I recently gave a keynote address on Science and Leadership for the Future to a small group of major media and corporate clients of New Scientist magazine.
Given the context, I was able to delve a little deeper into the issues than I would for most audiences.
The video of my presentation was sliced into a number of brief segments. Below is the video of the section of my presentation on Networks.
Here is a summary of the points made in the video:
* Networks are fundamental not just to our communications but to many aspects of our lives.
* There are many network topologies. One of the most important is ‘scale-free’ networks, in which the structure is identical irrespective of its size.
* The internet has maintained the same scale-free structure throughout its growth over the last 21 years.
* Scale-free networks develop through ‘preferential attachment‘, in which better-connected nodes tend to get more connections, in a version of ‘the rich get richer’.
* Gaussian curves describe so-called ‘normal distributions’. Financial market variations are often assumed to follow normal distributions, but in fact they often exhibit heteroskedasticity, meaning that events that are thought to be unlikely in fact happen far more than expected. In fact these ‘black swans’ are increasingly likely.
* Scale-free networks, in contrast to Gaussian distributions, have power-law distributions, in which a few nodes have many connections, and down an exponential curve, many have few connections, a concept popularized as ‘the long tail.’
* There are many factors, such as positive feedback loops and self-organisation, that are shifting systems towards power-law distributions.
* This is happening in organisations today. As we move from hierarchical structures that define specific and limited communication paths, to open communication across many levels and functions, organisations are shifting to power-law distributions.
* The shift towards scale-free networks is visible across many domains, including social networks, industries, and politics. However these types of networks are also evident in biological systems such as cells and brains.
* In influence networks, messages and ideas can ‘cascade’ to reach many people. The network characteristics that shape whether and how messages can cascade, or to use more popular jargon, ‘go viral’, include network density and network homogeneity.
* Cascades are hard to engage, with well-connected influencers with heterogenous networks the most likely source of widespread message dissemination.