devices such as mobile phones. Queries are rewritten using annotations from a user-specific knowledge graph, which can be built using data from a variety of mobile applications.
The patent describes how rewritten queries may provide more personalized results and make use of data from the knowledge graph it is building in a way that is transparent to a searcher.
I’ve written about rewritten queries before when discussing Hummingbird and Rankbrain.
The recently granted Google patent introduces the concept of rewritten queries to protect personally identifiable information on mobile devices such as phones and programs that may be used on them, which are built with a user-specific knowledge graph. That data is referred to as private knowledge graphs.
User-Specific Knowledge Graph
In 2012, Google debuted its massive knowledge graph. In a previous post titled User-Specific Knowledge Graphs to Support Queries and Predictions, I discussed smaller user-specific knowledge graphs. I also discussed a Google white paper on Personalized Knowledge Graphs titled Where Will We Go With Personalized Knowledge Graphs?
Mobile web searches have become commonplace in recent years. Web search engines attempt to rank results in the most relevant order for given search queries. Before being processed, a query is annotated and rewritten.
This allows it to include additional query terms, such as synonyms, that were not present in the original query, or to otherwise prepare the query for processing by the search engine. Google recognizes that rewritten queries and user-specific knowledge graphs can result in better mobile device searches by using rewritten queries from data saved on mobile devices and collected into a user-specific knowledge graph. Those rewritten queries may not even be visible to searchers on those devices.
When a search engine understands the searcher issuing a query, it can become more helpful. A server-side search engine may find the query “where to have dinner with David” meaningless unless the search engine has detailed information about the searcher and their contacts.
Mobile devices make greater use of user-specific knowledge than desktop computers.
In many cases, handheld devices have more user-specific knowledge than a server-side search engine. This knowledge may be derived from what a searcher has previously typed, what he has seen on-screen, or from his current surroundings. This method of searching may cause a mobile device to generate rewritten queries.
Mobile search queries may be rewritten on a client device using the device’s extended knowledge. This extensive knowledge could be based on:
- A searcher’s previous text entry
- On-screen information
- Device sensors