- From: Karl Dubost <karl+w3c@la-grange.net>
- Date: Wed, 24 Mar 2010 10:05:55 -0400
- To: public-xg-socialweb@w3.org
We will love our social network future. Our data in silos ;) Well, it depends where you are on the cable. In Investigation software snoops through social networks - Techworld.com At http://news.techworld.com/applications/3214195/investigation-software-snoops-through-social-networks/ Hone works in tandem with NarusInsight. By Nucci's own admission, however, it can do some pretty "scary" things. The software's user creates a target profile, and Hone then proceeds to link what Nucci calls "islands of information." Hone can analyse VOIP conversations, biometrically identify someone's voice or photograph and then associate it with different phone numbers. "I can have a sample of your voice in English, and you can start speaking Mandarin tomorrow. It doesn't matter; I'm going to catch you." It uses artificial intelligence to analyse emails and can link mails to different accounts, doing what Nucci calls topical analysis. "It's going to go through a set of documents and automatically it's going to organise them in topics. I'm not talking about keywords as is done today, I'm talking about topics," he said. That can't be done with today's technology, he said. "If you search for fertilisers on Google... it's going to come back with 6.5 million pages. Enjoy," he said. "If you want to search for non-farmers who are discussing fertiliser... it's not even searchable." Hone will sift through millions of profiles searching for people with similar attributes, blogger profiles that share the same email address, for example. It can look for statistically likely matches, by studying things like the gender, nationality, age, location, home and work addresses of people. Another component can trace the location of someone using a mobile device such as a laptop or phone. Bit by bit, it pieces together the subject's different identities on the Internet. Narus is still testing the waters with Hone. Working with a consortium of universities, the company has used Hone to sift through massive amounts of public information. "We started to collect data three years ago and we've gone through several programs," Nucci said. "We have something like 75 million users in our system." With the permission of users, Nucci's team also analyzed data on about 50,000 private profiles. -- Karl Dubost Montréal, QC, Canada http://www.la-grange.net/karl/
Received on Wednesday, 24 March 2010 14:06:16 UTC