Shaul Levi

Recently I see a lot of new initiatives for social search some of them are of large companies like Google that its claim to fame initially was ranking, now is talking about a social ranking of it's results. Also a bunch of small startups that the opportunity of a new ways to rank data based on social graph to create more relevancy and a result that is part of the circle of trust. The Idea of me getting results based on my friends user generated content is important. Actually when you look at what we are looking in real life to find it’s actually what our friends are doing, it is very important to me that my friend just bought a new pair of Nike more than the fact that the brand is less considered today or they used more for walking than running... So when I am using social search what do I want to get my friends or the data they have written??? The question is a tough question considering the fact that all the ranking is done using the social graph that is inaccurate as it is not u

During my recent work on creating a searching tool that uses the information from the social graph to rank the search results I came across a major problem... the social graph as you define in your social network service is accurate today but inaccurate tomorrow... and you have some people there that you barely know and some other that aren't there but happened to be the closest people to you... would ranking based on what I put in facebook myspace linked or others is the right way to rank results... how do I keep it accurate updated and what is the right or most relevant way for me to approach people I want to get introduced. Definitely there is a need for technology with better algorithms that can learn and anticipate the right results, the group of algorithms that can cover that end are behavioral algorithms that have a learning machine with success parameter in them. these set of algorithms are used to measure human behavioral interaction with the search results anticipate t