Apple on Tuesday was reassigned a patent from Topsy Labs, the social analytics firm it purchased in 2013, covering a computer search system that cuts through the noise often found in crowdsourced recommendation services by surfacing search results based on user ranking and influence.
Source: USPTO
Published by the U.S. Patent and Trademark Office, Apple's U.S. Patent No. 9,135,294 for "Systems and methods using reputation or influence scores in search queries" was originally assigned to Topsy Labs, a social network analytics firm that came under Apple's umbrella in late 2013. The patent is one of many that offers some insight into what Apple is planning to implement, or has already implemented, into its products as part of the acquisition.
This particular invention looks to solve the problem of presenting search users with reliable, trustworthy information, arguably one of the biggest problems facing services that aggregate user-sourced reviews and ratings. Someone looking for restaurants on Yelp, for example, might find a glowing review by "User A," only to find a contradictory one-star rating from "User B."
Further, without reliable information about a reviewer, and in turn their knowledge of a subject (restaurant, book, album or other reviewed item), it is difficult to assign appropriate weight to a given rating. The document boils the issue down to one of two computational realities, classifying one node as subjects (reviewers) and another as objects (reviewed items). By calculating associations between subject and object, Topsy proposes integration with an intelligent reputation platform, another of its inventions.
The system assigns reputation indicators, more commonly known as reputation scores, based on the degree and number of relationships between "reputation targets," or other users. This process is further refined to include experience in sub-environments, or specific topics like "movies," "music," "business" and others.
Conducting a graph search on direct links between subjects, objects, sub-environments, community feedback data, platform guidelines and other metrics results in finely tuned search query responses that are both relevant and, hopefully, more accurate than what current systems can provide.
Applied to Apple's current product lineup, Topsy's patent could see immediate use in iTunes and the App Store, or perhaps as a crowd-sourced Apple Music search engine. It is thought that Apple is currently leveraging a modified version of Beats Music's curation methodology, powered by humans and smart algorithms. The revamped Spotlight search, now called Siri Suggestions in iOS 9, is another service ripe for integration, especially as Apple delves into real-time location-based recommendations.
Apple might also use the patent to tap into Twitter, Topsy's bread-and-butter, for real-time recommendations on a litany of topics, but for now the company relies on data gathered from Yelp, TripAdvisor and other such services.
Apple's reassigned reputation-based search engine patent was first filed for in January 2014 and credits Rishab Alyer Ghosh and Vipul Ved Prakash as its inventors.