{"id":634044,"date":"2020-02-04T08:24:48","date_gmt":"2020-02-04T16:24:48","guid":{"rendered":"https:\/\/www.microsoft.com\/en-us\/research\/?p=634044"},"modified":"2020-02-04T08:24:48","modified_gmt":"2020-02-04T16:24:48","slug":"the-personal-web-connecting-information-for-better-search-and-recommendation","status":"publish","type":"post","link":"https:\/\/www.microsoft.com\/en-us\/research\/blog\/the-personal-web-connecting-information-for-better-search-and-recommendation\/","title":{"rendered":"The personal web: Connecting information for better search and recommendation"},"content":{"rendered":"
<\/p>\n
In the digital era, almost everyone struggles with the mountains of information they have. Every activity in our lives seemingly generates more information: emails, files, receipts, photos\u2014the list goes on. We have trails of digital information from each of our work projects, vacations, hobbies, and kids\u2019 schools, including websites, files, and calendar appointments, not to mention contacts\u2014the people we work, play, and live with in each of these spaces. We spend a significant amount of time figuring out when to share something with them, when to ask them for something, and where to find information they might have already sent us. Don\u2019t you wish there were a way to automatically and effortlessly organize everything? To have search as simple as people\u2019s associative memory\u2014the minute you start looking at one item all of the related information you need is magically at your fingertips?<\/p>\n
This is just the problem we set out to solve with University of Michigan PhD student Tara Safavi (opens in new tab)<\/span><\/a>, our Microsoft product partners and research colleagues, and Safavi\u2019s PhD adviser.<\/p>\n To make the challenge more amenable to algorithmic analysis, we view all the information people interact with as a personal web<\/em> of entities surrounding them. Just like the information on the World Wide Web, these entities\u2014contacts, projects, files, emails, and other digital information\u2014can be thought of as nodes in a graph connected through how they interact by edges. For example, a project may have different project members, a file may have different authors or readers, and an email has senders and recipients. In addition to proposing a graph-based approach to personal information, we also introduce novel computationally efficient techniques for updating how information propagates through a graph as the graph evolves with new edges, nodes, and information. Safavi will be presenting the paper on this work, \u201cToward Activity Discovery in the Personal Web (opens in new tab)<\/span><\/a>,\u201d at the 13th ACM International Conference on Web Search and Data Mining (opens in new tab)<\/span><\/a>.<\/p>\n While we envision this graph-based approach as a solution for organizing all aspects of one\u2019s life, in our study, we applied it to people\u2019s personal web at work\u2014the emails, documents, and contacts individuals access in their professional lives. You might be wondering why graph connections are needed at all when many digital entities have content. Looking at people, or contacts, provides one clear reason. By simply viewing personal information entities as a graph and looking at the number of connections each type has, it\u2019s clear that contacts are the hubs<\/em> of personal webs.<\/p>\nThe case for graph connections and how to make them<\/h3>\n