I’m about midway through my India journey. I don’t have much time travelling between conferences in Chennai and Coimbatore, so my post is short. The wifi at the University is also spotty, which makes connectivity a challenge. I’m posting on Tuesday (not my usual Wednesday), as tomorrow is a travel day.
Chennai seems to be bouncing back from the floods. The ICMCE conference was almost cancelled, and a speaker did not come owing to a US travel ban to Chennai. The floods did impact attendance. But the organizers charged on and I am here now.
At the beginning of the conference, the custom here is to light a candle which banishes ignorance. It was a beautiful and solemn way to begin the conference.
My talk was on Tuesday morning and went well. I spoke on dimensionality of on-line social networks like Facebook and LinkedIn. The Logarithmic Dimension Hypothesis is something I co-discovered; it’s more of a thesis than a conjecture. It posits that the dimension (i.e. smallest number of attributes needed to uniquely identify agents in the networks) is a logarithmic function of the number of agents. So in a big network like Facebook, even 6 or 7 dimensions would suffice to embed the graphs accurately.
We know very little about so-called Blau space, and the dimensionality of it was the theme of my talk. Blau space views a social network as embedded in a multi-dimensional space, with nodes close in the space if they have similar attributes. A deeper knowledge of Blau space may help us with everything from community detection, to modelling viral marketing in social networks.
Tomorrow I leave for Coimbatore for a keynote at a second conference at PSG College of Technology. I have visited there three times before, and look forward to seeing old friends at the college. First, I will speak at Stella Maris College in Chennai on Cops and Robbers games.
My time in India is always too short, but I enjoy the friendliness of the people and the great, spicy food. I miss it when I leave.
Anthony Bonato
We always love to hear graph theory from you.
LikeLike