Tuesday, August 04, 2009


I remember writing a blog post sometime back on the phenomenon of symmetry that characterizes our social lives and actions. This was a time prior to when I started working on the "social synchrony" problem back in last December. It's interesting on how on the same note, I am interested in the phenomenon of asymmetry, so apparent in the social relationships we inculcate in our lives!

Characterizing relationships in online social networks is of my current research interest; and I plan to look at it from the manner individuals communicate over these different media. However as most of the literature in the social sciences goes, people have mostly considered undirected edges, that is, assuming relationships are symmetric. While this can be useful to simplify the network representation, or to compute elegant network features, I am driven by my simple life experiences; and the experiences that say that 99% of the time relationships are not symmetric!

Sometime back, I heard someone saying that in a non-platonic relationship, one person always has the upper hand than the other; and that one person has the upper hand to terminate it over the other, i.e. is in a more advantageous position than the other; and also that the implications of termination are likely to be worse to one compared to the other. I don't have a good reason to think otherwise. I think while this is most true about non-platonic relationships, there could be validating evidences in most other relationships, like parents-kids, brother-sister, employee-employer etc, however good friends you can mutually be!

In current state-of-the-art, unfortunately relationships in social networks, have been over-simplified, especially when it comes to characterizing them. I would imagine that though Alice and Bob both label each other as "friends" on Facebook, it is likely that the properties of that friendship are not symmetrical to each. Such asymmetry can for example, be explicitly found in their communication patterns or can be implicit in evidences like who is more likely to follow the other - revealing sort of a dependency relationship from Alice to Bob, which might not be the case from Bob to Alice. These are just some brain-storming thoughts of characterizing the asymmetry. There could be more...

Basically this thought springs from two causes, and both of which interestingly clicked to me today in the train back home: (a) that most of the times relationships have been quite asymmetric for me, the context being different every time, and (b) social ties have rather been poorly characterized so far - we need a more realistic model that captures the directional dynamics between a pair of people, specifically based on their communication and their actions in the social front.

However, finally while I do think modeling the asymmetry is useful, how is homophily, i.e. the fact that you and your friends are alike in some manner, related to this asymmetry? Maybe a future post!

No comments: