Technology
Cascaad has developed patent-pending social-graph and semantic relevance ranking technologies that specifically address the nature of the real-time social web, making it easier to organize and surface personally-relevant news, conversations and other topics.
Information discovery so far…
Two dominant modes of information discovery have characterized the web 2.0 era so far:
- Wisdom of the friends: the people we connect to can act as a source of recommendation and filtering, for instance through their posts and retweets on Twitter.
- Wisdom of the crowd: all users contribute on the same footing in deciding what are the most popular items, e.g. through a global count of the number of times a message gets liked or shared.
…And its obvious limitations
- Friends sometime can help in validating content relevance, but who’s validating their judgment, and what about the vast majority of content that doesn’t catch their attention?
- Crowds are great, but should a post about international policies by Barack Obama have the same weight as that of all other users? And should users who more closely match my specific interests and relationships have more influence in the content rankings that affect me?
Going beyond the current approaches
Cascaad exploits the “wisdom of the graph”, where each user has a different influence on all other users that is dictated by the observed social paths that join them. The “wisdom-of-the-graph” method behind Cascaad’s real-time social ranking provides the basis for a massively-scalable personalized filtering of relevant items that reflects an individual’s extended sphere of influence. In order to increase statistical significance, this scoring takes into account all types of explicit and implicit cues found in activity streams about the attention received by a message. Moreover, the Cascaad engine also involves automatic semantic profiling to merge into the social graph the interest graph for individual users and thus further distill the items that match their personal interests.
What makes Cascaad’s technology unique
As a result, the key advantages of Cascaad compared with existing content discovery services are:
- filtered content is personalized to match a user’s specific interests and social affinities;
- an individual item’s relevance is picked up quickly by Cascaad’s algorithm so it is rapidly made available;
- items comprise not only news and other links, but also conversations and specific topics of interests (e.g. a book, a movie, etc);
- social messages are automatically enhanced with rich contextual information.
Although Cascaad provides relevance-ranked real-time search, the most natural modality for current information consumption of content on the real-time Web is push rather than pull and push-based personalized discovery is indeed where we are focusing our attention!
